Subscribe now

Environmental exposures play a central role in shaping both the microbiome and immune responses, influencing asthma susceptibility and progression. Major exposures include allergens, air pollution, viral infections, tobacco smoke, diet and obesity, occupational and indoor irritants, as well as climate-related factors

Share

Airway inflammation in asthma is typically diffuse but can appear patchy, involving both the large (central) and small (peripheral) airways. In classic allergic asthma, eosinophilic inflammation predominates, primarily affecting the bronchial walls from central bronchi to terminal bronchioles. In contrast, neutrophilic asthma is dominated by neutrophil infiltration. Regardless of the inflammatory type, chronic untreated inflammation eventually leads to airway remodeling—permanent structural alterations that contribute to airflow limitation and airway hyperresponsiveness

Subscribe now

The asthmatic inflammatory milieu is sustained by a complex cast of immune and structural cells orchestrated by a network of cytokines and lipid mediators. Traditionally, asthma has been viewed as a Th2 cell–driven disease. In many patients, CD4⁺ Th2 lymphocytes play a central role by secreting the canonical type 2 cytokines IL-4, IL-5, and IL-13, which drive IgE production, eosinophilia, and goblet cell hyperplasia, respectively. It is now evident, however, that a wide array of both innate and adaptive immune cells contribute to the asthmatic response, particularly in non-T2 phenotypes

Subscribe now

The cytokine networks in asthma are extensive, but some key pathways include

Share

The engagement of receptors on the surface of immune and structural cells in asthma initiates a cascade of intracellular signaling pathways. These interconnected networks regulate cellular activation, cytokine gene transcription, proliferation, and survival, thereby sustaining the chronic inflammation and airway hyperreactivity characteristic of the disease. Several major signaling mechanisms have been identified.

Subscribe now

Asthma is increasingly recognized not as a single disease entity but as a heterogeneous syndrome encompassing multiple phenotypes and underlying endotypes. Over the past decade, a paradigm shift has occurred from a purely clinical classification to one based on identifiable phenotypic clusters and distinct molecular endotypes. This approach aims to enable precision therapy by targeting the specific disease mechanisms present in each individual, a concept referred to as “treatable traits.”

Share

Asthma endotypes from research to practice

Broadly, asthma endotypes are often categorized according to the presence or absence of type 2 (T2) inflammation

Subscribe now

Phenotypes and endotypes of asthma

Given this heterogeneity, the framework of “treatable traits” has been proposed to guide individualized management. Treatable traits are clinically or biologically defined characteristics that can be identified and specifically targeted in each patient. These may be pulmonary (e.g., eosinophilic inflammation, airway hyperreactivity, small-airway dysfunction), extrapulmonary (e.g., obesity, chronic rhinosinusitis, anxiety), or behavioral (e.g., poor adherence, incorrect inhaler technique)

Share

Conventional pharmacotherapy

The foundation of asthma management has long relied on inhaled pharmacotherapy designed to suppress chronic airway inflammation and alleviate bronchoconstriction. These conventional treatments are generally divided into two main categories: controller medications, which are taken regularly to maintain long-term disease stability, and reliever medications, which are used as needed to rapidly relieve acute symptoms

Subscribe now

Currently available biologics for asthma

The era of biologic therapy in asthma began with the recognition of IgE and eosinophils as key drivers in subsets of patients. Biologics are typically monoclonal antibodies designed to neutralize specific cytokines or to block cell surface receptors, thereby interrupting discrete inflammatory pathways . They have shown remarkable efficacy in reducing exacerbations and steroid requirements and improving lung function and quality of life in patients with severe T2-high asthma.

Subscribe now

Source: https://doi.org/10.1038/s41392-026-02853-8

What exactly is so scientific about political science and social sciences? 😉

Perhaps medicine has spoiled me.

When somebody calls something “science,” my medical brain instinctively starts asking inconvenient questions.

Share

What is your hypothesis? Where is your control group? What are the confounders? Can I reproduce your findings? And, most importantly, what is the level of evidence?

This works reasonably well when studying a drug.

Politics, however?

Imagine submitting an election as a clinical trial…!

Intervention: charismatic politician.
Controls: unavailable.
Blinding: absolutely impossible.
Confounders: unemployment, inflation, caste, class, religion, ideology, war, weather, scandals, social media, historical grievances, celebrity endorsements, WhatsApp forwards and whatever somebody says at 10:47 p.m. the night before polling.

Primary outcome: Who won?

Authors’ conclusion: Further research is required.

As a medical educationist, I would probably send the manuscript back asking for major revisions…!!

And sociology isn’t much kinder to the experimentalist…!

The fundamental particles of sociology are human beings, and human beings possess one deeply unscientific characteristic:

They know they are being studied (most of the times).

Subscribe

An electron does not read the latest paper about electron behaviour and announce:

“I find this model reductionist and shall behave differently henceforth.”

Humans can.

Tell people that consumers behave according to Pattern X and consumers may change. Tell politicians that voters respond to Strategy Y and every politician adopts Strategy Y. Voters notice. Journalists notice that voters have noticed. Politicians notice that journalists have noticed that voters have noticed.

Your beautifully significant p value has now entered intensive care…!!

Then comes my favourite problem:

Reproducibility.

Suppose political scientists conclude that Policy X caused Party A to win an election.

Wonderful.

Now reproduce it.

Same country. Same candidates. Same voters. Same economy. Same scandals. Same international environment. Same weather. Same television debates etc..

Except this time, remove Policy X.

Unfortunately, history does not have a “repeat experiment” button.

You cannot reload India 2014 from a previous save point.

At this stage, the clinician in me becomes suspicious.

Share

But then the medical educationist in me whispers something uncomfortable:

“Professor, your own field isn’t quite that tidy either.”

Touché.

Medicine knows this problem rather well.

Patients are not identical laboratory mice. Children certainly aren’t. The same intervention does not produce precisely the same outcome in every child. Genetics differs. Nutrition differs. environments differ. Families differ. Adherence differs etc..

And medical education is even closer to the social sciences. Try proving that a teaching method “caused” better doctors twenty years later while controlling for intelligence, motivation, teachers, institutions, clinical exposure, culture and approximately half of human existence.

Share

Good luck with the randomized controlled trial….!!

So perhaps demanding that every legitimate science resemble experimental physics is itself unscientific.

Astronomers cannot reproduce a supernova on request. Geologists cannot rerun the Himalayas. Epidemiologists cannot recreate a pandemic simply because Reviewer Two requested replication.

Science is not merely about repeating phenomena.

It is about method: systematic observation, measurement, hypothesis testing, transparent analysis, reproducible methods and, crucially, allowing evidence to prove us wrong.

Subscribe

By that definition, political scientists and sociologists certainly “can do science.”

But this creates another delightful problem.

If scientifically studying something makes the subject itself scientific, then religion must also be science. 😂

We can measure religious attendance, belief, demographics, behaviour and associations with health or voting.

Yet nobody would suggest that running a regression analysis on prayer converts theology into biochemistry…😉

And therein lies the distinction.

Share

Politics isn’t science. Society isn’t science. Religion isn’t science.

But all three “can” contain phenomena that are investigated scientifically….!

Perhaps, then, my discomfort is really with the prestige carried by the word science.

Science sounds wonderfully authoritative.

But the farther we travel from molecules toward human beings, institutions and societies, the more conditional our conclusions become.

Medicine taught me something valuable here: evidence does not abolish uncertainty; good science tells us how uncertain we should be.

Perhaps that is the standard we should demand from every discipline bearing the name.

So I have made peace with Political Science and Social Science.

“Mostly”….

I will allow them the word science—with the same professional instinct with which I read the limitations section of a medical paper.

Political Science*

*Results may vary according to country, century, culture, economy, institutions, ideology, religion, charismatic leaders, irrational voters, social media, historical accidents and events nobody included in the original protocol.

Confidence interval: rather wide.

Further research, naturally, is required. 😉

A large amount of useful information in healthcare is written as free text, such as:

• clinical notes

• case sheets

• discharge summaries

• radiology and pathology reports

• nursing notes

• adverse-event reports

• research interviews

• student feedback

• reflective writing

• examination comments

• patient feedback

• scientific articles

These narratives often contain important information.

But there is a problem.

Two people may read the same text and reach different conclusions. An AI system may also interpret the same text differently from a clinician or researcher…!

This does not always mean that one person is right and the other is wrong.

There are several different reasons why disagreement occurs.

Understanding why disagreement occurs is important before deciding whether the problem lies with:

• the data,

• the person interpreting the data,

• the AI system,

• the definitions being used,

• or the question itself.

The different causes can be understood through six broad categories: A to F. Each category below is illustrated with one representative example; in practice, each one covers a family of related sub-patterns rather than a single fixed pattern — the point of the framework is to correctly diagnose which family a given disagreement belongs to, not to memorise an exhaustive list.

Subscribe now

A. The Information Is Missing

The problem

Sometimes the information required to make a decision simply is not present in the text.

No amount of careful reading or sophisticated AI can recover information that was never recorded.

Medical example

A case sheet states:

“Patient developed reaction after medication.”

But it does not mention:

• which medication,

• what reaction occurred,

• how soon it occurred,

• whether another drug was given,

• or whether the reaction was considered allergic.

A researcher cannot confidently classify this as an adverse drug reaction.

Another example

A student’s feedback says:

“The clinical posting was not good.”

We do not know whether the student means:

• poor teaching,

• inadequate patient exposure,

• excessive workload,

• lack of supervision,

• poor organisation,

• or something else.

Key question

Is the information required to answer the question actually present?

If not, the correct response may be:

“Insufficient information.”

It is better to acknowledge uncertainty than to invent an interpretation.

Share

B. The Information Is Present, but We Miss It

The problem

Sometimes all the required information is available, but the reader overlooks, misunderstands, or gives too much importance to one part of the text.

This is mainly an attention or interpretation error.

Medical example

A discharge summary initially states:

“Suspected acute myocardial infarction.”

Later it states:

“Serial troponins negative; final diagnosis: acute pericarditis.”

If someone reads only the beginning of the report, the patient may incorrectly be classified as having myocardial infarction.

Research example

An abstract may initially report:

“The intervention significantly improved blood pressure.”

But the discussion later states that the difference was only 2 mmHg and was not clinically meaningful.

Reading only the statistical result may lead to an exaggerated conclusion.

Medical education example

A faculty assessment says:

“Student communicates well with patients but repeatedly fails to recognise critically abnormal investigation results.”

If the positive comment receives more attention, the major competency problem may be overlooked.

Key question

Was the answer present in the text, but overlooked or incorrectly weighted?

If yes, the problem is not lack of data. It is the quality of interpretation.

Share

C. The Meaning Becomes Clear Only When We Look at the Wider Dataset

The problem

Sometimes one clinical note or response is difficult to understand on its own.

But when we examine many related records together, a pattern becomes clear.

Medical example

Different case sheets may contain:

• “HTN uncontrolled”

• “BP poorly controlled”

• “persistent high BP”

• “BP not responding”

• “uncontrolled hypertension”

Individually, these statements look different.

Across hundreds of records, however, they may represent the same underlying clinical issue:

Poorly controlled hypertension.

Research example

Consider student feedback collected after a curriculum change:

• “Too many sessions”

• “No time to study”

• “Schedule very packed”

• “Classes from morning to evening”

• “Need more self-learning time”

Taken separately, these look like different complaints.

Viewed together, they reveal a common theme:

Curriculum overload.

Why this matters

If every response is analysed independently, the researcher may count five different problems instead of one common problem.

Key question

Does this statement make more sense when compared with other records in the dataset?

This is particularly important when analysing:

• electronic health records,

• qualitative research,

• patient feedback,

• student feedback,

• incident reports,

• large clinical datasets,

• and AI-based text analysis.

  Share

D. External Medical Knowledge Is Required

The problem

Sometimes the information is clearly written, but understanding it requires knowledge that is not contained in the text itself.

This may include:

• clinical knowledge,

• guidelines,

• diagnostic criteria,

• medical abbreviations,

• institutional terminology,

• regulatory definitions,

• or speciality-specific conventions.

Medical example

A clinical note states:

“CR achieved after cycle 2.”

An oncologist may immediately understand CR as Complete Response.

Someone without oncology knowledge may not understand the meaning.

Another example

A note states:

“G3P2L2 at 38+2 weeks.”

An obstetrician understands this immediately.

A general-purpose AI system without sufficient medical context may misinterpret the notation.

Research example

A study may report:

“HbA1c = 6.6%.”

To interpret whether this supports a diagnosis of diabetes, external diagnostic criteria are required.

The number alone does not provide the complete interpretation.

Medical education example

A student is described as:

“At the Shows How level.”

Understanding this requires knowledge of Miller’s Pyramid of Clinical Competence.

Key question

Do we need specialist knowledge, guidelines, definitions, or institutional conventions to understand this correctly?

If yes, the solution is not necessarily a better AI model.

Sometimes what is needed is simply better contextual information.

Share

E. More Than One Interpretation May Be Reasonable

The problem

Some narratives genuinely support two or more possible interpretations.

Medicine frequently contains this type of uncertainty.

Trying to force every statement into one category may therefore be misleading.

Clinical example

A patient presents with:

• fever,

• tachycardia,

• hypotension,

• elevated inflammatory markers,

• and pulmonary infiltrates.

The deterioration could be related to:

• bacterial pneumonia,

• sepsis,

• viral infection,

• aspiration,

• or more than one condition occurring simultaneously.

A single label may oversimplify the situation.

Medical education example

A student writes:

“We need more clinical exposure.”

This could mean:

• more patients,

• more bedside teaching,

• more independent examination,

• more procedures,

• or more time in clinical postings.

Several interpretations may be valid.

Research example

An interview participant says:

“I stopped taking the tablets because they were difficult.”

“Difficult” might refer to:

• side effects,

• swallowing difficulty,

• cost,

• frequency of dosing,

• availability,

• or lack of perceived benefit.

Without further information, assigning one reason may create false certainty.

A better approach

Instead of asking:

“Which single category is correct?”

sometimes we should ask:

“Which interpretations are plausible, and how confident are we in each?”

This introduces the idea of probabilistic interpretation rather than forced classification.

Subscribe now

F. The Interpretation Depends on Perspective

The problem

Sometimes everyone agrees about the facts, but they still disagree about what those facts mean.

This is because interpretation depends on values, priorities, or professional perspective (ontology and epistemology).

More data may not completely solve the disagreement.

Medical example

Suppose a hospital reports:

“Average length of stay decreased from 6 days to 4 days.”

One administrator may conclude:

“Efficiency has improved.”

A clinician may ask:

“Are patients being discharged too early?”

A health economist may see:

“Reduced healthcare expenditure.”

A patient-safety researcher may ask:

“Did readmission rates increase?”

The number itself is the same.

The interpretation depends on the question and perspective.

Medical education example

Suppose examination pass rates increase from 75% to 95%.

One faculty member may conclude:

“Teaching has improved.”

Another may ask:

“Was the examination easier?”

Another may wonder:

“Have assessment standards fallen?”

All are looking at the same result.

Key question

Are we disagreeing about the evidence, or about what the evidence means?

This distinction is extremely important.

Share Vikram Sakaleshpur Kumar

The Six Problems at a Glance

Subscribe now

The Three Deeper Groupings

The six letters are useful as a checklist, but under pressure — mid-chart-review, mid-M&M-discussion — six categories is a lot to hold in working memory.

It helps to know that A–F collapse into three more fundamental types of problem, each demanding a different kind of fix:

• A and B — a data-quality problem. Either the information was never captured, or it was captured but the reader didn’t use it properly. The fix is better documentation practice, closer reading, or quality control — not a new classification scheme.

• C and D — a missing-context problem. The individual entry is genuinely under-interpretable in isolation; the fix is pooling records or supplying the specialist knowledge/glossary needed, not blaming the reader or the model for a failure that was never theirs to solve alone.

• E and F — an irreducible uncertainty or values problem. The facts and definitions may be entirely clear, but no additional data will produce a single correct answer — either because the text genuinely supports several readings, or because the disagreement is about what matters, not about what happened. Here the honest output is a distribution or an explicit statement of standpoint, not a forced single label.

  

A quick gut-check when disagreement shows up: is this a quality problem, a context problem, or a values problem? That question alone resolves most of the confusion about whether disagreement is something to fix, something to enrich, or something to simply report honestly.

Share

A Pediatric-Specific Lens: The Proxy Reporter Problem

An infant cannot narrate symptoms at all. A toddler’s distress is filtered entirely through a caregiver’s interpretation before it ever reaches paper. Even a competent teenager’s own account may be edited, softened, or replaced by a parent sitting beside them, or omitted deliberately for reasons of confidentiality. So before a pediatric note or transcript can be classified into A–F, there is a prior question worth asking explicitly:

Whose voice is actually in this text — the child’s, the caregiver’s, or the observer’s — and does that change what the words mean?

This single question resurfaces inside every one of the six categories:

• A — Missing information. “Baby not himself today” tells us nothing about feeding, fever, activity, or cry quality, and critically, it doesn’t tell us who is speaking — a first-time anxious parent and an experienced NICU nurse mean very different things by the same sentence.

• B — Overlooked information. A subtle change in cry character or feeding pattern, noted mid-way through a long triage note, is exactly the kind of detail an overstretched reader (or a rushed AI summariser) skips past — with disproportionate consequences in a population that cannot self-report deterioration.

• C — Wider dataset required. A single weight or height means very little in isolation; it only becomes meaningful against a growth curve — the child’s own trajectory across visits, or population percentile charts. This is category C in its purest pediatric form: the individual data point is uninterpretable, the trend is everything.

• D — External knowledge required. “Not meeting milestones at 18 months” is meaningless without developmental-milestone reference charts (WHO/CDC or local equivalents). NICU and PICU notes compress entire clinical pictures into shorthand — “Grade 3 IVH,” “Stage 2 ROP,” “on HFNC 6L/40%” — that are precise to a neonatologist and opaque to almost anyone else, including a general-purpose AI model.

• E — Multiple valid interpretations. “Fussy, poor feeding, unsettled” could plausibly represent normal infant colic, GERD, cow’s-milk protein allergy, an early viral illness, or — and this is where pediatrics carries extra weight — an early safeguarding concern. Forcing this into a single confident label is exactly the situation where reporting a distribution of plausible explanations is more honest than manufacturing certainty.

• F — Perspective-dependent. “Child is attention-seeking” reads completely differently to a busy ward nurse, a parent, and a child psychologist. “School refusal” can be framed as a behavioural problem, an anxiety disorder, or an undiagnosed learning difficulty depending entirely on who is doing the framing — the facts (the child isn’t attending school) are identical; the meaning is not.

  

  Leave a comment

Why this deserves special attention from your committee, not just your classroom

The safeguarding example under Category E is not a hypothetical. Ambiguous documentation of an injury — “bruising, explanation inconsistent” — sits at the exact intersection of C (does this pattern recur across the child’s visit history, or across siblings?), D (mandated reporting frameworks and non-accidental-injury criteria are external conventions, not inferable from the note alone), and E (the note itself may genuinely support more than one explanation, and prematurely collapsing it into a single confident label — in either direction — carries real consequences). This is precisely the kind of interpretive disagreement an ethics committee should want documented as a type of uncertainty, not adjudicated away as someone’s error.

A teaching example for your OSCE and assessment audience

History-taking stations built around a parent-as-informant, rather than the patient, are assessing something subtly different from adult-medicine history stations: not just what the trainee elicits, but whether they can triangulate an account that has already passed through someone else’s interpretation before it reached them. An examiner’s comment — “struggled to get history” — is itself an ambiguous piece of narrative data (is this Category B, the trainee missed cues that were present; or Category D, the trainee lacked a technique for eliciting information through a reluctant or over-protective proxy?) — which is a fitting, slightly recursive illustration of the whole framework applied to the assessment of the framework’s own subject matter.

Subscribe now

A Simple Method for Medical Researchers

Whenever two researchers, clinicians, or AI systems disagree about a piece of narrative data, do not immediately ask:

“Who is correct?”

Instead, ask these six questions in order.

Subscribe now

Why This Matters for AI in Medicine

This framework becomes particularly important when using large language models and other AI systems to analyse medical text.

Suppose an AI incorrectly classifies a clinical note.

It is tempting to say:

“The AI made an error.”

But we should first determine what kind of problem occurred.

Subscribe now

An Important Lesson for Medical Research

Researchers often assume that every observation has a single ground truth.

For laboratory measurements, this may sometimes be reasonable.

But narrative medical information is different.

For example:

“Patient is doing better.”

What exactly does better mean?

It might refer to:

• pain,

• mobility,

• laboratory values,

• oxygen requirement,

• mood,

• functional status,

• or the clinician’s overall impression.

Therefore, before measuring whether an AI system agrees with a human expert, researchers should first ask:

Is there actually one objectively correct interpretation of this statement?

This question is especially important when developing AI systems using clinician-labelled datasets.

Subscribe now

Inter-Rater Agreement Does Not Tell the Whole Story

In medical research, we frequently calculate agreement between observers using measures such as:

• percentage agreement,

• Cohen’s kappa,

• Fleiss’ kappa,

• or intraclass correlation coefficients.

Low agreement is often interpreted as poor performance.

But this framework suggests something more important.

We should first understand why the observers disagree.

For example, low agreement could occur because:

1. important information is missing;

2. one observer overlooked information;

3. observers saw different amounts of contextual data;

4. one observer had greater specialist knowledge;

5. several interpretations were genuinely possible;

6. the classification itself depended on judgement.

These situations should not all be treated as the same type of error. A low kappa caused by Category A (missing data) calls for better documentation; a low kappa caused by Category E or F (genuine multiplicity or values) may not be a flaw in the study at all — it may be an accurate reflection of the subject matter, and forcing it toward a higher kappa risks manufacturing false certainty.

Share

Implications for Medical Education

The framework is also useful when evaluating students.

Consider the statement:

“The student is not clinically competent.”

Before accepting this judgement, we should ask:

• What behaviour was actually observed?

• Was sufficient clinical exposure available?

• Was important evidence overlooked?

• Which competency definition was used?

• Were different assessors using different standards?

• Could the student’s performance reasonably fall into more than one category?

• Is the judgement influenced by assessor expectations?

This is why good assessment requires:

• clearly defined competencies,

• multiple observations,

• multiple assessors,

• adequate clinical context,

• and explicit assessment criteria.

  Subscribe now

Implications for Qualitative Medical Research

This framework is particularly valuable in qualitative research.

When two researchers code the same interview differently, disagreement should not automatically be considered a mistake.

Instead, the research team can ask:

What type of disagreement is this?

It may arise because:

• the participant did not provide enough information;

• one researcher overlooked part of the transcript;

• another interview provides important context;

• specialist cultural or clinical knowledge is required;

• the statement supports several themes;

• or researchers are approaching the data from different theoretical perspectives.

This makes disagreements analytically useful rather than merely inconvenient.

Share

The Central Message

When humans or AI disagree about medical narrative data, there are at least three very different possibilities:

1. There is a correct answer, but somebody missed it.

This is mainly an accuracy problem.

2. A correct answer exists, but additional context is required.

This is mainly a context problem.

3. There may not be one uniquely correct answer.

This is an uncertainty or judgement problem.

These situations require different solutions.

Subscribe now

A Useful Rule for Medical Academia

Before asking:

“How accurate is the human or AI classification?”

ask:

“Is this information actually classifiable with certainty?”

And before declaring disagreement an error, ask:

“Why did the disagreement occur?”

That small change in thinking can improve:

• clinical data analysis,

• qualitative research,

• medical education research,

• chart review studies,

• AI validation studies,

• clinical decision-support research,

• patient-feedback analysis,

• and institutional healthcare analytics.

  Share

In one sentence

Not every disagreement between clinicians, researchers, or AI systems represents an error; some arise from missing information, missing context, genuine uncertainty, or differences in perspective.

The six-category structure (A–F) is adapted from a framework originally developed by the data-analytics firm Wholesum for interpreting narrative text disagreement across pharma, finance, and other sectors; their original work describes 27 recurring sub-patterns falling within these six categories, of which the examples above illustrate one each. This version reframes the categories and all examples for a medical, research, and education audience.

Same salon. Same owner. Same unspoken arrangement in which he cuts my hair and I keep his family supplied with medical advice, folded into the conversation somewhere between cricket and the weather. I have made peace with the second half of that arrangement. What I have not made peace with is watching it happen in real time and still doing nothing about it…!

This week, midway through the haircut, his phone rang.

His daughter. A medical problem serious enough that he did not think to ask whether I minded. He switched to video and put the phone in my hand. Cape still on. Scissors still mid-air. My hair at that exact unflattering stage where one side is finished and the other looks like it lost an argument.

When it ended, the haircut resumed like nothing had happened, because to him nothing had. The massage happened. The bill came. Five hundred rupees, for the parts of the service that are priced. I paid it. He said thanks on the way out, the same thanks he would give anyone handing him five hundred rupees, the kind that exists to close a transaction rather than acknowledge one.

Subscribe now

The detail that changes the shape of this story is money, or rather the complete absence of a money problem. This man runs five or six businesses alongside the salon, including a real estate agency. He is, by any reasonable measure, doing considerably better than most of the people who queue up at my OPD…!

So the generous explanation, that he simply could not have paid or did not think to, does not survive contact with his bank balance. He was not being stingy. He genuinely did not register that anything billable had occurred.

The haircut has a price because he charges for haircuts. The diagnosis did not, because in his mental ledger it was never a service to begin with. It was just talking. A man who negotiates property deals for a living, who knows exactly what his own time is worth down to the square foot, handed me thirty minutes of clinical responsibility and never once wondered if it cost anything. Not because he is careless with value. Because nobody has ever taught him to see this particular kind of value as a line item.

Share

Economists have a name for this, and it explains more than any amount of professional indignation does.

Darby and Karni, writing in 1973, divided goods into three kinds.

• Search goods, which you can assess before you buy.

• Experience goods, which you can assess after you use. And

• credence goods, which you cannot reliably assess even after you have consumed them. Car repair. Legal advice. Structural engineering. Medicine is the textbook case.

The usual concern with credence goods is fraud — the mechanic who charges for work he did not do. But there is a quieter consequence that nobody warns young doctors about. If the buyer cannot evaluate the work, the buyer frequently cannot see the work.

When a carpenter finishes, the table is the evidence. When a tailor finishes, the shirt is the evidence. When my barber finishes, the evidence is sitting in the mirror looking mildly improved.

When I finish, the evidence is four sentences. And four sentences look precisely the same whether they took ten seconds or twenty years to arrive at.

This is why the pricing follows the visibility, and why the visibility follows the physical. He charged for the massage because he could point at it. He did not charge for the diagnosis because there was nothing to point at, and nobody has ever taught him or most people, that the absence of a visible object is not the absence of work.

Subscribe now

Look at it from his side for a moment, because that is only fair. In his mind this was not a stranger walking in off the street demanding free healthcare. This was a decade of familiarity, of me sitting in that chair every few weeks, of health questions already folded into the conversation alongside cricket and the weather.

Familiarity accelerates this. A stranger books an appointment. An acquaintance sends fifteen reports on WhatsApp. A relative calls at 11:45 PM.Somewhere in that decade my advice stopped being a professional act and became a feature of the friendship, the way some friends always help you move house and others always know a good electrician.

Once something is absorbed into a relationship like that, nobody in the relationship thinks to price it, and asking to be paid at that point feels less like fair compensation and more like a betrayal of the friendship itself. I understand this. I even sympathise with it. It does not change the fact that the friendship extracts in one direction only.

Share

Then there is society's side, because he did not invent this attitude on his own. Ask any Surgeon about weddings. Ask any dentist about flights. Ask any Specialist about family WhatsApp groups. The pattern is identical everywhere. Consultation hours become wherever the healthcare professional happens to be standing, and the request always arrives wrapped in the same disarming phrase, just one small doubt, which has never once in recorded history actually been small.

A lawyer reading two pages of a contract gets billed by the hour and nobody blinks. A doctor reading a child's entire history over video, in real time, under pressure, with a parent's anxiety on the other end of the call, gets a thank you if he is lucky…!

The head massage bundled into that same five hundred rupees probably costs more per minute than the diagnosis did, and nobody finds that arrangement strange, least of all the man who priced both.

And my own side, because I do not get to write this as though I were merely a victim of circumstance. Nobody forced that phone into my hand. I could have said, finish the haircut first, then call me properly and I will sit with this the way it deserves.

I did not say that. I took the call, gave good advice, solved the problem, and only afterward felt something I am now willing to call resentment. Healthcare professionals are catastrophically bad at declining these moments, not because we lack the words but because refusing feels like it might cost someone their health, and we would rather absorb the cost ourselves than risk being wrong about that.

The instinct is admirable in an actual emergency. It is a liability everywhere else, because the people asking cannot tell the difference between the two, and we have never once bothered to teach them.

Subscribe now

None of this is an argument that every hallway consultation should come with an invoice. A society that runs entirely on billing between neighbours is a worse society than one that occasionally runs on goodwill, and I would rather live in the second one.

The mechanic still gets his blood pressure checked for free, and the world is better for it. But goodwill only stays goodwill when both sides know it is being extended. The moment one side starts treating it as a standing entitlement, quietly folded into the price of a haircut, it has stopped being generosity and become an unpaid shift that nobody agreed to work…!

And lastly the part that this genre never writes, because it is the part that is not flattering to us.

Let me describe those thirty minutes honestly.

An unregistered teleconsultation with a minor. No consent, hers or mine. Conducted on a personal phone, in a commercial premises, with a third party listening. No verification of identity. No history beyond what could be volunteered in a hurry by a frightened father. No examination worth the name. No written record. No prescription. No defined pathway if I turned out to be wrong.

India has had a framework for exactly this since March 2020. The Telemedicine Practice Guidelines were gazetted in May 2020 as Appendix 5 to the Indian Medical Council (Professional Conduct, Etiquette and Ethics) Regulations, 2002, and they remain the governing framework today under the National Medical Commission. They cover consent, identification, records, prescribing limits, and the circumstances in which a remote consultation is simply inappropriate. A registered practitioner is liable for negligence in telemedicine exactly as in physical practice.

None of that stops applying because the doctor happens to be wearing a cape…!

But regulation is the smaller problem. The larger one is that informality degrades the advice itself, and we have known this for over a decade.

In a prospective cohort published in the Journal of Hospital Medicine, Burden and colleagues compared curbside consultations with formal consultations performed on the same patients on the same day. The information conveyed in the curbside version was inaccurate or incomplete in roughly half the cases. The management advice differed between curbside and formal in about six out of ten patients. Where the information had been inaccurate or incomplete, the advice differed almost every single time.

Now hold that study up against my thirty minutes.

Share

So this is written for every healthcare professional who has taken that phone.

We take the call at the wedding. We read the reports at the airport. We answer the just one small doubt in the lift, at the gym, in the queue, at 11:45 PM, holding a plate of food. We tell ourselves it would be unkind to refuse and unsafe to decline. We tell ourselves we will follow up. We rarely do.

I repeat again “We are catastrophically bad at saying no…!” — not for lack of words, but because refusing feels as though it might cost someone their health, and we would sooner absorb the cost ourselves than risk being wrong about that.

The instinct is admirable in a genuine emergency. Everywhere else it is a liability, because the people asking cannot distinguish an emergency from a Tuesday — and we have never once bothered to teach them.

A febrile child does not wait for the blow-dry. A wrong dose does not care that you were being polite. A missed red flag does not become less missed because it was given away for free. And a corridor is not a clinic, however good the intentions of everyone standing in it……

None of this is an argument for billing your neighbours.

I will go back to that salon. He is decent at his job, and I am not about to relearn how someone else handles a blade near my scalp.

But the next time the phone rings mid-haircut, the answer is going to sound different. Finish the haircut. Then we sit down and do this properly…!

Not because I want to be paid for it.

Because thirty minutes squeezed between a trim and a blow-dry is not a consultation. It is a favour wearing a consultation’s clothes — and the person it deceives most dangerously is not the doctor. It is the patient on the other end of the screen, who has no way of knowing the difference.

The head massage was the small thing.

That is exactly why it mattered.

A system reveals itself in what it agrees to put a price on, and what it agrees not to notice. In that salon, a head massage was a service and a child’s assessment was a chat. Nobody decided that. Everybody accepted it.

We should read that before the patient does.

Sources referenced:

Darby, M. R. & Karni, E. (1973), Free Competition and the Optimal Amount of Fraud, Journal of Law and Economics 16(1):67–88 — origin of the “credence goods” category. https://www.journals.uchicago.edu/doi/abs/10.1086/466756

Burden M, Sarcone E, Keniston A, Statland B, Taub JA, Allyn RL, Reid MB, Cervantes L, Frank MG, Scaletta N, Fung P, Chadaga SR, Mastalerz K, Maller N, Mascolo M, Zoucha J, Campbell J, Maher MP, Stella SA, Albert RK. Prospective comparison of curbside versus formal consultations. J Hosp Med. 2013 Jan;8(1):31-5. doi: 10.1002/jhm.1983. Epub 2012 Oct 12. PMID: 23065716.

Board of Governors in supersession of the Medical Council of India (2020), Telemedicine Practice Guidelines, notified 25 March 2020, gazetted 12 May 2020 as Appendix 5 to the Indian Medical Council (Professional Conduct, Etiquette and Ethics) Regulations, 2002; currently administered by the National Medical Commission. https://pmc.ncbi.nlm.nih.gov/articles/PMC8106416/

It’s a beautiful picture. And today, on Guru Purnima, in the middle of a Gen Z reckoning with institutions, an AI that can answer almost anything in seconds, and an education system groaning under its own bureaucracy — it’s worth asking honestly what that picture still means, and what it doesn’t.

A senior clinician takes the stethoscope from a student's ears, places it back on her own for a moment, then returns it. "Don't chase the sound," she says. "First understand where it belongs." Nothing in this exchange appears in a textbook. It cannot be looked up, tested for, or downloaded. The student is being taught how to listen — to a murmur, and to a patient, and eventually to her own uncertainty. This is the guru–shishya bond at its most defensible. It is also, not incidentally, the version we choose to remember every Guru Purnima, on a day built almost entirely out of gratitude and very little else.

I want to hold two things at once here, because I think both are true. I revere this tradition — genuinely believe our civilisational character survives, in no small part, because of it. And I also think reverence without scrutiny is how good things curdle. So here’s an attempt at a mature, unsentimental, multi-angle look — not to settle the question, but to actually sit with it, the way the word upanishad itself suggests: sitting close, in inquiry.

Subscribe now

It was never mentorship. That’s the first honest thing to say.

A mentor helps you do something better — pass an exam, land a role, navigate a department. It’s a real and useful relationship, organised around career and competence.

A guru enters somewhere deeper: not just what you know, but what you’re becoming. A resident doesn’t just learn how to examine a frightened child from a textbook — she learns it by watching whether her senior sits down before speaking to the parents, whether uncertainty can be admitted without shame, whether credit is given fairly after a difficult case. None of that is on any syllabus. It shapes her anyway, maybe more than the syllabus does. That’s the part the corporate mentorship language simply cannot hold.

Share

The realist’s view: the tree got replaced by a logbook

Here’s the uncomfortable, unromantic truth about why this bond has thinned in most institutions — and it isn’t that today’s teachers have less character or today’s students less reverence. It’s that the conditions for sustained attention have quietly disappeared…!

Faculty are buried under clinical load, administrative reporting, publication targets, accreditation paperwork. Mentoring gets bolted on as one more unpaid, unprotected duty. Trainees, meanwhile, are surviving exams and deadlines, not searching for transformation — not because they don’t want it, but because nobody built them the time for it.

Institutions respond the only way institutions know how: they formalise it. Assigned mentors, logged meetings, signed forms. And here’s the trap — a mentoring system can be fully compliant on paper and completely empty in practice. You cannot schedule trust for the third Friday of the month. A relationship that shapes a human being doesn’t show up in a register just because three meetings were entered into one.

If the tradition is going to survive institutions, the realist’s demand is simple: protected time, real recognition of mentorship as serious labour, and — where possible — separating the person who nurtures you from the person who has unchecked power over your career. Care and evaluation sitting in the same hands is where things quietly go wrong…!

Subscribe now

The pessimist’s warning, which deserves to be heard in full

Reverence, when it’s the only ethical foundation of a relationship, is dangerous — because it doesn’t announce itself as power. It just feels like devotion, gratitude, loyalty. Which is exactly how it can also become the vehicle for someone senior denying a junior their due credit, humiliating a resident in front of a patient’s family, or punishing a question asked at the wrong moment.

And there’s an older, harder truth underneath: this tradition has a shadow. Historically (if the contemporary academic historians are to be beileved), access to it was not universal — class, gender, wealth, and lineage probably decided who was permitted to sit beneath that tree. We don’t honour a civilisational inheritance by pretending it never excluded anyone. We honour it by widening the circle it once kept narrow.

The pessimist isn’t attacking the bond. They’re refusing to let sacredness be used as cover for harm — and that refusal is, itself, an act of respect for what the relationship could be.

Subscribe now

The idealist’s answer to “but AI can teach that now”

Here’s where I think the idealist is simply, factually right, not just sentimentally right.

A textbook can list the features of a murmur. An LLM can generate ten empathetic ways to speak to a grieving family in under a second. Neither can carry the responsibility for that conversation. Neither can sit beside a resident at two in the morning after a death, let the silence stay exactly as long as it needs to, and help her walk back onto the ward without becoming numb or breaking.

This is tacit knowledge — judgment, timing, restraint, the sense that something is wrong before you can articulate why. It’s transmitted through apprenticeship: watching, then attempting with support, then finally acting alone. Cognitive science calls this modelling, scaffolding, fading. The Guru-Shishya tradition was doing this, quietly and precisely, long before anyone gave it that name…!

Subscribe now

The optimist’s twist: AI doesn’t kill the guru — it strips away their disguise

For a long time, a teacher’s authority came partly from simply knowing more. That’s over. A student can retrieve, compare, and synthesise faster than most seniors alive.

But that doesn’t make the guru obsolete — it makes them legible for the first time in decades. When information is free, judgment becomes the scarce resource. When answers are instant, knowing which question actually matters becomes the valuable skill. When a machine can produce polished, plausible language effortlessly, intellectual honesty becomes both harder to fake and more urgent to model.

The future guru doesn’t compete with the machine on memory. They become the person who helps you decide what’s worth remembering, questioning, resisting, and doing — which was arguably always the real job, just obscured by how much of the old role was information-delivery.

The relativist’s correction: the guru is becoming a constellation, not a sovereign..!

No single senior can now hold all the clinical judgment, research integrity, technical fluency, and emotional steadiness a young professional needs. A resident might get ethical grounding from a veteran in a rural hospital, computational thinking from a peer on another continent, and humility from a nurse who’s seen more suffering than any professor. Sometimes the student teaches the teacher — and a mature guru experiences that not as humiliation but as evidence that their authority never depended on omniscience in the first place. “Teach me” may be one of the highest things a senior can say to a junior.

This doesn’t make the bond shallower. It just means no one person should claim total ownership of another person’s becoming..!

What should actually survive — and what has to be let go (purely my opinion)

Subscribe now

Putting all five lenses in the same room, here’s the practical shape that emerges, not as theory but as something you could actually try to live:

Keep: sustained attention, apprenticeship, gratitude, the recognition that some things — judgment, courage, presence — can only be learned in close human proximity to someone who has them.

Release: unquestioned authority, coerced loyalty, the confusion of family language (”she’s like my daughter”) with an excuse for having no boundaries, and — quietly, but non-negotiably — the idea that the learner owes permanent dependence in exchange for what they were given.

Add, explicitly, because the old model never made this a formal value: the right to leave. A postgraduate should be able to change guides without professional retaliation. A junior should be able to decline an unethical demand without losing a recommendation. A relationship isn’t fully voluntary if leaving it threatens someone’s future. The best gurus have always understood this instinctively — they build their students toward independence from the very first day, and they measure their own success not by how devoted their students remain, but by how capable those students become of living without their approval.

For today

If you’re a shishya today, the honest questions aren’t just “who taught me.” They’re: Who corrected me without diminishing me? Who opened a door and didn’t demand lifelong loyalty for it? Who allowed me to become someone they hadn’t planned for?

And if you’re a guru — and most of us are, to someone, whether we’ve noticed or not — the harder questions run the other way: Did I give credit fairly? Did I make it safe for someone junior to disagree with me? Did I mistake obedience for learning? Did my student become more capable of living without my approval, or less?

The image beneath the tree is still beautiful. I don’t think its future lies in rebuilding the ashram, and I don’t think it lies in a corporate mentorship dashboard either. It lies somewhere quieter: in a guru who gives a student enough knowledge to stand, enough humility to keep learning, enough courage to disagree, and enough freedom to walk toward a place the guru will never see.

Maybe that’s the real gurudakshina — not repayment, not repetition, not permanent dependence, but carrying the knowledge forward with gratitude and judgment: repairing what was inherited, widening the circle that was once kept narrow, and becoming capable of handing the next generation something a little better than what we received.

Best wishes on Guru Purnima. To everyone who taught me to look again — thank you. And to everyone I’ll one day be responsible for — I hope I remember to let you go.

This is the prevention paradox, and it has a name and a paper, both usually skipped over in Indian medical training in favour of the diagrams. Geoffrey Rose, a cardiovascular epidemiologist at the London School of Hygiene, wrote “Sick individuals and sick populations” in 1985, and it remains one of those rare ideas in medicine that is counterintuitive and completely correct at the same time. Rose separated two questions that sound alike and are not: why did this particular person get the disease, and why does this population have this rate of disease. The determinants of individual cases and the determinants of incidence rate are different problems, and finding out why one child developed measles is not the same exercise as explaining why one district keeps having outbreaks and another does not.

Subscribe now

Prevention works, and then starts looking unnecessary

Rose’s own formulation was plain: “A preventive measure which brings much benefit to the population offers little to each participating individual.” He said it about immunisation, seat belts, and lifestyle change, and it still holds forty years on. Most people wearing a seat belt on any given drive were never going to be in the accident. Most children getting a vaccine were never going to catch that infection in that window. Multiply the small probability across millions of drives or millions of children and the number of deaths averted becomes real, even though no single beneficiary can ever be pointed to.

Communicable disease adds a layer static risk factors don’t have. The baseline itself moves with participation. The mother in my OPD could truthfully say nobody had seen measles, precisely because coverage had stayed high enough. If enough families make the same individually reasonable decision to opt out, that protection erodes and the individual risk climbs back up. The absence of disease was partly evidence the programme was working, not evidence it was no longer needed.

Share

Where the number needed to treat makes this uncomfortable rather than abstract

Number needed to treat, for anyone who has not sat through the biostatistics lecture recently, is simply the number of people who must receive an intervention, over a defined period, for one additional person to avoid the outcome being measured, compared with not treating them. It is the inverse of the absolute risk reduction. Nothing more mysterious than that, though it gets treated as either magic or an indictment depending on who is quoting it.

Take statins for primary prevention, since the numbers are worked out properly and I cannot be accused of picking something obscure. A persistence-adjusted analysis of Cochrane primary prevention data found five year NNTs of 146 in lower risk populations and 53 in higher risk populations to prevent one coronary event, the gap driven mostly by differing baseline risk. In JUPITER, the quoted five year NNT for the composite endpoint was 25, though that trial was stopped early at a median follow-up under two years and the number shifts with which endpoint you pick, so I would not lean on it too hard in a talk.

I used to read an NNT of 25 as twenty-four losers and one identifiable winner. That is the wrong picture, and I want to correct it plainly rather than bury it in a hedge. Nobody can point to the one person in twenty-five whose outcome changed. NNT is an average expression of a counterfactual that never gets observed in any single patient, not a lottery with a visible name on the ticket. Some of the other twenty-four may be getting benefit the trial’s chosen endpoint never measured. All twenty-five carry the pill burden and the cost regardless.

Subscribe now

Where I think the paradox gets misused

I have seen this argument turned in exactly the wrong direction, usually by someone selling something. “Most people who take this don’t benefit” is true and I am using it honestly above, and it also gets used to argue against vaccination or salt iodisation in favour of something individualised and expensive.

Rose’s own point cut the other way. Small shifts across a whole population produce more net benefit than large shifts confined to a small high-risk group. That is not always true and Rose never set population and high-risk strategies against each other as enemies, we still need high-risk clinical care, but where most of the disease burden sits in the broad middle of the distribution rather than the extreme tail, chasing only the tail misses most of it.

The paradox cuts both ways as a discipline, though. If a great many low-risk people must participate so that a few benefit, the intervention has to be genuinely safe, cheap, and light, and the harms owe the same honesty as the benefits. Population benefit does not get to make individual burden disappear from the account.

Share

What I have actually changed in practice

These carry the obvious bias of someone doing maternal-cord serosurveillance and thinking about AMR stewardship for a living, so treat them as one clinician’s rules rather than guidance:

• For counselling families, I stopped trying to inflate a small population probability into a felt personal benefit, because it cannot be done honestly. I say plainly that the present risk is low and that it is low partly because coverage has held.

• For teaching purposes, I now separate relative risk reduction, absolute risk reduction, and NNT explicitly in the T-L sessions, since learners/office practitioners used to high-risk clinical work often read a modest relative effect as unimpressive without seeing the population arithmetic behind it.

• Before agreeing to scale anything down to “high-risk only,” I ask where the actual case burden sits, because Rose’s argument depended on the exposure being reasonably continuous across the population, and that is not automatically true for everything I work with.

  

I had a child born to a naturopthy practising couple last year whose mother had declined the hepatitis B birth dose because her knowledge told her it was unnecessary for a baby born at home to an apparently healthy mother. Nothing happened to that child. Nothing was ever statistically likely to happen to that child, and a mother looking healthy is no guarantee she is not carrying the virus, and infection acquired in infancy has a much higher chance of turning chronic than infection acquired later. Nothing happening was not proof the policy was wrong. It was the paradox doing exactly what Rose described, one family at a time, in a programme that only makes sense when you stop looking at one family.

I am not convinced I explain any of this well in the room, in the two minutes I actually have with a parent before the next patient. I would like to know how others here are doing it differently…!

Over the past year, in my capacity reviewing research and data-sharing proposals for medical colleges, I have watched a particular kind of document arrive with increasing frequency. A health-tech company — well-funded, well-spoken, slides in the right shade of blue — proposes a "collaborative research partnership." The institution will "digitise and structure" its patient records. In exchange, the company will use the de-identified data to improve a diagnostic model. There is usually a data-sharing agreement. There is usually a paragraph about future co-authorship. There is rarely a paragraph about who owns the finished model, who profits from it, or whether the originating institution will ever see it again except as a customer…!

I have watched capable, well-meaning colleagues prepare to sign such documents. Not out of carelessness. Out of the ordinary, unglamorous scarcity that defines under-resourced Indian medical colleges: little in-house AI or data-governance expertise, ethics committees stretched thin across a hundred other responsibilities, and a genuine, understandable hunger for anything that looks like partnership with people who have resources the institution does not. This is not naivety in the way the word usually implies. It is a rational decision made inside a system that has never given these institutions the leverage to negotiate better terms.

The more accurate word for what is happening across underfunded Indian institutions right now may not be exploitation of ignorance. It may be extraction, dressed as collaboration, aimed precisely at the places least equipped to price what they are giving away.

Subscribe now

Two Indias, two trajectories

I have argued elsewhere that AI is unlikely to make Indian healthcare cheaper overall. I want to go further here, because "overall" hides the more important story, which is divergence. India does not have one healthcare system into which a new technology arrives. It has several, stacked on top of each other and barely on speaking terms:

• five-star corporate chains with the capital to license the best diagnostic tools and the fee-for-service incentive to deploy them liberally;

• mid-tier private nursing homes chasing whatever technology promises differentiation; and

• government and rural teaching hospitals running on numbers that would be a rounding error in the first category's annual report.

The corporate tier will likely use AI to add cost — more flagged findings, more confirmatory scans, more billable next steps, absorbed easily by a payment structure built to convert efficiency gains into revenue. That is a real problem, but one experienced mostly by patients with insurance or savings to absorb it.

The trajectory that concerns me more runs backwards through the second tier: not AI arriving to add cost to care already delivered, but AI arriving to extract value from the one asset resource-poor institutions hold in surplus — patient volume, diagnostic diversity, and the density of undocumented clinical experience that a data-hungry model needs, and that a smaller, more curated corporate caseload often cannot supply as cheaply.

Subscribe now

The extraction economy has a familiar shape

Indian medicine has lived this pattern before, in a different currency. Government and rural hospitals have long been where clinical trials recruit fastest, where postgraduate theses find their sample sizes, where research eventually cited worldwide gets its raw material — often with informed consent that is technically valid and substantively thin, given by patients navigating illness, poverty, and a language barrier from the consent form itself.

The incentive structures now forming around AI training data appear to be reproducing that shape, faster, and with even less institutional memory of what fair benefit-sharing should look like. A model trained substantially on the pathology load of underfunded hospitals — tuberculosis, rheumatic heart disease, malnutrition-complicated illness, the diagnostic long-tail a well-resourced urban hospital sees less of — becomes commercially valuable precisely because of that data…!

The finished, licensed product is then typically priced for institutions that can pay: the corporate chains, again. The originating institution is, at best, offered a discounted subscription to a refined version of its own patients' data. At worst, it is offered nothing beyond the original proposal's vague promise of "future collaboration."

This is not a hypothetical harm sitting downstream. Ethics committees at under-resourced colleges across the country are, right now, evaluating proposals of exactly this kind — framed as capacity-building, assessed by reviewers without the specific expertise to price data as an asset, approved by people acting in good faith under real institutional pressure. This is not a failure of individual judgment. It is what happens when an entire tier of institutions is asked to negotiate data value without ever having been given the tools, precedent, or leverage to do so…

Share

Even the honest case for AI carries this shadow

I want to be fair to the genuine counter-example, because I raised it before and it deserves to stand: an AI triage tool that lets a technician in a primary health centre flag a probable TB pattern on a chest film, before it ever reaches a radiologist three hundred kilometres away, is not a cost driver — it may be the only realistic way that film gets read at all. I still believe that. But I no longer think it is separable from the extraction question.

Ask where that triage model’s training data came from. Ask whether the institutions that supplied the chest films that taught it to recognise TB were compensated, credited, or consulted about how the finished tool would be priced and to whom it would be sold. Often, the honest answer is that they were not asked, because nobody thought to ask, because the data-sharing agreement that made the model possible was signed the way the one on my desk nearly was — quickly, hopefully, by people doing their genuine best inside a system that has never taught them to say no on commercially informed terms.

Subscribe now

What the divergence actually looks like, five years out

If this trajectory holds, the disparity AI produces in Indian healthcare will not simply be the familiar one — corporate hospitals get better tools, government hospitals get worse ones.

It will be stranger and more self-defeating than that: government and rural institutions will have supplied the raw material that made the tools possible, received little of the financial return, and then be expected — by policy, by accreditation bodies, by the general cultural pressure to look “AI-enabled” — to adopt the finished products anyway, at prices set by a market their own data helped create.

The return on investment for the institutions doing the actual data-generating work will be poor almost by design, because the value was priced and captured somewhere else in the chain, by people who never treated a single one of the patients whose records built the model…!

The proposals keep arriving — reasonable-sounding, well-intentioned on both sides, evaluated by good people without the specific expertise or leverage the moment requires.

Institution by institution, we are generating the data, signing away the data, and quietly reinforcing a system that will use our own openness against the very patients and institutions that made it possible.

Nobody in any of these rooms intends harm. That is precisely what allows the pattern to survive. We are building, with good faith and very little malice, the exact asymmetry we will spend the next decade trying to explain to ourselves.

Note: examples in this piece are composite and illustrative, drawn from patterns visible across multiple institutions rather than any single proposal or review.

But the opposite story — AI has made associations obsolete — is just as lazy. It mistakes information for judgment, and access for trust. The real question isn’t whether associations should exist. It’s whether they can prove they’re worth what they cost. Right now, almost none of them can.

There was a time when joining a medical association did not need much explanation. You joined because your seniors were members. You attended conferences because that was where knowledge moved. You listened to lectures because information was scarce. You paid the annual fee because belonging to the professional body was part of becoming a doctor in the full social sense of the word. The association was not just a body. It was a guild, a fraternity, a network, a classroom, a political instrument, and sometimes even a shelter.

That world has changed.

Today, knowledge is no longer locked inside conference halls. A postgraduate student can hear a world expert on YouTube before breakfast. A private practitioner can ask an AI tool to summarize a new guideline before evening OPD. A young pediatrician can compare global protocols, download open-access papers, join case discussions, listen to podcasts, and receive more updates in one week than an earlier generation received in months. The problem today is not lack of information. The problem is too much information, too little judgment, too little time, and too little trusted local meaning.

This is where the uncomfortable question begins.

Subscribe now

What exactly is the ROI of being a member of a medical association today?

Debates about institutional relevance tend to collapse into two camps: defenders who point to legacy and scale, and reformers who point to stagnation and waste. Both camps are half right, which is precisely why the debate never resolves.

What almost nobody in either camp does is the boring, unglamorous work of finding out which specific activities are actually earning their budget line. That's not a defect of one side's argument. It's a defect of the entire sector's accounting.

The easy answer is to reduce ROI to personal benefit.

• Did I get CME points?

• Did I get a certificate?

• Did I get a registration discount?

• Did I meet useful people?

• Did I get a chance to speak?

• Did I get visibility?

• Did I recover the value of my membership fee?

These are not irrelevant questions. Members are not saints. They spend money, time, travel, attention, and emotional energy. They deserve value. But if ROI is reduced only to what the individual member receives, then a medical association becomes just another service provider. At that point, it is competing with AI tools, online courses, YouTube lectures, WhatsApp groups, private CME platforms, pharmaceutical events, and professional networking apps.

I feel, that is a losing game…!!

A medical association cannot justify itself merely by saying, “We provide information.” The internet already does that. AI now does it faster. The association has to provide something more difficult: judgment, trust, context, solidarity, local relevance, ethical courage, and collective strength.

I feel, that is the real ROI….!

Subscribe now

Activity Is Not Impact, and Vision Is Not Evidence

Here’s the trap: an association under pressure to look modern doesn’t necessarily fix its measurement problem — it just rebrands the same unmeasured activity as “innovation.” A five-year AI strategy document with no KPI (Key Performance Indicator) attached is exactly as much of a ritual as a webinar series nobody evaluates….! Both are institutional theater. Neither has earned the right to call itself accountable.

If that sounds harsh, consider the test: could the organization tell you, with a straight face and real numbers, which three programs from last year measurably changed how a clinician practiced? Not attendance. Not satisfaction scores. Practice.

Most associations — Across all the medical domains; take your pick — cannot answer that question…!

The old model of associations was built in an information-scarce world.

• Journals, conferences, CMEs, annual meetings, lectures, and guidelines had obvious value because access itself was valuable.

• A senior person had read what others had not.

• A conference brought information that had not yet reached the periphery.

• A journal carried authority because few alternatives existed.

The association was a gate through which knowledge passed…!

But we now live in a post-scarcity knowledge world.

• More lectures do not automatically mean more learning.

• More webinars do not automatically mean better practice.

• More guidelines do not automatically mean better patients.

• More conferences do not automatically mean stronger professional life.

In fact, more can sometimes mean worse….!

Doctors are already overloaded. They are drowning in forwarded messages, half-read guidelines, medico-legal warnings, drug updates, institutional pressure, parental anxiety, algorithmic medicine, public mistrust, and endless administrative demands. In such a world, an association that merely adds more noise is not helping. The associations are worsening the cognitive burden of its own members….!

Share

The useful association of the present age must become a filter, not an amplifier

It must tell members what matters, what does not matter, what is evidence, what is fashion, what is locally applicable, what is dangerous, and what can be ignored. It must help cut the clutter…!

That is knowledge ROI.

Instead of asking “should this program exist,” the sharper question is “what tier of evidence has this program cleared?”

I can picture five rungs:

1. Did people show up? (attendance, downloads, credits — easy, proves nothing)

2. Did people come back? (retention, NPS — a pulse, not a proof)

3. Did anyone’s practice change? (pre/post testing, audit data — the rung almost everyone skips)

4. Did the system change? (policy adoption, protocol citations, new consortia)

5. Did outcomes change? (proxied patient/societal impact, tracked honestly, attribution caveats and all)

Nearly every institution can show rungs one and two on demand, in a glossy annual report. Almost none attempt rung three. That gap between what’s reported and what’s proven is where all the real skepticism should be aimed — at the traditionalist’s webinar and the reformer’s AI pilot in exactly equal measure.

This is not an argument against conferences. Conferences have value. They create memory, friendship, energy, professional identity, and informal learning. They allow younger doctors to be seen. They allow older doctors to pass on experience. They build networks that cannot be fully replaced by screens.

But conferences cannot remain the central proof of relevance….!

If an association becomes active only around annual events, it is not functioning as a living professional body. It is functioning as an event machine with a medical logo.

The real test is what happens between conferences….

Subscribe now

What Actually Counts as Reinvention (few association do it)

Not a new logo. Not an AI taskforce that produces a white paper and no dashboard. Reinvention that would actually mean something looks like:

• A public dashboard, updated yearly, showing how many flagship programs cleared rung three or above — not a curated highlight reel, the full unflattering list.

• Sunset clauses, written down in advance, retiring any recurring program that can’t produce rung-three evidence within a fixed window.

• A ring-fenced evaluation budget, small but real, spent on independent assessment rather than more programming.

• Leadership terms tied to evidence, not attendance — renewal contingent on what got measured, not what got scheduled.

• Separate scorecards for reach and for outcome, so equity arguments and impact arguments have to each earn their keep.

None of this requires abandoning tradition or chasing every AI trend. It requires the one thing institutions of every kind resist most: keeping score in public, including on themselves.

These reveal ROI more honestly than attendance numbers…!

This is especially important in India. Our medical reality is not clean. We work across government hospitals, private clinics, corporate institutions, teaching hospitals, nursing homes, rural centres, semi-urban setups, and overstretched public systems. The gap between textbook medicine and daily practice is often large. Families arrive with internet knowledge and deep mistrust. Media trials simplify clinical complexity. Courts, administrators, hospitals, insurance companies, pharmaceutical interests, and public expectations pull doctors in different directions.

Share

A member should therefore ask: what can this association do that I cannot do alone?

• Alone, I can read a paper.

• Alone, I can attend a webinar.

• Alone, I can build my CV.

• Alone, I can post opinions online.

But alone,

• I cannot easily influence policy.

• I cannot create district-level disease surveillance.

• I cannot negotiate with institutions.

• I cannot build a referral ecosystem.

• I cannot produce consensus statements with legitimacy.

• I cannot defend professional standards during a public crisis.

• I cannot create a credible public health voice.

That is where the association becomes valuable….!

But this is also where members must be honest. Many members want the association to be strong while they remain passive.

• They want advocacy but do not contribute data.

• They want better leadership but do not vote seriously.

• They want meaningful academic work but do not review, write, audit, mentor, or organize.

• They want protection during crisis but do not help build the institution in ordinary times.

• They want reform but do not participate.

This is the consumer mindset….!

It weakens associations from within.

A medical association is not a subscription package. It is not a professional version of a streaming service where one pays and then waits for content. It is a professional commons. Its value depends on what members put into it, what they demand from it, and what they allow it to become.

If thoughtful members withdraw, associations do not remain neutral. They get captured by the loud, the ceremonial, the sponsor-friendly, the politically skilled, and the permanently ambitious. Then the institution slowly fills with posts, banners, plaques, garlands, predictable speeches, election arithmetic, and photo documentation. Serious members become cynical. Young members become bored. The community sees only symbolism. The association loses moral authority.

So ROI is not only a question members should ask of the association.

It is also a question members should ask of themselves.

Subscribe now

The Question Worth Sitting With

In pediatrics especially, the ultimate stakeholder is not the office-bearer, the speaker, the sponsor, or even the member. It is the child and family.

If a pediatric association spends large amounts of money, time, and professional energy but cannot show how its work eventually improves child health, then members have the right to ask hard questions. Not because they are anti-association. But because they are serious about the purpose of association.

Members must stop asking only, “What did I get for my fee?”

They must ask better questions.

• Did this association reduce my confusion?

• Did it improve my practice?

• Did it protect my professional dignity?

• Did it help me serve patients better?

• Did it create value I could not create alone?

• Did it use money and influence transparently?

• Did it build younger doctors?

• Did it speak when silence was easier?

• Did it improve the health system around me?

• Did it matter to the child and family?

That is ROI.

Not “are associations still relevant” — that’s a debate that can run forever because it’s unfalsifiable by design. The sharper question, and the one that actually has an answer:

What did this organization prove it changed last year, and would it publish that number even if the answer embarrassed it?

Everything else — the webinars, the five-year visions, the AI strategies, the term limits — is just detail until that question gets answered honestly.

As a pediatric professor, a lifelong student of medical education, and a doctor who still tries to practice ethically in an increasingly transactional world, I find myself pausing this year. Not in gratitude. In quiet, uncomfortable reflection.

What are we actually celebrating? The ideals of a profession that once stood apart from the marketplace — or the rituals that persist, hollowed out, while the profession itself has been industrialised around us?

A Different Kind of Trust

There was a time — romanticised now, perhaps, but rooted in real memory — when medicine in India carried a different weight. Doctors weren't professionals in the modern sense; they were part of the social fabric, often the first and last resort a family had. The relationship rested on trust, not timestamps, insurance pre-authorisations, or star ratings. "Vaidyo Narayana Hari" wasn't marketing. It was a society naming, out loud, the weight of holding someone's life in your hands.

Subscribe

Fast forward to 2026. The profession has been thoroughly industrialised. Corporate hospital chains run like efficient factories — patient throughput, revenue targets per bed, upselling of diagnostics and packages. Insurance companies and hospital administrators shape protocols as much as clinical guidelines do. The Consumer Protection Act, meant to empower patients, has also layered on defensive medicine: more tests, more documentation, more fear of litigation than of missing a diagnosis. Capitalism didn't just enter healthcare. It rewired its incentives from the inside.

Share

I see the inequality daily. In teaching hospitals, we work with constrained resources, overstretched residents, and patients who've travelled hundreds of kilometres for basic care. A few kilometres away, private centres gleam with equipment gated by the ability to pay. The data bears this out — stark urban-rural and rich-poor divides in outcomes, year after year. And when colleagues who've trained abroad describe how an MI or a septic child is met in many high-income systems — a coordinated, resource-backed response rather than a scramble for the next ICU bed — the contrast isn't imagined. It's a difference in what the system does *automatically*, without depending on any one person's heroics. That said, it's not a clean comparison: those systems have their own well-documented strains — corridor care, ambulance delays, staffing shortfalls — and I don't want to trade one myth (the selfless Indian doctor) for another (the seamless Western system). What I mean, specifically, is this: in many HIC settings, the *infrastructure* of an emergency response — bed allocation, oxygen supply, triage protocol — doesn't depend on which hospital you happened to reach first. That specific kind of reliability is what I see missing here, more than any general claim of superiority.

Subscribe

India's system delivers heroic individual effort against systemic odds. But the aggregate picture is one of patchy excellence inside widespread strain. We take pride in medical tourism and the size of our doctor pool, while quietly losing people to burnout and brain drain, and losing public trust one bad headline at a time.

Saviours in Suits

Picture it: a doctor moves between OPD, ward rounds, and an emergency, and is greeted at day's end with a corporate hamper and a LinkedIn post thanking "our frontline warriors." The same system, the same week, pushes them to hit targets, sit through "relationship-building" dinners with pharma reps, and justify every clinical call to a utilisation review team. Saviours? More like gladiators in an arena where the crowd cheers selectively and the organisers count the ticket sales.

The consumer lens deepens the satire. Patients — rightly more empowered than a generation ago — sometimes approach a consultation like a delivery app order: *this scan, this injection, quick relief, or a bad rating.* Defensive practice blooms in response: caesareans that weren't strictly needed, polypharmacy, referrals made to cover a file rather than treat a patient. Not out of malice. Out of survival, in a system that has made survival and integrity harder to hold at the same time. We became "service providers." The divine aura faded as the balance sheet rose — not because capitalists set out to villainise doctors, but because they commodified the encounter, and the encounter is what the aura was ever about.

Subscribe

Here's the raw nerve, since we're naming them today: not every doctor is a saint, and the profession carries its own rot — quackery, unnecessary procedures, people who game the system for profit. Pretending otherwise does no one any favours. But painting the whole fraternity as greedy, while ignoring the structural forces that reward exactly that behaviour, is convenient scapegoating. It's easier to run an outrage cycle about one clinician than to examine underfunding, regulatory capture by large players, or a decade of policy inertia. The Consumer Protection Act brought real accountability — and also buried clinicians in paperwork that has nothing to do with the patient in front of them.

As an educator, I watch this land on the next generation first. Bright, idealistic residents, crushed by unrealistic hours, debt, and the gap between what they were taught — patient first — and what they live — throughput first. Lifelong learning starts to feel like a luxury when survival already demands constant upskilling in medicine *and* in navigating a business you didn't choose to be part of.

Share

What This Isn't

This isn't a blanket case against private enterprise. Corporate hospitals have built specialties, drawn in talent, and raised the technical ceiling of Indian medicine faster than the public system alone ever could have. Plenty of ethical doctors do good work inside that system by holding their own line. And public healthcare is no utopia either — it has its own corruption, inefficiency, and political interference.

Nor is the comparison with HICs a simple one. India's scale, its epidemiological transition from infectious to lifestyle disease, its federal complexity — none of it maps cleanly onto smaller, wealthier, more centralised systems. Ayushman Bharat and similar efforts show real promise and real implementation gaps in the same breath. India's strengths show up in different places — disaster response, epidemic containment — even as its everyday reliability lags.

The cost of all this is not abstract. Eroding trust harms everyone: patients delay care because they fear the bill or the misdiagnosis; doctors wear burnout like a badge because the alternative is admitting the system is unsustainable. Medical education risks turning out graduates who are excellent at exams and starved of mentorship. The "saviour" narrative feels good for a day, but it quietly excuses the system from the work of actual reform — better-balanced regulation, fairer reimbursement, real investment in primary care and prevention, less administrative bloat, and a recalibration of what patients, providers, and policymakers each owe one another.

Subscribe

Maybe, too, we've over-medicalised the whole conversation, while under-investing in the things that actually determine health at scale — nutrition, sanitation, education, equity.

Doctors cannot fix what policy and economics have broken. And yet we're held to a god's standard by a system that treats us as replaceable line items.

A Harder Mantra

I'm not writing this for sympathy, or for more garlands. I'm asking, as a practitioner and a teacher: what does it mean to heal inside an industrialised system? How do I teach the next generation integrity, in an environment that keeps rewarding volume instead? How do I advocate for a patient when the system I work in sometimes sets us against each other?

Share

The profession isn't dying. It's been transformed, and pretending otherwise doesn't help anyone. The truth underneath the satire is this: we were never meant to be saviours standing alone, and we were never meant to be vendors either. Real healing has always required a partnership — between doctor and patient, profession and policymaker, ethics and economics.

Perhaps the mantra we actually need this year isn't *Vaidyo Narayana Hari*, but something closer to *Vaidya, Patient, Society — Sahakarinam*: collaborators, not saviours and customers.

To my colleagues: keep learning, keep questioning, keep resisting the burnout that gets sold to you as a badge of honour.

To patients: hold us accountable, but extend grace where the system fails an individual doctor who is fighting it too.

To the regulators and the boardrooms: healthcare is not just another industry. Commodify it past a certain point, and you erode the very trust the whole enterprise depends on.

The downfall isn't inevitable. But reflection without reform is just another performative post. On this Doctors' Day, let's aim for something harder than gratitude — an honest reckoning…!

What a Discharge Summary Quietly Reveals

The post travelled. People shared it. Doctors wrote back. Parents wrote back. Hospital administrators said it was eye-opening. Residents said they had forwarded it to their seniors. It went across cities, across specialties, across countries. Someone told me it was discussed in a quality meeting. Someone else said it was pinned in a department WhatsApp group…!!

Since then, I came across many kids who came back from tertiary care centres with a discharge summary carrying various inappropriate Final diagnoses…...

The diagnoses columns read as usual:

1. Global developmental delay

2. Fever under evaluation

3. Acute malnutrition

4. Failure to thrive

5. Seizure under evaluation etc etc etc…!

And so I am writing again. Not because I think this post will do what the last one could not.

Share

The Age of AI Has Arrived Everywhere Except the Diagnosis Column …!!

Let us be precise about where we stand in 2026.

Hospitals in India are using AI to detect diabetic retinopathy on fundus images. AI is reading ECGs. AI is triaging radiology worklists. AI is predicting sepsis scores. AI is transcribing OPD conversations. AI is generating billing summaries, analysing insurance claims, flagging drug interactions, and personalising patient reminders.

Hospitals are also spending crores on AI-powered CCTV systems to monitor footfall, on chatbots to answer patient queries, on dashboards that track bed occupancy in real time, on tools that predict when a ventilator will next need servicing.

We have AI for revenue. We have AI for operations. We have AI for marketing.

We apparently do not have AI for the one document that travels with the child after she leaves the building.

This is not a gap in technology.

This is a gap in what we have decided matters.

Share

ICD-10 Has Existed for Thirty Years. ICD-11 Is Already Here.

The International Classification of Diseases, Tenth Revision, was released in 1994. Thirty years ago. It has codes for almost every condition in clinical medicine. It has specific codes that distinguish between febrile seizures and epilepsy, between severe acute malnutrition with complications and without, between developmental delay with a known cause and developmental delay that is unspecified. It has the granularity to make vague language unnecessary.

ICD-11 is already in use. It is more precise. It is digitally native. It is designed to work with modern hospital information systems.

Every hospital in India that files insurance claims uses ICD-10 codes. The billing department knows these codes. The medical records department knows these codes. The software knows these codes. The insurer checks these codes.

But the treating unit writes “fever under evaluation” in the diagnosis column and walks away…!!

This is not ignorance. A hospital that uses ICD-10 for billing cannot claim it does not know ICD-10 exists. It knows. It uses it selectively. It uses it where money is at stake.

It does not use it where only clarity is at stake.

And clarity, apparently, is not worth the same as money.

Subscribe now

Why Are Some Medical Colleges Still Producing This?

Medical colleges are where diagnostic habits are born and where they calcify.

Every junior doctor who writes “fever under evaluation” as a discharge diagnosis learned to do so in a teaching hospital. They saw seniors do it. They saw consultants sign off on it. They saw the discharge summary go out with that language and nobody complained. So they did the same. And they will teach the same to the next batch.

This is how institutional habits survive. Not through active instruction in bad practice. Through silence in the presence of bad practice.

A postgraduate resident writes a discharge summary at 11 pm after 36 hours of work. She copies the format from last month’s summary. The diagnosis column has the same words. The consultant countersigns it in 90 seconds during a break between two procedure calls. It goes through. The family is given the file. The next patient is admitted to the same bed.

Nobody failed in their intention. Everyone failed in their accountability…!

Medical colleges train residents in procedures. They train them in examinations. They train them in presentations. They do not train them, with any seriousness, in clinical writing. The discharge summary is treated as paperwork. It is assessed, if at all, for completeness of physical data fields. Not for clinical coherence. Not for diagnostic honesty. Not for the quality of reasoning it transfers to the next provider.

If a postgraduate student submits a thesis with a vague hypothesis and an unsupported conclusion, it is rejected.

If the same student submits a discharge summary with a vague diagnosis and no reasoning, it is stamped and filed.

We have decided that writing for examiners requires rigour. Writing for patients does not.

Subscribe now

Why Are Some Corporate Hospitals Doing This?

This question is harder to excuse.

Corporate hospitals have resources. They have EMR systems. They have quality teams. They have NABH accreditation processes that run for months. They have committees for everything. They have compliance officers.

They also have brand decks that say things like “patient-centred care” and “clinical excellence” and “outcomes that matter.”

And yet the discharge summary comes out saying “seizure under evaluation.”

What did the unit think was causing the seizures? Was it structural? Metabolic? Genetic? Idiopathic? … Where is the ILAE…!!?

None of this is in the diagnosis column. Sometimes none of it is anywhere in the summary.

But the bill is itemised to the last capsule. The signature of the department head is on the prescription. The discharge package has a welcome kit and a branded folder and a feedback form.

The thinking of the clinical team during the admission is random.

This is not a documentation problem. This is a trust problem. The family has trusted this institution with their child and their money. They are leaving with a folder that does not tell the next doctor what this institution actually concluded.

Share

The EMR Is Not the Villain. The Decision Not to Use It Properly Is.

Every time I raise this, someone tells me about the limitations of electronic medical record systems. Forms are rigid. Coding takes time. The system crashes. The interface is built for billing, not for clinical documentation.

These are real complaints. They are not excuses.

A hospital that wanted to fix this would build a discharge summary template that separates: final diagnosis, working diagnosis, comorbidities, unresolved problems, pending investigations, red flag symptoms, follow-up date, and family counselling notes. This takes one conversation with the software vendor and one policy decision from the medical director.

A hospital that wanted to fix this would configure ICD-10 or ICD-11 mandatory fields at the point of discharge. The doctor cannot close the encounter without selecting a coded diagnosis. This is already done in insurance-linked discharges. It can be done routinely.

A hospital that wanted to fix this would use any afforadble frontier AI Models — not to replace the thinking, but to audit the output. A simple language model can be prompted to flag if the discharge diagnosis is a symptom rather than a diagnosis, if unresolved problems have no follow-up plan, if the admitting complaint has no explanation in the summary. This is not complicated. This is a prompt and a checklist.

Hospitals that wanted to fix this could fix it in a single day…..!

They have not fixed it because they have not wanted to fix it.

Subscribe now

Who Bears the Cost?

Not the hospital…!

The hospital has already been paid. The bed is already occupied by someone else. The summary is in the file. NABH will check that a discharge summary exists. It will not check whether the diagnosis is clinically coherent.

The cost is borne by the child, who may return to a peripheral centre with the same unresolved problem and no documentation of what was already done.

The cost is borne by the family, who paid for an admission and received a description of their anxiety rather than a clinical conclusion.

The cost is borne by the next doctor, who must reconstruct the diagnostic reasoning from scratch, often without the investigations that were done, often without knowing what diagnoses were considered and excluded.

The cost is borne by the healthcare system, which repeats evaluations, reschedules tests, and readmits children because the clinical thinking of one admission never successfully transferred to the next point of care.

This is an expensive failure. It is invisible on no spreadsheet except the patient’s own.

Subscribe now

What Would Change This?

I have thought about this more carefully after the first post.

The first blog embarrassed nobody with actual authority. The families who read it had no leverage. The residents who forwarded it could not change the culture that produced the problem. The administrators who called it insightful continued to run systems that produced the same outcome.

For this to change, someone with power must decide it matters.

That could be a National Medical Commission that mandates discharge summary quality standards as part of Undergraduate/postgraduate training assessment. That could be NABH revising its documentation audit criteria to include diagnostic coherence rather than just field completion. That could be insurance companies refusing to process claims where the principal diagnosis is a symptom code without a qualifying clinical narrative. That could be hospital medical directors adding discharge summary quality to their monthly unit audit rather than only to their annual NABH preparation sprint.

Any one of these would begin to move the needle.

All of them together would transform it.

None of them requires new technology. None of them requires new money. All of them require a decision that the quality of clinical thinking, as expressed in written documents, is a legitimate performance metric and not merely the aesthetic preference of doctors who write blogs.

Subscribe now

A Specific Request

This may be a feasible audit to begin….for medical teams reading this. Read this

Audit as a Quality Improvement Tool

Open any ten random discharge summaries from your paediatric ward. From the last month. Look only at the diagnosis column.

Ask these questions:

• Is the principal diagnosis a confirmed disease, a working diagnosis, or a symptom?

• If the admission complaint was fever, does the diagnosis tell us what the fever most likely represented?

• If the diagnosis is “developmental delay,” does the summary tell us what workup was done, what was found, what remains pending, and what intervention was started?

• If the diagnosis includes “under evaluation,” does the summary tell us what the evaluation revealed so far and what happens next?

• If the answer to most of these is no, you have a systemic problem. It is not a junior doctor problem. It is not an overwork problem. It is a standards problem. And it has a solution.

• If you find that ten out of ten summaries are clinically coherent, honest, and useful to the next doctor — please write to me. I will publish your institution’s name and what you did. Not as a promotional act. As evidence that this is possible.

Share

The Dustbin Is Still There

Children are still going home with files full of unfinished sentences.

The child with seizures leaves without her family knowing if the hospital thought this was epilepsy or something else.

The child with fever leaves without the next doctor knowing what dangerous diagnoses were excluded.

The child with developmental delay leaves without a single sentence about what should happen next.

The hospital has been paid. The beds have been filled.

The AI has been deployed for the CCTV and the billing and the chatbot.

The diagnosis column is still a dustbin…!

I am not going to pretend that writing this twice changes what writing it once could not. I am not going to end with a call for optimism that I do not currently feel.

But the institutions must know this:

The families are not powerless forever. Literacy is rising. Consumer courts exist. Social media exists. The gap between what hospitals claim in their branding and what they deliver in their documentation is a gap that will eventually be filled — either by the institution choosing to close it, or by someone outside the institution forcing the question.

The diagnosis column is a place where the hospital’s thinking becomes a matter of record.

Right now, in too many places, it is a record of the absence of thinking.

That should embarrass everyone who signed the summary.

It should embarrass everyone who stamped it.

It should embarrass everyone who built the system that allowed it to leave the building.

The child has gone home.

The question is still open.

Someone should answer it.

This is Part Two of a series on discharge documentation in paediatric practice. Part One can be read here.

This week, Nature Medicine published a 25-page study evaluating how well large frontier AI models perform on clinical reasoning tasks. The models tested included GPT-5, Gemini 2.5 Pro, Claude 3.5, and several others. The paper was submitted seven months ago, in November 2025. By the time it landed in your inbox or journal feed, at least two of those named models had already been updated or replaced by newer versions with substantially different capabilities.

The numbers in that paper’s comparison tables were aging while the paper was in peer review.

This is not a failure of that research team. It is not an editorial lapse. It is a structural collision between two timelines that are simply incompatible: AI systems now iterate every few months, while peer-reviewed journals operate on cycles of six to eighteen months. The result is a growing body of literature that is simultaneously scientifically honest and practically stale — sometimes before the ink dries.

For medical educators teaching the next generation of clinicians, and for healthcare professionals trying to make evidence-informed decisions about AI tools, this creates a genuine problem that requires a new kind of reading literacy. You need to know not just what a paper says, but whether what it says will still matter in six months.

This article gives you the framework to make that judgment.

Subscribe now

Part One: Understanding the Problem

What “Redundancy” Means — and Doesn’t Mean

When we say an AI research study is “redundant,” we do not mean it was poorly conducted or should not have been written. We mean that by the time it reaches publication, its central claim; usually a performance comparison between specific AI models, describes a world that no longer exists.

Think of it this way. A study published in early 2026 that reports GPT-5 achieved 84% accuracy on a medical image question set is, in a narrow sense, true. That model, on that dataset, at that moment, did achieve that score. But by the time a clinician reads the published version, a successor model may already be in clinical evaluation. The 84% figure is a historical record, not a benchmark for current decision-making…!

The problem compounds when that paper appears in a prestigious journal. High-impact publication confers an authority that the finding may not have earned over time. A medical educator who incorporates it into curriculum, or a hospital administrator who cites it in a procurement decision, may be relying on a snapshot that has already been superseded.

Share

Why Medicine Has This Problem Worse Than Other Fields

Other disciplines studying AI — computer science, engineering, economics — also grapple with rapid model iteration. But medicine has atleast three additional layers that make the obsolescence problem acutely serious.

The clinical validation gap. Before an AI system can be responsibly used in patient care, it must be validated across real clinical environments: diverse patient populations, varying equipment quality, real-world noise, and integration into clinical workflows. This takes time — often years. The model a team begins validating is frequently not the model they finish validating. By publication, it may not even be the model currently deployed.

The benchmark-to-bedside gap. A study showing an AI model achieves 90% accuracy on a curated medical question set tells us almost nothing about whether that model reduces misdiagnoses, shortens time-to-treatment, or prevents adverse events in practice. The questions medicine most needs answered — does this tool change patient outcomes? — require study designs that are inherently slow, expensive, and difficult to replicate as models evolve underneath them.

The proprietary opacity problem. Many of the most powerful clinical AI systems are developed inside large technology companies and released commercially before any independent peer-reviewed evaluation exists. When academic researchers do publish independent evaluations, the commercial product has often already been updated — sometimes without public announcement. Researchers are, in a real sense, studying a moving target through a narrow window.

Subscribe now

Part Two: Why We Still Need These Studies Published

Before concluding that AI research in medicine has become a journal-publishing exercise in futility, it is important to understand why the system still needs to exist — and why abandoning it can create more problems than it solves.

Patient Safety Requires Accountable Evidence

Medicine cannot operate on the technology sector’s principle of “move fast and break things.” A radiology AI that systematically misses early-stage lung cancers, deployed without peer-reviewed scrutiny, causes patient harm that no subsequent correction can undo. High-impact journals — for all their delays — provide layers of oversight that informal dissemination channels do not: independent peer review, statistical auditing, mandatory conflict-of-interest disclosure, and a permanent, traceable scientific record. Removing that filter from clinical AI is not a reform. It is a patient safety risk.

Regulators Require Published Evidence

The U.S. Food and Drug Administration, the European Medicines Agency, and their counterparts worldwide increasingly require published clinical evidence as part of the approval pathway for AI-based medical devices. A preprint — however well-conducted — carries no regulatory weight. If academic medicine retreats from high-impact journals, the evidentiary void is filled by internal validation data produced and controlled by the companies seeking approval. That is not a system that serves patients or clinicians.

Failure Studies Have Long Half-Lives

Not every AI study in medicine aims to demonstrate what a model can do. Some of the most important — and most durable — research documents what AI systems get wrong, where they fail, and which populations they underserve. A study showing that a clinical AI systematically underperforms on images from portable equipment common in resource-limited settings remains instructive for years, regardless of which specific model it evaluated. The failure pattern transfers across model generations even when the model itself is replaced. These studies belong in high-impact journals because their findings compound in value over time rather than decaying.

Subscribe now

Part Three: What Has Genuinely Become Redundant

Intellectual honesty requires acknowledging what critics of the current system are right about. A specific and identifiable category of AI research in medicine is, at the moment of publication, producing diminishing returns.

The Model Comparison Snapshot Problem

Studies whose core contribution is a head-to-head comparison of named AI models — Model A achieved 88%, Model B achieved 79% on benchmark X — have a useful life measured in weeks - months, not years. They are informative at the moment of submission and increasingly archival by the moment of publication. When they appear in prestige journals with high impact factors, they create a misleading impression of settled evidence. Educators who teach from them and clinicians who cite them in practice decisions are working from a photograph of a landscape that has since changed…!

The Reproducibility Problem With Named Models

Commercial AI models change continuously through fine-tuning, infrastructure updates, and prompt/context/loop modifications — usually without public announcement and without version locking. A study evaluating “Frontier GPT-5” as though it were a stable, reproducible entity analogous to a drug formulation is making an assumption that is not true. Unlike a clinical drug trial, where the compound being tested is chemically defined and can be replicated, AI model evaluations may describe a system that no longer exists in the tested form by the time a reader tries to replicate the work. The scientific record preserves not a reproducible finding but a time capsule.

The “Benchmark Mirage”

Performance on a controlled clinical evaluation dataset does not equal clinical readiness. Scores on USMLE questions, curated radiology image sets, or standardized medical vignettes reflect a model’s ability to perform on a specific, tightly controlled task format. They do not reflect how that model behaves when images are low quality, when clinical context is ambiguous, when documentation is incomplete, or when patients present atypically — which is to say, they do not reflect real clinical conditions. Publishing benchmark comparisons in high-impact journals lends them an authority that systematically overstates their clinical meaning.

Share

Part Four: A feasible Framework — Three Tiers of AI Research

Not all AI studies in medicine age at the same rate. The most useful question is not “should this be published in a high-impact journal?” but “what type of contribution is this study making, and how long will that contribution remain actionable?”

This can be a working framework for medical educators and clinicians reading the literature.

Tier 1 — Enduring Contributions

Read it. Teach it. Cite it with confidence.

These are studies whose core findings transfer across model generations because they address the evaluation ecosystem, the clinical deployment context, or structural patterns that recur regardless of which specific models are being tested.

Examples include: novel evaluation methodologies and adversarial testing frameworks; analyses of systemic bias or demographic inequity in AI training data; prospective clinical trials using AI as an intervention with patient outcomes as primary endpoints; failure mode taxonomies linked to clinical consequences; benchmark design critiques that reveal what existing assessments actually measure versus what they are assumed to measure; and systematic reviews or meta-analyses that synthesize evidence across model generations rather than profiling a single one.

These belong in high-impact journals. Their value compounds over time.

Tier 2 — Moderately Durable Contributions

Read it. Use it with context. Revisit it within two years.

These studies address real clinical questions but are partially anchored to the capabilities of current model generations, meaning their conclusions will need updating as models improve.

Examples include: clinical workflow integration studies that embed AI into real healthcare environments and measure operational downstream effects (time savings, error rates, staff workload); comparative effectiveness studies pitting AI-assisted clinical performance against clinician-only performance on real-world, multi-site data; and interpretability analyses that explore how a class of model approaches a clinical task, rather than evaluating a single named version.

These belong in journals, but readers should treat specific quantitative findings as estimates valid for current-generation models rather than permanent benchmarks.

Tier 3 — Time-Stamped Observations

Read it as context. Do not teach it as established fact. Check the date.

These studies make claims that are accurate when submitted and increasingly historical by publication. Their value lies in establishing a moment-in-time record, not in providing durable clinical guidance.

Examples include: head-to-head accuracy comparisons of named commercial or open-source AI models on standardized test sets; prompt engineering evaluations on clinical question banks; single-institution retrospective accuracy assessments using proprietary datasets; and leaderboard-style publications that rank current model performance without a methodological framework that transfers to future evaluation.

These are better suited to preprint servers, conference proceedings with explicit versioning, or journal formats designed for rapid technical communication — not full research articles in prestige publications.

Subscribe now

Part Five: Decoding a Real Example — The Nature Medicine Paper Published recently

The Nature Medicine study that opened this piece offers a nearly perfect illustration of why these tiers matter — because it contains contributions from all three within a single paper.

What Looks Like Tier 3 (But Isn’t the Whole Story)

On the surface, the paper reads like a model comparison study: GPT-5 scored X% on NEJM questions; Gemini 2.5 Pro scored Y%; Claude 3.5 did Z on JAMA benchmarks. These specific figures will be historical data within twelve months. They represent exactly the kind of time-stamped snapshot that does not, on its own, justify a 25-page article in a flagship journal.

What Is Actually Tier 1 — The Durable Contributions

Beneath the model comparison lies a set of contributions that will outlast every model named in the paper.

An adversarial testing framework. The paper introduces six structured stress tests — assessing how models behave when clinical images are removed, when answer order is randomized, when distractor options are introduced, when visual content is substituted — that can be applied to any future model. This is a methodology, not a snapshot. It transfers directly to evaluating the next generation of systems, and the one after that.

A clinical failure taxonomy. Table 1 of the paper maps eight categories of model failure — including visual misperception, heuristic dependence, logical inconsistency, and unsafe recommendation — directly to their downstream clinical consequences, such as missed findings, diagnostic misguidance, and treatment harm. No equivalent structured vocabulary exists elsewhere in the literature. This taxonomy will be cited and extended long after GPT-5 is a historical footnote.

A benchmark profiling method. The authors created a clinician-annotated rubric that positions each medical AI benchmark on two axes: reasoning complexity and visual complexity. This reveals, for the first time in structured form, that widely used benchmarks measure fundamentally different cognitive demands — which explains why a model can perform well on one and poorly on another despite similar headline accuracy scores. This insight applies to every benchmark and every model that follows.

The shortcut learning finding. Perhaps most important for clinical educators: the paper shows that several tested models maintain high accuracy on medical image benchmarks even when the clinical image is completely removed from the input. The models are relying on text cues, answer-position patterns, and memorized associations — not genuine visual reasoning. This failure pattern will recur in future model generations until the benchmarks themselves are redesigned. It is a finding about a structural flaw in how we evaluate clinical AI, not about any particular system’s capabilities.

Subscribe now

The Lesson From This Paper

The paper deserved to be published in Nature Medicine. Its methodological contributions are genuinely important. But those contributions were framed as supporting material for a model comparison that will age poorly. Had the paper led with “we propose an adversarial evaluation framework and a clinical failure taxonomy for health AI” rather than “we evaluated GPT-5 and Gemini,” it would read as foundational work rather than a benchmark snapshot that happens to contain foundational work…!

The redundancy was not in the paper’s existence. It was in its framing choices..!

Part Six: What This Means for You

For Medical Educators

When incorporating AI research into curriculum, we need to explicitly teach students to distinguish between study type and study age. A 2024 paper with a transferable methodology may be more relevant today than a 2026 paper reporting model comparison results. We have to train students to ask, before accepting a finding: “Is this a claim about a specific model, or a claim about how AI behaves structurally?” The latter survives model updates. The former does not.

We have to update ourselves and consider building AI literacy modules around landmark methodology papers and failure taxonomies rather than around model performance comparisons. The adversarial testing framework and clinical failure taxonomy described above are the kind of conceptual tools that will serve clinicians across every model generation they encounter in their careers.

For Clinicians Reading the Literature

When you encounter an AI study, four questions will help in orienting your reading before you reach the results:

• First, which specific model version was tested, and when was the study submitted — not published? A seven-month gap between submission and publication is standard. Much can change.

• Second, is the claim about a specific model’s performance, or about a structural behavior pattern? One ages; the other does not.

• Third, was the evaluation conducted on a curated controlled dataset, or in a real clinical environment with real patients? Benchmark performance and clinical performance are not the same thing, and should not be treated as equivalent.

• Fourth, does the study include validation under adversarial or perturbation conditions — degraded images, ambiguous inputs, distractor options — or only under optimal conditions? Performance under stress reveals clinical readiness in a way that optimal-condition benchmarks do not.

For Healthcare Administrators and Policy Decision-Makers

It is never wise to use a single peer-reviewed AI benchmark study as the primary evidence base for procurement or deployment decisions. Ask vendors for independent evaluation data, stress-test results, and failure mode documentation. The existence of a journal publication does not mean the specific model version evaluated is the one being sold or deployed. Require version-locked documentation and ongoing post-deployment monitoring as part of any AI contract.

Share

Part Seven: Red Flags — How to Spot a Time-Stamped Paper

When reading any AI study in medicine, these signals suggest you are looking at a time-stamped observation rather than a durable contribution:

1. The paper’s primary contribution is a head-to-head accuracy comparison between named commercial models.

2. The dataset is a standardized multiple-choice benchmark (USMLE, NEJM quiz, JAMA Clinical Challenge) used without modification.

3. There is no adversarial or perturbation testing.

4. The paper evaluates a single institution’s dataset.

5. The word “version” does not appear in the methods section in relation to model specification.

6. The gap between submission date and publication date exceeds four months and no model versioning information is provided.

7. The conclusion includes phrases like “model X outperforms model Y” without specifying the conditions under which that comparison holds.

None of these signals mean the paper is worthless. They mean the findings should be read as a dated observation rather than a clinical standard.

Subscribe now

Part Eight: What the System Needs to Change

Individual reading skills are necessary but not sufficient. The publication infrastructure itself needs reform in three areas.

Journals can start categorising AI submissions by study type at intake. A benchmark comparison study and a methodological innovation are different scientific artifacts. Journals in other fields already make this distinction through article type classifications. Medical journals should explicitly designate AI benchmark studies as time-sensitive technical reports — with appropriate formatting, shorter length, and explicit versioning requirements — distinct from full research articles making durable clinical claims.

Model versioning should be mandatory, like drug lot numbers. Pharmaceutical trial reports require detailed documentation of the compound’s formulation, batch number, and storage conditions because reproducibility depends on knowing exactly what was tested. AI model evaluations should require equivalent disclosure: the exact model version, API parameters, inference conditions, and any known changes to the model between submission and publication.

Academic incentives should reward compounding contributions. A researcher who develops an adversarial evaluation framework that fifty subsequent teams apply to new models has contributed more durable scientific value than one who publishes a model comparison that no one can replicate. Career advancement structures in academic medicine — hiring decisions, grant evaluations, promotion criteria — should explicitly value open methodologies, reusable frameworks, and living review contributions, not only high-impact single-paper output.

Subscribe now

Five Things to Remember

1. Publication date is not the same as relevance date. Always check submission date and model version.

2. Methodology papers outlast model comparison papers by years, sometimes decades.

3. Benchmark performance and clinical performance are not the same thing. Treat them as distinct.

4. Failure mode and bias studies have longer shelf lives than capability demonstration studies.

5. A finding about how AI behaves structurally is more valuable than a finding about how one named model performed last quarter.

   Thanks for reading! This post is public so feel free to share it.

   Share

Every June, India sends its children back to school with new uniforms, polished shoes, sharpened pencils, labelled water bottles and a strange confidence that the academic year begins on a clean slate.

It does not.

In pediatric outpatient departments, the new school year announces itself differently. It arrives as fever. It arrives as cough, cold, wheeze, sore throat, vomiting, fatigue, poor feeding, breathlessness, and that familiar phrase parents bring with them: “Doctor, school has just reopened.”

Subscribe now

When Schools Reopen, Viruses Return

Within a few weeks of school reopening, the pattern becomes hard to miss. Daycare children start falling sick. Nursery children follow. Primary school children carry infection back home. Pediatric OPDs swell. Wards begin to fill. Step-down PICUs see more children with viral wheeze, bronchiolitis, influenza-like illness and Pneumonias. Some children recover in two days. Some need oxygen. Some need high-flow support. Some fragile infants and children with underlying disease get pushed into severe illness by viruses that began their journey in crowded classrooms and daycare rooms.

And the chain does not stop with the child…!

This happens every year. We discuss it in OPD corridors, pediatric meetings, parent WhatsApp groups and school offices. We complain about it. We warn each other about it. But we still do not have a simple, widely used, school-facing protocol for the June to September viral surge.

That is the failure.

India has disease surveillance and influenza like illnesses data. India has pediatricians who can recognize the seasonal rise before any dashboard does. India has schools that can mobilize parents overnight for fees, uniforms, transport changes, annual day costumes and examination circulars. Yet when it comes to preventing predictable viral spread after school reopening, most schools still operate on improvisation, denial and attendance pressure.

We are not dealing with one virus. That is the first point schools must understand. The child with fever and cough in July may have influenza, RSV, rhinovirus, adenovirus, human metapneumovirus, parainfluenza, SARS-CoV-2, or a combination of viruses. Clinically, many of these infections look similar at the beginning. A mild cold in one child can become bronchiolitis in an infant, wheeze in an asthmatic child, pneumonia-like illness in another, and hospitalization in a high-risk child.

Subscribe now

The second point is more uncomfortable. Schools and daycare centres are perfect transmission machines unless they are deliberately managed. Young children sit close, share toys, touch faces, wipe noses, cough openly, forget hand hygiene, sleep in groups, eat together, cry into caregivers’ shoulders, and often cannot explain symptoms early. Nursery and daycare children cannot be expected to behave like trained adults. Infection control in these spaces cannot depend on the child’s discipline. It has to depend on adult systems.

The third point is that school reopening in India overlaps with the monsoon respiratory-virus season in many regions. This is not just an anecdote from one pediatric OPD. Indian surveillance and hospital studies have repeatedly shown that influenza-like illness and severe acute respiratory infection rise in seasonal patterns, with monsoon and post-monsoon peaks in several parts of the country. The exact virus changes. The geography changes. The timing varies between North India, South India, Western India, Eastern India and the Northeast. But the larger truth remains: June to September is not a neutral period. It is a biologically risky window.

Yet our school calendar behaves as if infection risk is irrelevant….!

We reopen crowded institutions into a known transmission season. We allow febrile children to attend because parents cannot miss work, because exams are near, because daycare is essential, because “it is only a cold,” or because schools quietly reward attendance more than responsibility. Then we wait for pediatric OPDs to absorb the consequences.

Subscribe now

This is not a call to close schools. That would be irrational and inappropriate. Children need school. They need learning, food, play, language, friendships, discipline, teachers, routine and escape from the loneliness that screens have created. Repeated closures damage children in ways we have already seen during the pandemic.

The question is not whether schools should remain open. They should.

The question is why they remain open without a respiratory infection protocol…!

A sensible school-health policy does not begin with panic. It begins with thresholds. It tells schools what to do before the outbreak, during the outbreak, and after the outbreak. It tells parents when not to send a child. It tells teachers what to watch for. It tells daycare staff how to reduce spread. It tells pediatricians and public health authorities when a cluster deserves attention.

At the minimum, every daycare, nursery, kindergarten and primary school should have a June to September respiratory illness plan.

Share

The first step is pre-reopening preparation. Schools should send a health advisory before reopening, not after half the class is coughing. Parents should be told clearly that fever, active cough, breathing difficulty, vomiting, loose stools, lethargy or poor feeding are valid reasons to keep a child home. Schools must stop demanding medical certificates for every short illness absence. That single habit drives unnecessary OPD visits, exposes other children in clinics, burdens parents, and turns pediatricians into attendance clerks…!

The second step is vaccination awareness. Influenza vaccination is not a magic shield, but it is a rational preventive tool, especially for children at higher risk of severe illness, children with asthma or chronic lung disease, congenital heart disease, neurodevelopmental disability, immunosuppression, prematurity history, and those living with elderly or medically fragile adults. Staff in daycare and nursery settings should also be part of this conversation. A school cannot claim to care about child health and then ignore preventable influenza morbidity.

The third step is the sick-child rule. It has to be simple enough to print on a school notice board. Fever means stay home. A child with fever should return only after being fever-free for at least 24 hours without antipyretics and clinically better. A child with persistent cough, breathlessness, fast breathing, poor oral intake, drowsiness, chest indrawing, bluish lips, or oxygen saturation concern needs medical evaluation, not classroom sympathy. Toddlers with active respiratory symptoms should not be sent to daycare merely because adults have no backup plan. That is harsh to say, but necessary. Daycare convenience cannot become community transmission policy.

The fourth step is class-level surveillance. Schools already maintain attendance registers. They need one additional column during June to September: reason for absence. Fever. Cough. Cold. Wheeze. Vomiting. Loose stools. Rash. Unknown. If three or more children in one class develop influenza-like symptoms within a few days, the school should alert its medical advisor or local health contact. If absenteeism rises sharply in a class, the school should not wait for rumours. It should act.

Subscribe now

The fifth step is layered control. During clusters, schools can reduce assemblies, improve ventilation, avoid indoor crowding, stagger meals, clean high-touch surfaces, disinfect shared toys in daycare, keep symptomatic staff away, and encourage masks for older children and staff during active outbreaks. None of this is dramatic. None of this requires shutting the school; but requires seriousness.

The sixth step is honest parent communication. Schools should stop sending vague circulars saying “viral fever is spreading, take care.” That means nothing. Parents need clear instructions: do not send a febrile child; inform the class teacher if your child has influenza-like illness; seek care urgently for breathing difficulty, poor feeding, persistent high fever, lethargy, seizures, dehydration or worsening wheeze; protect infants and elderly household members if one child is sick. Good communication reduces panic. Bad communication creates gossip.

The seventh step is hospital-public health linkage. Pediatricians are often the first to see the signal. When multiple children from the same school or locality present with similar illness, there should be a simple reporting pathway. Not every cluster needs laboratory testing. Not every outbreak needs media attention. But repeated clustering should not remain trapped as OPD anecdote. It should become surveillance intelligence.

This is where India’s current approach is incomplete. We have ILI and SARI definitions. We have national surveillance networks. We have laboratory capacity in many regions. We have outbreak investigation experience. But these systems often sit above the everyday school setting. Schools need a practical bridge between clinical medicine and public health.

Share

A school respiratory protocol should not be a 90-page document that nobody reads. It should be a two-page operational checklist.

• Before reopening: advisory, vaccination awareness, sick-child rule, ventilation check, daycare cleaning plan.

• During routine weeks: absenteeism tracking, hand hygiene, cough etiquette, symptom-based exclusion, parent communication.

• During clusters: class-level alert, temporary reduction of crowding, masking where feasible, enhanced cleaning, medical consultation, public health intimation if severe or unusual.

• During severe outbreaks: targeted class-level pause only if needed, not blind school closure.

• After the season: review what happened, how many children were absent, how many were hospitalized, what worked, and what failed.

It protects learning by preventing avoidable disruption. It protects children by reducing viral load and repeated exposure. It protects hospitals by flattening seasonal surges. It protects grandparents by recognizing that the school-child-household chain is real.

The opposition to such a protocol will be predictable. Some will say children fall sick anyway. True. Children do fall sick. A young child may have several respiratory infections a year. But inevitability is not the same as helplessness. Road accidents happen too. We still use helmets, seat belts, speed limits and school-zone rules. Viral infections will continue. That is not an argument against prevention. It is an argument for adult competence.

Some will say schools cannot manage this. That is false. Schools already manage far more complex systems when they want to: transport routes, fee collection, examinations, uniform rules, digital portals, parent meetings, CCTV, annual functions and competitive coaching schedules. A fever policy is not beyond institutional capacity. It is beyond institutional priority.

Some will say parents will not cooperate. Some will not. That is exactly why schools need written policy. A parent who sends a febrile child should not be treated as merely “adjusting.” They are placing other children and households at risk. Working parents need support, but the solution cannot be to convert daycare into an infection exchange.

Some will say this is alarmist. It is not. Alarmism is shouting after the ward is full. Public health is acting before the ward fills.

The June to September pediatric viral surge is not a mystery. It is a recurring event hiding in plain sight. Every pediatrician knows it. Every school has seen it. Every parent has lived it. Every hospital has absorbed it. What we lack is the courage to stop calling it normal.

The first lesson of the school year should not be that fever spreads faster than policy.

If we can reopen schools with new timetables, new books and new fee structures, we can reopen them with a respiratory illness protocol. The children are already giving us the data. They give it through crowded OPDs, full pediatric wards, anxious parents, exhausted nurses and oxygen beds occupied by illnesses that began as “just a cold” in a classroom.

Every June, the virus enters school with the children.

The question is whether adults will finally be ready before it does…..!

Further readings:

1. https://pmc.ncbi.nlm.nih.gov/articles/PMC10624221

2. https://link.springer.com/article/10.1186/s12879-024-09536-1

3. https://ijmr.org.in/viruses-causing-severe-acute-respiratory-infections-sari-in-children-5-years-of-age-at-a-tertiary-care-hospital-in-rajasthan-india/

4. https://journals.lww.com/jfmpc/fulltext/2024/13030/an_outbreak_of_h1n1_influenza_a_among_the_students.47.aspx

5. https://ncdc.mohfw.gov.in/uploads/glimpse_pdfs/1770281948_Pandemic-Influenza-H1N1-2009.pdf

6. https://www.cdc.gov/children-and-school-preparedness/infection-prevention/docs/IPC-Science-Brief_508.pdf

I did not ask immediately.

In medicine, some silences arrive already diagnosed.

After a few moments, he told me the story. His biology professor from his pre-university college days had called him from far away. He was 65 years old now. A year earlier, this man had still been teaching biology to pre-university students. He was working, explaining life to young minds, using his voice, his hands, his memory, his body, and his years of accumulated confidence in the classroom. Then came an illness that began as weakness and later acquired a name: Guillain-Barré syndrome.

The name is tidy. The disease is not…!

On paper, Guillain-Barré syndrome is an acute immune-mediated neuropathy, often triggered after an infection, in which the body’s immune system attacks peripheral nerves. In real life, it can be a terrifying betrayal. A person who was walking, speaking, swallowing, eating, teaching and breathing normally may, over days, begin to lose the ordinary functions that make life feel secure. Legs fail. Hands weaken. Voice changes. Swallowing becomes unsafe. Breathing, that quiet background rhythm of existence, becomes an excessive work...!

We speak of “ascending paralysis” in textbooks. But no family experiences it as a phrase. They experience it as panic moving upward.

Subscribe now

The family first took him to a local hospital. The doctor there assessed him and referred him to a tertiary care hospital in Bengaluru. The patient was shifted by road in an ambulance. Somewhere on the way, near Chitradurga, his condition worsened. He was gasping. A doctor in Bengaluru, guiding them over the phone, advised them to stop at a nearby teaching medical college hospital. He was stabilized there for a few hours. Then the family continued the journey and finally reached a corporate tertiary care hospital in Bengaluru.

That journey took more than forty hours…!

Forty hours is not a number when a nerve disease is moving toward the lungs. Forty hours is not distance on a highway. It is biology (pathology) unfolding without permission. It is oxygen, panic, traffic, waiting, pleading, calling, guessing, hoping. It is a family sitting beside a deteriorating teacher, trying to understand whether they are moving toward safety or merely moving.

In the months (to be precise 13 months) that followed, he went through tracheostomy, cardiac arrests, prolonged intensive care, multiple sessions of plasmapheresis, supportive treatment, nursing care, physiotherapy and a year of uncertainty that no discharge summary can contain. His body fought. Medicine fought. His family fought. There must have been days when one vital sign felt like victory, and days when a small setback must have felt like betrayal.

Subscribe now

He survived.

That sentence deserves respect.

He survived because modern medicine exists. He survived because ventilators exist. He survived because intensive care exists. He survived because plasmapheresis exists. He survived because tracheostomy, monitoring, antibiotics, suctioning, nutrition, physiotherapy and rehabilitation exist. He survived because nurses, intensivists, neurologists, physiotherapists, technicians, family members and invisible hands kept him alive when his own body had almost withdrawn cooperation.

It is fashionable now to speak of hospitals, doctors, pharma and modern medicine only through suspicion. Some suspicion is earned. Healthcare has commercial excess, institutional arrogance, poor communication, avoidable delay and ethical blindness. But criticism becomes foolish when it forgets the obvious: a 65-year-old man with severe Guillain-Barré syndrome, respiratory failure, cardiac arrests and prolonged ICU dependence surviving for a year is not a small thing. It is a triumph of science, skill, endurance and human labour…!

Modern medicine did not fail this man. It pulled him back from death.

But survival is not the same as return.

Today, after one year, he is wheelchair-bound. He has vocal cord palsy. He has lost mobility. His life has been rearranged by one illness. His family’s life has also been rearranged. In chronic neurological disability, the patient is not the only patient. The family becomes an unpaid ICU, an informal rehabilitation centre, a financial shock absorber, a nursing unit, a transport service, a rehabilitation coordinator, and often, silently, a casualty.

Someone must feed. Someone must lift. Someone must clean. Someone must arrange physiotherapy. Someone must understand prescriptions. Someone must call hospitals. Someone must borrow money. Someone must stop working. Someone must sleep lightly, because the patient may need help at night. Families do not merely “support” such patients. They are absorbed into the illness.

Share

This is what disturbed my colleague. And this is what disturbed me…!

Not because Guillain-Barré syndrome is always preventable. It is not. Not because every patient with GBS can be brought back fully. They cannot. Not because every bad outcome is someone’s fault. That is immature thinking. Disease has its own violence. Biology is not obedient. The human body has limits, and medicine, even at its best, does not erase them.

But some cases do not allow us to hide behind disease alone.

• What if the first recognition had been sharper?

• What if respiratory weakness had been anticipated earlier?

• What if the ambulance had been equipped to support a deteriorating neurological patient, rather than function mainly as a vehicle with a siren?

• What if the receiving hospital had been alerted before transfer?

• What if the referral had been not just to a bigger place, but to the right place, at the right time, with stabilization and handover?

• What if a region with so many hospitals, medical colleges and corporate centres had a referral network humane enough that a frightened family did not have to depend on a stranger’s voice on the phone to find its way?

These questions are uncomfortable because they produce no clean villain.

The problem is not one doctor, one hospital or one ambulance. The problem is a system in which responsibility becomes blurry exactly when clarity is needed most. Everyone does a part. Nobody owns the whole journey. The patient moves from one institution to another, but the illness does not wait for paperwork, bed availability, geography, institutional ego or the family’s ability to negotiate.

Subscribe now

This is one of the great paradoxes of modern healthcare. We have ventilators, but not always timely ventilation. We have neurologists, but not always reachable ones. We have plasmapheresis, but not always where the patient first needs it. We have hospitals everywhere, but not always a system that connects them intelligently.

The body is fragile. That is the first truth.

Most days we forget this. The body obeys so quietly that we mistake its obedience for simplicity. We walk, speak, eat, sleep, teach, drive, write, swallow and breathe without gratitude. Then one immune-mediated storm attacks the peripheral nerves and the fiction collapses. Suddenly, every ordinary act reveals itself as a sophisticated biological achievement. Breathing is not simple. Swallowing is not simple. Standing is not simple. Voice is not simple. Life is not simple. It only appears simple when the body is kind.

The second truth is that modern medicine is astonishing.

A century ago, many patients with severe neuromuscular respiratory failure would simply have died. Today, a patient can be ventilated, fed, monitored, protected from infection, treated with plasma exchange or immunoglobulin, rehabilitated and kept alive through a storm that would once have been fatal. This is not sentiment. This is science turned into infrastructure. A ventilator is not just a machine. It is physiology made mechanical. An ICU is not just a room. It is organized vigilance.

But the third truth is harsher.

Modern medicine can save a life, while the healthcare system wounds the same life through delay, confusion, poor triage and abandonment disguised as referral…!

Subscribe now

This is where the ethical question begins.

Medical ethics is not only consent forms, confidentiality and signatures. Those matter, but they are not enough. In an emergency, ethics lives in recognition, communication, stabilization, transfer and ownership.

• Do we recognize severity early?

• Do we tell the family the truth without either panic or false reassurance?

• Do we stabilize before sending?

• Do we call ahead? Do we document clearly?

• Do we transfer to the correct level of care? Do we make sure the ambulance can handle what may happen on the road?

Referral is not clerical work. Referral is a clinical intervention…!

Done well, it saves. Done badly, it becomes risk displacement. A patient is moved, but responsibility is not. A name is written, a destination is suggested, an ambulance leaves, and everyone feels that something has been done. But the patient’s biology continues in real time. In Guillain-Barré syndrome, time does not behave like a calendar. It behaves like nerve conduction, respiratory muscle strength, swallowing safety and autonomic instability. A few hours may mean little in one patient and everything in another.

This case also exposes a middle-class (in particular) illusion about healthcare: that Grand looking hospitals mean healthcare exists….

They do not.

A city may have medical colleges, private hospitals, specialists, diagnostic centres and ICUs, and a patient may still be functionally alone inside that abundance. Healthcare is not the presence of buildings. It is the reliable movement of the right patient to the right care at the right time, without forcing the family to become logisticians during a medical emergency.

We see versions of this all the time. A child with sepsis. A pregnant woman with eclampsia. A stroke patient. A newborn with respiratory distress. A teacher with ascending paralysis. The story changes, but the pattern is familiar. The disease begins. The family searches. The first facility hesitates or refers. The ambulance is uncertain. The receiving centre is not ready. The definitive care comes late. Science rescues what it can. The family pays for what the system did not organize. Later, everyone calls it fate.

Some of it may be fate.

But not all of it.

That distinction matters. If we call everything negligence, we become unfair and stupid. If we call everything fate, we become morally useless. The truth is harder. Some damage comes from disease. Some comes from delay. Some comes from the randomness of access. Some comes from the fact that healthcare is still designed around institutions rather than around the patient’s journey through them.

Subscribe now

What should disturb us is not only that such stories happen. What should disturb us is that we have grown used to them. We discuss them in corridors, shake our heads, praise the final centre, pity the family and move on. We have blame culture when something becomes public, and silence culture when it does not. What we do not have enough of is just culture: the discipline to examine a bad journey honestly, without immediately hunting for a scapegoat or protecting reputation.

A serious healthcare system would study such journeys.

• Where was severity first recognized?

• Was respiratory function assessed?

• Was the transfer appropriate?

• Was the ambulance equipped?

• Was the receiving unit informed?

• Was the family counselled?

• Could the patient have reached definitive care earlier?

Not to punish blindly. Not to shame individuals. But to stop pretending that a patient’s path through the system is someone else’s problem.

This one case contains everything: the fragility of the human body, the power of modern medicine, the exhaustion of families, the ethics of referral, and the unfinished accountability of our healthcare system.

The professor survived. His family survived in another way. My colleague was shaken because he knew the patient not as a bed number, but as a teacher from his own past, a human being whose life had been permanently altered. I was shaken because the story felt painfully familiar. Every doctor has seen versions of it. Every family fears becoming the next version.

There are cases that teach diagnosis. There are cases that teach humility. There are cases that expose the system.

This was all three.

Kudos to the patient and his family for enduring what no family should have to endure. Kudos to modern medicine for pulling him back from death. Kudos to every nurse, doctor, technician, physiotherapist and caregiver who helped keep him alive through that long year.

But shame on us if we stop at admiration…!

The correct response to such a story is not sympathy alone. It is referral discipline. It is early recognition. It is equipped transfer. It is accountable handover. It is regional coordination. It is the refusal to let another family travel forty hours through uncertainty while a treatable neurological emergency climbs silently toward the lungs.

A human body can fail without warning.

A healthcare system should not…!

Cardiometabolic multiple long-term conditions (MLTC), defined as the coexistence of two or more cardiometabolic diseases such as diabetes, cardiovascular disease, and chronic kidney disease, are increasingly prevalent and represent a growing challenge for health systems worldwide.

Evolution of terminology of multimorbidity

•The term multimorbidity was first described by Brandlmeier in an urban German general practice in 1976.

•In view of the confusion in terminologies between comorbidity and multimorbidity, a distinction was proposed in 1996 with the term comorbidity implying combined effects in reference to an index chronic condition (eg, stroke in a person with diabetes) and multimorbidity meaning co-occurrence of conditions without a single condition holding a priority over other co-occuring conditions.

•The concept of multimorbidity as two or more chronic conditions was adopted by WHO in 2008 in recognition of its growing burden in an ageing society.

•In 2018, the Academy of Medical Sciences published a landmark international policy report titled Multimorbidity: a priority for global health research. The report highlighted the rising burden of multiple long-term conditions and the lack of coherent evidence on its causes, prevention, and management.

•Patients do not find multimorbidity an acceptable term to describe their condition and associate it with a negative image, giving the impression to patients that they are in poor health, their body is failing them, they may be dying soon, and there is nothing they can do about their condition. The UK National Institute for Health and Care Research Strategic Framework therefore emphasised their preferred term as multiple long-term conditions.

Cardiometabolic conditions most strongly contributing to observed cardiometabolic multiple long-term condition clusters

Subscribe now

MLTC across the life course

Subscribe now

Determinants of cardiometabolic MLTC

Cardiometabolic MLTC arise from a complex interplay of biological, behavioural, social, and environmental determinants; genetic susceptibility; metabolic pathways; ageing; adiposity; insulin resistance; and chronic inflammation, with some determinants being modifiable . One genome-wide association study showed that there are 11 independent genetic variants associated with psycho-cardiometabolic MLTC, with a polygenic risk score being predictive of co-occurrence of depression, coronary artery disease, and type 2 diabetes, and suggested eight causal factors (BMI, body fat percentage, LDL cholesterol, total cholesterol, fasting insulin, income, insomnia, and childhood maltreatment).

Subscribe now

The sequencing of conditions and effect on outcomes

Share

Methodological challenges and future research gaps

Subscribe now

Biological and mechanistic pathways of cardiometabolic multiple long-term conditions

Share

Role of genetic and epigenetic biomarkers

Subscribe now

Interventions for the prevention and management of cardiometabolic multiple long-term conditions

Multifaceted approach to cardiometabolic MLTC

Subscribe now

Evidence base for prevention and management interventions in cardiometabolic multiple long-term conditions

Share Vikram Sakaleshpur Kumar

Key research gaps

Population-level interventions

•High-quality evaluations using real-world data, rather than modelling studies, are needed for assessing the effect of population-level interventions on the prevention and management of cardiometabolic multiple long-term conditions (MLTC).

•There is limited evidence on populations at high risk, such as those from socioeconomically deprived areas and minority ethnic communities.

•Most studies have been in older populations and further research is needed for younger populations who are facing a growing burden of MLTC.

•There is a scarcity of high-quality trials in low-income and middle-income countries.

Individual-level interventions

•Clarity is required as to what are the most meaningful primary endpoints for intervention trials in individuals with, or at risk of, cardiometabolic MLTC.

•There is insufficient high-quality evidence that in-person behavioural interventions in people with cardiometabolic MLTC produce sustained improvements in risk factor control and cardiometabolic outcomes.

•Evidence for digital and artificial intelligence-based delivery of lifestyle interventions to improve clinical outcomes and address health inequalities remains scarce.

•Evidence is needed to understand optimal timing and intensity of treatment with novel pharmacotherapies in different risk groups across the life course in cardiometabolic MLTC.

•Further evidence is required to understand the effect of rational polypharmacy and deprescribing on clinical outcomes in cardiometabolic MLTC across the life course.

System-level interventions

•Further research is required on pragmatic models of care for improving outcomes of people with established cardiometabolic MLTC.

•There is a lack of multimorbidity-specific outcomes, with most studies prioritising single behaviours or conditions rather than outcomes relevant to MLTC (eg, treatment burden, functional status, or integrated risk reduction).

•Future studies should include complex interventions integrating behavioural interventions, digital tools, care coordination, and health system infrastructure.

•Evidence for system-level digital interventions, including artificial intelligence and large language model-enabled decision-support systems, remains scarce.

Share

Key evidence gaps and future research priorities in cardiometabolic MLTC

Shared risk factors and mechanisms

•Identify shared causal molecular pathways for therapeutic targeting

•Delineate the neuroendocrine–immune axes that drive the development of cardiometabolic multiple long-term conditions (MLTC)

•Delineate the temporal stages in the evolution of cardiometabolic MLTC to identify critical windows for intervention

SDOH

•Standardise consensus-based criteria for the uniform definition of social determinants of health (SDOH) and objective measurements of cardiometabolic MLTC and disease severity

•Broaden the scope of SDOH by capturing socioeconomic, environmental, behavioural, and structural factors (eg, social support, housing quality, and neighbourhood safety) using standardised instruments

•Investigate the link between SDOH and cardiometabolic MLTC by conducting prospective real-world studies in diverse populations, including in marginalised groups and different health-care settings

Early-life exposures and programming

•Delineate the mechanisms by which early-life exposures alter the composition and functions of the gut microbiome

•Define the window-specific contribution of different developmental periods, from preconception, in utero, infancy, and adolescence to the final risk of cardiometabolic MLTC

•Analyse longitudinal trajectories for determining the crucial transition point at which cardiovascular and metabolic risk factors solidify into cardiometabolic MLTC

Genetic and epigenetic biomarkers

•Identify genetic modifiers within high-burden cardiometabolic MLTC, such as cardiovascular–kidney–liver–metabolic conditions

•Identify possible rare and structural variants and characterise their contributions to account for missing heritability

•Determine the relative influence of genetic, phenotypic, and environmental crosstalks on interindividual differences in MLTC progression

•Determine the cell-specific, persistent epigenetic changes (eg, DNA methylation and histone modifications in adipocytes or endothelial cells) that explain the effect of early-life exposures

•Measure and validate mitochondrial damage and dynamics markers (eg, cell-free mitochondrial DNA and key metabolites) in ethnically diverse cohorts

Emerging non-genetic biomarkers

•Integrate multiomics data for the discovery of novel, cross-disease biomarkers for improved risk stratification, prediction, and prognostication of cardiometabolic MLTC

•Validate multiomics signatures that can predict the transition from a single disease to cardiometabolic MLTC and the disease trajectories

•Delineate circulating molecular ageing markers that can predict and mediate inflammation-driven organ damage across various tissues

Further reading:

1. https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(26)00606-9/fulltext

2. https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(26)01078-0/fulltext

3. https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(26)00607-0/fulltext

4. https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(26)00608-2/fulltext

He had just had a baby. The delivery happened in a famous hospital. The mother was fine. The baby was fine. The rooms were good. The staff were polite. The process was smooth. The family was reassured. The bill was around five lakh rupees.

Then he asked me, “Did my spending five lakhs for my child’s birth assure me of something more than a private delivery in a government hospital for free? Did it change any outcome?”

I did not have a clean answer…!

That embarrassed me a little, because I work in a medical college. I Head the Pediatric department. I sit in academic sessions, mortality meetings, morbidity meetings, briefings, debriefings, audits, departmental reviews, institutional discussions, and all the small rituals through which medical systems try to convince themselves that they are serious about quality. I know many government and teaching hospital teams that are deeply committed, technically strong, underpaid, overworked, and still deliver remarkably good care. In many such places, childbirth happens for free, or at a cost that may be less than what one pays for a family dinner in a large city hotel. Women get reasonable labour support. Newborns get essential care. Emergencies are handled. Complications are referred or managed. Life goes on without marble flooring and valet parking…!

Subscribe now

And yet, when this friend asked me whether five lakhs had bought him better medical outcomes, I could not answer with the confidence that an academic should have.

That is the discomfort.

In India, childbirth is never just childbirth. It is biology, family anxiety, social signalling, fear of regret, reputation of the doctor, pressure from relatives, fear of government hospitals, aspiration for privacy, and the deep middle-class desire to say, “We did the best possible thing.” The hospital is not chosen only for its clinical protocols. It is chosen for the feeling it gives the family when everyone is afraid. A polished reception desk, a known obstetrician, a clean room, a nurse who answers quickly, a neonatologist available on call, a lift that works, a bill printed with confidence, all of this becomes part of the experience of safety.

But experience of safety is not the same as safety….!

That is where we confuse hospital service with hospitality….!

Hospitality is not useless. It is not some shallow luxury to be mocked by those of us sitting in teaching hospitals. Privacy matters. Clean toilets matter. Respectful speech matters. A calm labour room matters. A husband or mother being treated with dignity matters. A woman not being shouted at during labour matters. A newborn not being passed around in a chaotic corridor matters. A family receiving clear explanations matters. These are not “extra” things. They are part of humane care.

The problem begins when we assume that because the hospitality is better, the healthcare outcome must also be better.

Subscribe now

A five-lakh birthing package may give comfort, privacy, speed, reputation, and emotional reassurance. It may also give excellent medical care. Many private hospitals do. But those are two different claims. One is visible. The other must be measured.

The visible things are easy to sell. Room category, consultant name, attendant bed, food, nursing attention, package inclusions, luxury, convenience.

The invisible things are harder. What is the hospital’s primary caesarean section rate? How often is labour induced? How many babies are admitted to NICU? How many of those admissions are medically necessary? What is the rate of postpartum hemorrhage? How many mothers need ICU care? How often are antibiotics used? How early is breastfeeding started? How many babies are readmitted with jaundice, sepsis, or feeding problems? What is the episiotomy rate? How many women report respectful care? How many felt coerced? How many were actually offered a trial of labour when appropriate?

These are not abstract academic questions. These are the questions a family should be able to ask before spending five lakhs.

Share

But in the Indian birthing market, families rarely get such answers. They get reputation. They get word of mouth. They get WhatsApp recommendations. They get Google reviews. They get the name of a famous doctor. They get the comfort of a known brand. They get a package estimate. They get fear converted into a purchasable plan.

And maybe that is enough for many families. Maybe they do not want an outcome dashboard. Maybe at the time of childbirth, nobody wants to read hospital statistics. They want the mother and baby to come home safely. They want someone competent to be present if things go wrong. They want dignity. They want predictability. They want not to be trapped in the chaos that many Indians associate with public hospitals.

That fear is real. It cannot be dismissed.

Government and teaching hospitals may have clinical strength, but many of them still fail the dignity test. They may manage hemorrhage well but communicate badly. They may save a baby but humiliate the family. They may have good doctors but broken systems. They may have experience with complications but no time to explain. A poor woman is expected to tolerate this because she has no choice. A middle-class family pays to escape it. A rich family pays to never even see it.

So the moral answer is not simple…!

It is not enough to say that expensive hospitals are exploiting people. Some are. But some are also filling the gaps that public systems have left open for decades: dignity, comfort, privacy, communication, and the feeling that the family is seen.

It is also not enough to say that expensive care is automatically better care. That is a ridiculous assumption. In medicine, price is not a biomarker of quality. A large bill does not prove better outcomes. A famous name does not prove rational care. A suite room does not prove safer labour. A polished discharge summary does not prove better newborn follow-up.

Subscribe now

This is where academia should help. But does it?

I searched for clear academic answers. I looked for literature that could help me respond to my friend in a responsible way. I found many articles on maternal health, institutional delivery, caesarean rates, public-private differences, respectful maternity care, out-of-pocket expenditure, and health system inequity. But I did not find the kind of answer an ordinary Indian family actually needs.

Not a theoretical answer. Not a moral answer. Not a slogan.

A practical answer.

“If I spend five lakhs for childbirth in a famous hospital, what measurable advantage am I buying over a competent public or teaching hospital?”

That answer is still missing.

And this absence itself says something about us.

We produce academic work. We conduct meetings. We discuss audits. We speak about quality. We publish papers. Many teaching hospitals publish more than several corporate hospitals. Yet when a family asks a basic question about the value of an expensive birth, our system becomes strangely vague. We have data in fragments, but not in a form that helps citizens compare care. We have claims, but not enough public accountability. We have reputation, but not enough transparency…!

Share

In a serious health system, hospitals offering childbirth care should be able to publish meaningful outcome indicators in a way families can understand. Not as marketing brochures. Not as selective success stories. Not as photographs of smiling mothers and babies. Real numbers. Caesarean rates. NICU admission rates. Maternal complications. Neonatal outcomes. Breastfeeding support. Readmissions. Infection rates. Patient-reported dignity. And these numbers should be risk-adjusted, because a referral hospital seeing high-risk pregnancies cannot be compared crudely with a centre handling mostly low-risk deliveries.

This is not to shame hospitals. It is to make trust measurable…!

Because trust without data becomes branding….

My friend was not asking whether his hospital was good. His wife and child were safe, and that matters more than any argument. He was asking something more difficult: whether the extra money bought better medicine, or mainly a better experience around medicine.

I still do not know.

Maybe he bought faster access. Maybe he bought a lower threshold for escalation. Maybe he bought the assurance that a senior doctor would appear quickly. Maybe he bought privacy and dignity. Maybe he bought protection from the uncertainty and crowding of an overstretched public system. Maybe he bought an atmosphere where his family felt looked after. Maybe he bought hospitality. Maybe he bought healthcare. Most likely, he bought a mixture of both.

But I cannot honestly tell him how much of each.

And that is where my reflection ends, not with an answer, but with a discomfort I cannot shake off. For all the academia I have around me, for all the meetings, audits, reviews, journal articles, mortality discussions, and institutional vocabulary of quality, I still do not have a credible answer to one intelligent father’s question: when an Indian family pays five lakhs for a birth, what exactly has improved?

Until we can answer that clearly, families will continue to judge hospitals the way they judge hotels: by reputation, comfort, polish, and fear. And childbirth, one of the most human moments in medicine, will remain trapped between safety and status, between care and comfort, between hospital service and hospitality.

Lets uncover a few points in detail, with practical advice, context on why each matters, resources, and some insights from the academic publishing landscape.

The goal: empower you to publish ethically, legitimately, and at zero or minimal cost while targeting Scopus-indexed venues.

Subscribe now

Why Free Publishing Matters

• Equity: High APCs (often $1,000–$5,000+) exclude talented researchers from the Global South or underfunded institutions.

• Quality isn’t tied to cost: Many diamond OA journals are rigorously peer-reviewed and indexed in Scopus, DOAJ, and more.

• Sustainability: Diamond models are often community- or institution-supported, reducing reliance on commercial profits.

Always verify everything yourself—policies change, and predatory journals prey on desperation.

The simple Tips: Expanded and Actionable

1. Start with this list
Begin your search at journalsearches.com/free-publishing-journals.php. It curates Scopus/SCI-indexed free journals across dozens of subjects (e.g., Engineering, Medicine, Social Sciences). The site lists titles, publishers (many SpringerOpen, Elsevier, SAGE), ISSN, and review processes. It’s an excellent curated starting point with direct links.

Subscribe now

2. Use DOAJ :

https://doaj.org/

The Directory of Open Access Journals is the gold standard. Filter for journals with no APC and no other charges—it shows over 14,000 such journals. Apply additional filters for subject, license (e.g., CC-BY), and indexing. DOAJ vets for quality, making it a trustworthy resource for diamond OA.

Share

3. Use SCImago Journal Rank
On scimagojr.com, enable the “Only Diamond Open Access Journals” filter. Sort by subject, country, or quartile (Q1–Q4). This helps identify high-impact, free journals with transparent metrics based on Scopus data. Great for targeting reputable outlets.

4. Check curated no-APC journal databases
Platforms like NoAPC list thousands of free-to-publish OA journals indexed in Scopus and Web of Science (over 2,500 in Scopus alone). Use these specialized databases for quick, targeted discovery.

Subscribe now

5. Search by field
Use precise Google or database queries: “no APC journals in microbiology,” “free publication journals in education,” or “diamond open access journals in public health.” Combine with “Scopus indexed” for relevance. This surfaces society journals or regional platforms tailored to your discipline.

6. Look at society and university-supported journals
Many academic societies, universities, and research institutes sponsor journals without author fees. These are often high-quality, field-specific, and diamond OA. Examples include those hosted on university presses or society sites. Search your field’s main associations.

Share Vikram Sakaleshpur Kumar

7. Check SciELO
SciELO (Scientific Electronic Library Online) hosts reputable journals, particularly from Latin America, with strong diamond OA coverage. Excellent for social sciences, humanities, and regional research with global visibility.

Subscribe now

8. Check Redalyc
Redalyc focuses on Latin American journals in social sciences and humanities. Many are no-fee and provide strong indexing and visibility.

9. Explore publisher waiver policies
Even journals with standard APCs often offer full/partial waivers. Check the journal/publisher site (e.g., Springer, Elsevier, Wiley, SAGE) for eligibility. Many have dedicated waiver pages.

Subscribe now

10. Look for Research4Life or similar eligibility
Programs like Research4Life provide APC waivers/discounts for authors from eligible low- and middle-income countries. Check if your institution/country qualifies—it’s a game-changer for many researchers.

Subscribe now

11. Check special publisher initiatives
Some publishers offer country-specific, institution-specific, or journal-specific waivers. Explore these proactively.

12. Ask the editor directly
If a journal normally charges APCs, email the editor explaining your funding situation. Many consider case-by-case waivers, especially for strong manuscripts. Polite, well-timed inquiries work more often than expected.

13. Use your institution’s publishing support
Contact your university library or research office. Many have transformative agreements (Read & Publish deals) that cover APCs for affiliated authors or provide internal funds.

14. Collaborate strategically
Co-author with researchers from institutions with publishing agreements. This can legitimately cover costs while enriching the work through diverse perspectives. Ensure authorship follows ethical guidelines (e.g., ICMJE).

15. Watch for special issues and temporary fee waivers
Journals often waive fees for special/anniversary issues or promotional periods. Monitor calls for papers in your field.

Subscribe now

16. A list that many have found useful…. https://docs.google.com/spreadsheets/d/e/2PACX-1vSCfhHbtTQhnjdd0LmUwUZ6lMGSbCDQC_cx2qFjmw2v983pccv9k3yHJooQEcZNIn4BTOLV-9mGBpRM/pubhtml https://www.researchgate.net/deref/https%3A%2F%2Fcutt.ly%2FxeYtYkx?_tp=eyJjb250ZXh0Ijp7ImZpcnN0UGFnZSI6InF1ZXN0aW9uIiwicGFnZSI6InF1ZXN0aW9uIiwicHJldmlvdXNQYWdlIjoicXVlc3Rpb24iLCJwb3NpdGlvbiI6InBhZ2VDb250ZW50In19

17. Use Think. Check. Submit.
Before submitting, visit thinkchecksubmit.org to verify the journal is legitimate (not predatory). Check editorial board, peer review transparency, and indexing. Essential due diligence.

Subscribe now

18. Check indexing status
Confirm the journal is currently indexed in Scopus (or your target databases) and has a clear editorial board + peer review process. Indexing can change; always double-check on the official Scopus source list or the journal site.

19. Look beyond big commercial publishers
Excellent free journals come from smaller non-profits, scholarly associations, and regional platforms. Don’t overlook them—they often provide better fit and service for niche topics.

20. Search for diamond open access by subject
Targeted searches like “diamond open access journals in Medicine,” “pediatrics,” or “surgery” yield strong results. Combine with DOAJ/SCImago filters.

21. Always verify before submission
Policies evolve. A journal free last year might not be now. Always check the official “Instructions for Authors” or APC page on the journal website. Confirm three things: (1) Is it really free? (2) Is it indexed where you need? (3) Is it legitimate?

Subscribe now

Final Thoughts

Free publishing in reputable Scopus-indexed journals is achievable with smart searching, verification, and persistence. Resources like DOAJ, SCImago, NoAPC, and journalsearches.com democratize access. Focus on quality research and legitimate outlets—your work deserves visibility without financial barriers.

Start with one or two tips today (e.g., DOAJ search + Think. Check. Submit.). Share your successes and challenges in the comments. For more on research skills, thesis writing, or ethical AI, follow and subscribe

This post is for informational purposes. Always consult official sources and your institution for personalized advice.

Happy publishing! 🚀

Further reading:

1. https://www.researchgate.net/post/Where_can_I_find_legitimate_2025_free-to-publish_journals_no_APC_no_submission_fees_no_hidden_charges_indexed_in_Scopus_for_Biomedical_Health_Sc

2. https://jefflibraries.libguides.com/OpenAccess/NO_APC\

3. https://www.linkedin.com/posts/freejournals-noapc-openaccess-share-7429071413057134592-xiXW/

4. https://abcdindex.com/listings2.php?ProType=journal&keyword=&filter_indexing=3&filter_country=&filter_paid_free=free&filter_rating=&filter_review_staus=&search_data=Search

5. https://www.springernature.com/gp/open-science/journals-books/journals

6. https://en.wikipedia.org/wiki/Diamond_open_access

7. https://www.openaccess.nl/en/diamond-open-access/resources

8. https://lib.asu.edu/news/research/diamond-no-cost-open-access

   Thanks for reading! This post is public so feel free to share it.

   Share

Variables are the building blocks of medical research. They define what we are studying, what we are measuring, what may be influencing the result, and what outcome finally matters to patients. Without clear variables, research becomes vague, analysis becomes weak, and conclusions become unsafe.

In healthcare, variables are not just statistical labels. Age, severity of illness, nutritional status, socioeconomic background, timing of treatment, laboratory values, and clinical outcomes all shape real decisions. They decide whether a finding applies to a patient, whether an intervention is useful, whether a risk factor is truly causal, and whether a policy will work in the real world.

In health delivery, good variable thinking helps us move from data to action. It tells us whom to screen, whom to treat, what to monitor, what to control for, and where systems are failing.

A doctor, researcher, teacher, or policymaker who understands variables is less likely to be misled by associations and more likely to make decisions that are fair, precise, and clinically meaningful.

Subscribe now

1. The simplest way to think about variables

The medical research question always has this skeleton

Exposure / Intervention → Outcome

Example:

Early antibiotics in suspected neonatal sepsis → Mortality

But real patients are not clean equations. They bring age, gestation, birth weight, severity, culture status, referral delay, maternal risk factors, prior antibiotics, hospital capacity, and clinician judgment into the picture.

That is why variables matter. They decide whether your conclusion is clinically useful or dangerously misleading….!!

STROBE stresses that observational studies must clearly describe exposures, outcomes, confounders, effect modifiers, and statistical methods because poor reporting makes clinical interpretation weak. CONSORT similarly exists because randomized trials need transparent reporting before readers can judge whether the result is believable.

Subscribe now

2. Independent variable

What it is

The independent variable is the exposure, risk factor, treatment, test, or intervention whose effect you are studying.

It answers:

“What are we testing as the possible cause or predictor?”

How it affects decision-making

If you define the independent variable badly, the whole study collapses.

Example:

“Antibiotic use” is vague.

Better:

“Receipt of intravenous antibiotics within 1 hour of sepsis recognition.”

Why? Because clinical decisions depend on precision. A policy-maker cannot act on “antibiotics help.” They need to know which antibiotic, at what timing, for which patient, at what severity, and with what trade-off.

Share

3. Dependent variable

What it is

The dependent variable is the outcome.

It answers:

“What changed?”

How it affects decision-making

A weak outcome produces weak medicine.

Example:

A study shows that a drug reduces fever duration by 6 hours. But it does not reduce hospitalization, complications, antibiotic use, cost, or mortality. That result may be statistically significant, but clinically small.

This is why evidence-based decision-making does not stop at p-values. GRADE asks decision-makers to judge certainty of evidence, balance of benefits and harms, patient values, resources, equity, acceptability, and feasibility before making recommendations.

Share Vikram Sakaleshpur Kumar

4. Confounding variable

What it is

A confounder is a third variable that is related to both the exposure and the outcome, creating a false or exaggerated association.

It answers:

“Is something else actually explaining this result?”

Classic medical example

Research question:

Does formula feeding increase infant infections?

Possible confounder:

Socioeconomic status.

Why?

Socioeconomic status may affect feeding choice and infection risk through sanitation, crowding, maternal education, healthcare access, and nutrition.

So formula feeding may look harmful partly because the groups were different before feeding ever happened.

Confounding is one of the biggest weaknesses of observational studies because exposures are not randomly assigned. Observational studies can be rigorous, but lack of random assignment leaves them vulnerable to confounding and bias.

How it affects reasoning

Confounding is where bad researchers fool themselves.

They see an association and call it causation.

Wrong.

First principle:

An association is not a causal claim until alternative explanations are handled.

If you do not adjust for socioeconomic factors, you may blame the wrong thing.

Subscribe now

5. Mediating variable

What it is

A mediator explains the pathway between exposure and outcome.

It answers: “How does the effect happen?”

Medical example

Question:

Does maternal education reduce under-5 mortality?

Here, care-seeking behavior, immunization, and nutrition may be mediators….!

Mediation asks whether the exposure affects the outcome through an intermediate mechanism. Epidemiology literature separates mediation from effect modification because mediation explains pathway, while effect modification asks whether the size or direction of effect differs across groups.

How it affects decision-making

Mediators help you design better interventions.

Example:

If maternal education reduces mortality mainly through earlier care-seeking, then the practical intervention is not just “educate mothers.” It is also transport, referral linkage, danger sign recognition, and reducing health system delay.

Mediation tells you where to intervene in the chain…!

Subscribe now

6. Moderating variable / effect modifier

What it is

A moderator or effect modifier changes the strength or direction of the relationship between exposure and outcome.

It answers:

“For whom does this work better, worse, or differently?”

Effect modification occurs when the effect of an exposure on an outcome depends on another variable.

Medical example

Question:

Does zinc reduce duration of diarrhea?

Possible moderator:

Baseline nutritional status.

How it affects decision-making

Moderators prevent lazy universal conclusions.

Bad conclusion:

“This intervention works.”

Better conclusion:

“This intervention works best in children under 2 years with moderate severity and early presentation, but not clearly in late referrals with hypoxia.”

That is the difference between research and clinical intelligence.

The mediator sits in the pathway.
The moderator stands outside the pathway and changes the effect.

Subscribe now

7. Control variable

What it is

A control variable is a variable deliberately held constant, restricted, matched, stratified, or statistically adjusted.

It answers:

“What are we keeping stable so the comparison is fair?”

How it affects decision-making

Control variables protect fairness.

Example:

If one group has more preterm neonates and another has mostly term neonates, mortality comparison is useless unless gestational age is controlled.

Randomization is powerful because it aims to balance both known and unknown prognostic factors between groups, but reporting still matters. CONSORT was developed to improve RCT reporting so readers can appraise and interpret trial results properly.

Share

8. Extraneous variable

What it is

An extraneous variable is unwanted noise that may affect the outcome but is not central to the research question.

It answers:

“What else could disturb the measurement or result?”

How it affects decision-making

Extraneous variables increase noise AND noise reduces your ability to detect truth.

Example:

If bilirubin samples are collected at different postnatal ages without standardization, the result may reflect timing rather than disease severity.

Extraneous variables do not always create systematic bias, but they reduce precision and reproducibility.

9. Categorical variable

What it is

A categorical variable places patients into groups.

It answers:

“Which group does this patient belong to?”

Types

Nominal = categories without order

Example: blood group, sex, organism type, mode of delivery

Ordinal = categories with order

Example: mild/moderate/severe dehydration

low/medium/high risk

grade 1/2/3/4 murmur

How it affects decision-making

Categorization is clinically useful but statistically costly.

Example:

Hemoglobin as a continuous variable gives more information. But hemoglobin grouped as mild/moderate/severe anemia is easier for treatment decisions.

The danger is crude grouping.

A child with Hb 10.9 g/dL and another with Hb 7.1 g/dL should not be treated as if they are clinically similar just because both are “anemic.”

Subscribe now

10. Continuous variable

What it is

A continuous variable is measured on a numeric scale.

It answers:

“How much?”

How it affects decision-making

Continuous variables preserve information.

Example:

Oxygen saturation of 89%, 93%, and 97% are not the same.

But if you convert everything into “hypoxia: yes/no,” you lose clinical detail….!

Still, thresholds are necessary for decisions.

The threshold must be clinically justified, not invented after seeing the data.

Subscribe now

11. Demographic variable

What it is

A demographic variable describes the patient or population.

It answers:

“Who are we studying?”

How it affects decision-making

Demographics decide whether evidence applies to your patients.

A pneumonia prediction rule developed in a high-resource urban hospital may fail in a rural Indian hospital where malnutrition, delayed referral, prior antibiotic use, and oxygen access are different.

This is called indirectness in evidence assessment. GRADE considers indirectness when judging certainty of evidence, because evidence from a different population, intervention, comparator, or outcome may not translate cleanly to the local decision.

Subscribe now

One variable, many roles

Take age.

The same variable can play different roles in research an analysis

So never ask only:

“What type of variable is age?”

Ask:

“What role is age playing in this research question?”

Share

12. The most important distinction: confounder Vs mediator Vs moderator

This is where most PG students and reviewers get confused.

Example question

Does maternal education reduce child mortality?

Confounder

Household wealth ─────► Maternal education
       │
       └──────────────► Child mortality

Household wealth may influence both education and mortality. If not adjusted, education may look more powerful than it really is.

Mediator

Maternal education ─────► Better care-seeking ─────► Lower mortality

Care-seeking explains how education may reduce mortality.

Moderator

Maternal education ─────► Lower mortality
             ▲
             │
     Rural versus urban setting changes effect strength

Education may reduce mortality more strongly where health services are reachable.

Subscribe now

The three-question test

Share

13. How variables change healthcare decision-making

A. Clinical decision-making

Poor variable reasoning leads to wrong bedside decisions.

Example:

Study says:

“High CRP predicts antibiotic need.”

Problem:

CRP may simply reflect severity, duration of fever, referral delay, or previous antibiotics.

Better question:

“After adjusting for age, fever duration, respiratory rate, work of breathing, oxygen saturation, and danger signs, does CRP independently improve antibiotic decision-making?”

That is a useful clinical question.

B. Guideline-making

Guidelines should not ask only:

“Is the p-value significant?”

They should ask:

Is the exposure well-defined?
Is the outcome patient-important?
Were confounders handled?
Is there effect modification?
Is the result precise?
Is the population similar to ours?
Are harms measured?
Is it feasible?
Is it affordable?
Is it equitable?

GRADE Evidence-to-Decision frameworks were developed to make these judgments explicit and transparent for clinical, coverage, health-system, and public-health decisions.

C. Policy decision-making

A policy based on poor variable logic wastes money.

Example:

If neonatal deaths are associated with “home delivery,” the lazy conclusion is:

“Move all births to institutions.”

But the real drivers may be transport delay, poor referral, maternal anemia, prematurity, sepsis recognition, lack of SNCU beds, or poor postnatal follow-up.

The correct policy depends on identifying the causal structure, not just the visible association.

Subscribe now

14. Medical examples across common study designs

Randomized controlled trial

Question:

Does zinc reduce duration of acute diarrhea in children?

Decision implication:

If zinc reduces duration mainly in malnourished children, then universal use and targeted use may have different value.

Cohort study

Question:

Does prior antibiotic exposure increase risk of multidrug-resistant neonatal sepsis?

Decision implication:

If prior antibiotic exposure is a marker of severity rather than the true driver, antibiotic stewardship policy must target both prescribing and infection-control pathways.

Cross-sectional study

Question:

What factors are associated with anemia among school children?

Decision implication:

A cross-sectional study can show association, not clean temporality. It can guide screening and hypothesis generation, but cannot confidently prove cause.

Diagnostic study

Question:

Does procalcitonin identify culture-positive sepsis in children?

Decision implication:

A test may look good overall but fail in pretreated children, neonates, or low-prevalence settings. That affects real-world use.

Subscribe now

In Summary

Share

15. The common mistakes

Mistake 1: Treating all variables as equal

They are not equal.

-Some are causes.
-Some are outcomes.
-Some are pathways.
-Some are distortions.
-Some are measurement formats.

Mistake 2: Adjusting for mediators as if they are confounders

This can erase the true effect.

Example:

If breastfeeding reduces diarrhea through improved immunity, and you adjust for immune markers, you may remove part of the breastfeeding effect.

Mistake 3: Ignoring effect modification

A treatment can work in one subgroup and not in another.

Overall average effects can hide clinically important subgroup differences.

Mistake 4: Turning continuous variables into categories too early

You lose information.

Example:

Age in months is richer than “under 5.”

Mistake 5: Choosing outcomes that do not matter to patients

A biomarker improvement is not enough unless it translates into recovery, survival, function, reduced harm, or better care.

Share

The reviewer’s variable checklist

Share

A Final clean table for students

Share

In Conclusion….

In medical research, variables are not just columns in Excel. They are the logic of the study.

If you misclassify variables, you misread causality.
If you misread causality, you make weak clinical decisions.
If you make weak clinical decisions, patients pay for your bad reasoning.

The disciplined approach is simple:

Define the exposure.
Define the outcome.
Draw the causal pathway.
Identify confounders.
Identify mediators.
Identify moderators.
Choose the right measurement scale.
Analyze accordingly.
Translate only after checking certainty, applicability, harms, feasibility, and equity.

That is how variables move from research methods into real healthcare decision-making…..

Other publications :

Most families have one person who keeps saying the same thing.

Eat less sugar. Walk more. Finish your dinner early. Don’t keep eating after dinner. Don’t wait for diabetes to teach you biology.

And most families also have the same reply.

“Nothing will happen.”

“Our grandfather ate ghee every day.”

“I hardly eat anything.”

“Life is short, let us eat.”

The argument usually dies there. Not because the science is weak. Because the science is explained badly. We say calories, obesity, BMI, insulin, cholesterol. People hear judgement. They hear a doctor trying to remove joy from life.

So let us leave all that aside for a moment.

Let us talk about a house/home

Subscribe now

The house/home that does not collapse in one day

Imagine an old family house. Kitchen, storage room, wiring, water pipes, a courtyard, rooms where people live. When the house is new, it tolerates abuse. Extra sacks of rice pushed into one corner. Old vessels in another. Broken furniture in the attic. Nobody notices much.

Then the years pass.

The storage room becomes full. But the family keeps bringing in more things. The room can no longer take it. Things spill into the kitchen. Then into the bedroom. Then near the electrical wiring. Then near the pipes.

At first, nothing dramatic happens. That is the dangerous part.

The house adjusts. People step around the boxes. Someone says, “It has always been like this.” Someone says, “We will clean it after the festival.”

Then dampness appears. Then mold. Then insects. Then the wiring begins to fail. And now the house does not look like a storage problem anymore. It looks like a whole-house disease.

The human body works the same way.

Food is fuel. Movement is demand. Fat tissue is the main storage room. Muscle is the fuel-burning engine. The liver is the processing plant. The pancreas — a small organ sitting quietly behind your stomach — is the control office that decides how much fuel to release and when.

When we eat roughly what the body needs, the system hums along. Some fuel is used immediately. Some is stored safely. Some is released later. This is normal. The body was built for exactly this.

The problem begins when supply keeps exceeding demand.

This is the modern problem.

Subscribe now

Why your grandfather’s butter/ghee is not the argument you think it is

Most adults today are not doing the physical work their grandparents did. Not walking long distances daily. Not lifting, bending, climbing, farming, cycling. But food supply has not reduced. In many homes it has increased…!

Breakfast is heavy. Tea has snacks. Lunch is polished rice or refined wheat in large portions. Evening has more tea and more snacks. Dinner is late. Sweets come because someone visited. Fried food comes because someone ordered. Packaged food comes because children asked. Festival food comes even when there is no festival.

The body receives fuel as if we are labourers.

But many of us live like office machines.

That mismatch is the beginning.

For some time, the body manages. It is genuinely intelligent. It stores excess energy in fat tissue. Fat tissue is not useless or evil — it is a living organ that stores energy, protects against sudden fuel overload, and talks to the brain and hormones. The body is trying to help you.

But every storage room has a limit.

Once safe storage becomes overloaded, the body starts pushing fat into places where fat does not belong. The liver becomes fatty — this is not a dramatic event, it happens silently over years. Muscle begins to resist insulin.

Here it is worth pausing. Insulin is the key that unlocks cells so sugar from your blood can enter and be used as fuel. When cells stop responding to this key — a condition called insulin resistance — sugar has nowhere to go. It piles up in the blood. The pancreas panics and makes more insulin to compensate. It works harder and harder. For years, the blood report may look almost normal, because the pancreas is still managing. Until one day it cannot.

Meanwhile, blood vessels are becoming inflamed. Triglycerides — the fat particles in blood — are rising. Blood pressure is climbing. The waist is growing. Tiredness is becoming the new normal.

Then one day the report comes.

Fatty liver. Prediabetes. High triglycerides. Borderline sugar. High BP.

And the same person says, “But I was fine until last year.”

No. The body was not fine. The body was compensating.

The blood report is only the notice pasted on the door. The damage started when the storage room began overflowing — quietly, invisibly, years earlier.

Subscribe now

But here is the part families are not ready for

This is no longer only an adult story.

We are watching the adult metabolic story arrive in childhood. And the tragedy is that most families are not seeing it, because they are looking for the wrong signs.

A child’s body is growing, learning, adapting, repairing. That gives childhood enormous resilience. But resilience is not immunity. A growing body can also be overwhelmed when the food environment is continuously working against it.

Today, many children are surrounded by a food system designed — and this word is chosen carefully, it is not an accident — designed to make them eat frequently, quickly, and without thought. Sweet drinks. Cream biscuits. Chips. Chocolate-flavoured “health” drinks. Fast food. Birthday food every week. Screen food. Reward food. Food after every small achievement. Food after every small sadness.

The child’s body receives the same signal again and again: Store. Store. Store.

And at the same time, many children move less than any previous generation. School is sitting. Tuition is sitting. Screen time is sitting. Transport replaces walking. Playgrounds disappear. Apartments replace open spaces. Parents fear traffic, injury, exams, heat, strangers, wasted time. So children are kept safe indoors — and slowly made metabolically unsafe.

Metabolism, in plain terms, is just the body’s system for converting food into energy and deciding what to do with the excess. When that system is working well, the child grows strong, sleeps deeply, plays without tiring, and feels hungry at the right times. When it is disrupted by continuous overload and under-movement, the same damage that takes decades in adults can begin arriving in a child within years.

A child with excess weight may already have a fatty liver. A teenager may already have insulin resistance. A school-going child may already have high triglycerides, early high blood pressure, or poor sleep quality. A young adolescent may show signs — a darkening of skin around the neck, irregular periods in girls, constant fatigue, snoring — that a doctor recognises as early metabolic warning signals.

These are not “adult diseases” arriving unexpectedly early.

They are the same disease, in the same body, starting earlier because the overload started earlier.

This should disturb us deeply.

A child is supposed to be building muscle, bone, confidence, sleep rhythm, and a healthy relationship with food. Instead, many children are being trained into craving cycles, sitting cycles, sugar cycles, screen cycles, and reward-eating cycles.

We are not only feeding children.

We are programming their metabolism.

Subscribe now

Why families do not change even when they understand

Here is the part that most health writing skips.

Families are not stupid. They know fried food is not ideal. They know too much sugar is not good. They know children should move more. They have been told this hundreds of times.

And yet, the evening snack still appears. The sweet drink is still poured. The birthday celebration still involves cake plus chips plus sweet juice plus return gifts of more chocolate. The child still sits at the screen. The adults still sit at the screen too, right next to the child, asking why the child is sitting at the screen.

This is not weakness or ignorance. This is something more important: food in the Indian family is not just nutrition. It is love. It is welcome. It is apology. It is celebration. It is the answer to almost every emotion. A child cries — give a biscuit. A guest arrives — produce snacks immediately. Someone passes an exam — order sweets. Someone fails — order food to comfort.

Food has become our emotional vocabulary.

And changing that feels, to most families, not like a health decision but like an attack on how they express love.

This is the real storage room that is overflowing. Not just the pantry full of packaged snacks. But the family habit of reaching for food every time an emotion needs addressing. The belief — deep, unquestioned, passed across generations — that feeding is the most reliable form of care.

The person who says “eat less sugar” in that family is not heard as a health advisor. They are heard as someone who wants to remove warmth from the home.

This is why scolding does not work. This is why lectures do not work. This is why even fear — the diabetes diagnosis, the fatty liver report — works only briefly before the old patterns return. Because the food is not filling the stomach. The food is filling something else.

To change what the family eats, the family has to find other ways to say the things food was saying.

Sit with a child instead of giving a biscuit. Call instead of sending a box of sweets. Celebrate with an experience instead of always with sugar. Welcome guests with water and fruit without apology. Replace reward food with reward time, reward words, reward touch.

This is harder than buying low-calorie biscuits.

But it is the only change that lasts.

Subscribe now

What honest families actually need to do

Not with drama. Not with crash diets. Not with body-shaming children. Not with fear. With ordinary discipline, applied quietly over ordinary days.

The house. Remove constant food visibility. If snacks are always on the table, people will eat them. That is not weakness — that is how human biology works. Keep fruit visible. Keep fried snacks out of daily reach. Do not convert the dining table into a shop.

Meal timing. The body does not need to be fed continuously from morning to night. Children need nourishment, not constant packaged snacks between meals. Adults do not need endless grazing. Tea does not need a snack every time. Night does not need a second dinner.

Muscle after meals. Ten to fifteen minutes of walking after lunch or dinner is not a small thing. It tells the body that fuel has arrived and the engines are running. Muscle is the body’s most powerful tool for clearing sugar from blood after a meal. An unused muscle is like a generator kept in the corner — large, expensive, switched off.

Play is not optional for children. Play is metabolic education. Running, jumping, climbing, cycling, swimming, dancing, rough-and-tumble games — these teach the body how to use fuel. They build muscle, sleep pressure, appetite regulation, coordination, and confidence. A child who never moves is not simply missing play. That child is missing a biological curriculum that no tuition class can replace.

Liquid sugar is the easiest fix. Sweet tea, soft drinks, packaged juices, and chocolate-flavoured health drinks quietly and quickly overload the system. Liquid calories enter the blood faster than solid food, leaving the body less time to respond. For families with children, reducing liquid sugar is one of the highest-impact changes with the least disruption.

Make the plate honest. Half the plate should not be polished rice or refined wheat at every meal. Add vegetables, dal, eggs, curd, pulses, fish, chicken, sprouts, or whatever fits the home, culture, and budget. Protein and fibre slow the flood of sugar into blood. They also keep hunger quieter for longer — which means less reaching for snacks by evening.

Sleep is not laziness. A tired brain is a brain that craves quick fuel — sweet, salty, fried. Poor sleep worsens hunger signals, insulin resistance, mood, and decision-making in both adults and children. Children who sleep late are not becoming modern. They are becoming metabolically fragile.

Do blood tests before fear arrives. Waist circumference, blood pressure, fasting sugar, HbA1c, lipid profile, and liver enzymes are maintenance checks, not disaster reports. For children, testing should be guided by the paediatrician — especially when there is significant weight gain, strong family history, or early warning signs. The goal is to catch the storage room overflowing before it reaches the wiring.

Subscribe now

The mirror is not enough

Weight alone can mislead.

Some adults look thin but have fatty liver and high blood sugar. Some children look only “slightly healthy” but already have early metabolic risk. Some teenagers look normal but live on sugar, screens, poor sleep, and near-zero muscle use. Some people say, “I am not obese” — but their waist, their liver, their triglycerides, and their insulin response are telling a different story.

The weighing scale is not enough either.

The better question is not “How much do I weigh?”

The better question is: “Is my body still storing excess fuel safely — or has it started pushing that excess into organs where it does not belong?”

For children, the better question is not “Is my child thin or fat?”

The better question is: “Is my child becoming stronger, fitter, more active, better rested, and more regulated — or is the opposite happening?”

Subscribe now

The body gives warnings long before collapse

The body is generous. It gives warnings long before it fails.

A growing waist is a warning. Fatty liver is a warning. Sleepiness after meals, rising triglycerides, borderline blood pressure, prediabetes — all warnings. Poor stamina in a child. A child who cannot play for twenty minutes without exhaustion. A teenager with constant fatigue, or irregular periods, or snoring, or a darkening of skin on the back of the neck — that last sign, which a doctor calls acanthosis nigricans, is the skin’s way of showing that insulin is working too hard.

These are not random symptoms. They are the house speaking….!

The difference between a family that hears these warnings and one that does not is rarely knowledge. Most families have the knowledge. The difference is whether the family has been honest — genuinely honest — about what food is doing in their home. Not just to the body. To the relationships. To the emotions. To the children who are watching every adult’s habits and learning what normal looks like.

Children do not build their food habits from lectures.

They absorb environments.

The home is the first and most powerful food environment a child will ever have.

Subscribe now

What should change after reading this

Not everything. Not dramatically. Not all at once…!

But something should shift.

The next time someone in the family says “eat less sugar”, hear it differently. They are not attacking your joy. They are saying the storage room is getting full.

The next time a child’s paediatrician mentions weight or blood markers, do not deflect. Ask what the trajectory is. Ask what the family can change. Ask to check again in six months.

The next time the evening snack appears without thought, pause. Not to feel guilty. Just to ask: is this hunger, or is this habit?

The next time a child comes home tired and sad and the first instinct is food — notice it. Then sit with the child instead. Ask what happened. The child does not need a biscuit. The child needs you.

And the next time someone says “our grandfather ate ghee every day” — agree with them. He did. He also walked four kilometres to the field at dawn. He also slept by nine. He also ate no packaged food in his life. He also had muscles that worked all day, every day, clearing fuel the way muscles are built to do.

His body received fuel into a system with demand…!

Our bodies receive fuel into a system sitting in a chair.

That is the entire story.

Subscribe now

The body is not a dustbin.

It is a home.

Homes do not collapse in one day. They collapse after years of ignored storage, ignored warnings, ignored signals, and the comfortable belief that the dampness in the corner is someone else’s problem to fix.

The storage room is full.

The question — the only question — is whether your family will start clearing it while the house is still standing.

There are weeks in clinical practice that stay with you long after the ward has moved on.

This was one of them.

Two families. Seven days. Both were from Shivamogga — not distant villages, not without resources, not without education. Both had arrived at our paediatric unit after long diagnostic journeys, with children who were getting worse and not better. The first family had a 14-year-old boy with a severe congenital neuromuscular disorder who had developed an overwhelming pneumonia. He was ventilated. His muscles were not recovering. The ventilator was doing work his body could no longer sustain. The second family had a five-month-old infant with persistent diarrhoea that had begun at two weeks of age and never resolved. We had done everything the setting allows — genetic panels sent to outside laboratories, biopsies, a full metabolic and infectious workup. No diagnosis emerged. The child had been admitted twice, perhaps three times. The illness continued its own silent course.

Both families had arrived at the same difficult threshold: treatment could still be continued, but it was increasingly unlikely to change the direction of what was happening.

After a few conversations, both families chose to take their child home.

Their children died there….!

Both families later told us that the final hours had been manageable. The grief was profound; it would be wrong to suggest otherwise. But what I noticed, in the weeks that followed, was that these parents did not come apart in the way I have sometimes seen. They had been prepared. And that distinction between; a family that was accompanied toward a death and one that was not, is what this piece is about.

Preparation does not prevent grief. It prevents bewilderment. Those are not the same thing, and the difference is entirely within our control.

Subscribe now

Preparation Is Not What We Think It Is

Doctors often speak of families being in denial. Of waiting for them to accept reality. I have become cautious about both phrases.

A parent may understand every word we say about prognosis and still ask for one more treatment. They may speak of hope in the morning and funeral arrangements by evening. This is not irrationality. It is how people survive inside devastating uncertainty. Grief is not a corridor with a fixed endpoint, and the clinician is not a guide who must deliver everyone to the same destination on schedule.

There is a concept called anticipatory grief — the mourning that begins the moment a family understands a child’s life is threatened, long before the death itself. Clinicians sometimes treat it as a tool: something to facilitate in order to produce a smoother bereavement. The evidence is more cautious. Palliative interventions addressing anticipatory grief show benefit in some contexts, but families experience it in very different ways and the literature remains limited.¹ What anticipatory grief names is something families may already be living. Our task, maybe is to accompany it — not to manage it toward a target outcome.

What we can usefully do is more modest, and perhaps more important: tell the truth carefully, repeat it when necessary, ask what has actually been understood, make room for silence, attend to symptoms, and help parents think through the decisions that may soon confront them.

Preparation is not persuading families to agree with us. It is helping them see the road ahead clearly enough to make choices that remain their own.

Subscribe now

These Conversations Are Usually Many Conversations

We sometimes imagine that breaking bad news is a single decisive event — the senior clinician sits down, the prognosis is explained, understanding follows, and a plan emerges. In practice, it is almost never like that.

With the first family, we spoke in stages. About the pneumonia. About the neuromuscular disease. About what the ventilator could and could not change. We also asked what they feared most and what they still hoped to protect. With the second family, uncertainty was the central fact — no diagnosis, no label, no clean trajectory. That made the conversations harder, not impossible. We could say what we knew: the child remained severely unwell, repeated admissions had not changed the course, and recovery was becoming increasingly remote. We could also say, plainly, what we did not know.

Prognostic humility is not evasiveness. Families deserve clarity. But clarity does not require false certainty — and a doctor who pretends otherwise is protecting themselves, not the family.

Structured frameworks such as SPIKES — developed originally in adult oncology and widely adapted since — help clinicians organise these conversations: an appropriate setting, understanding of what the family already knows, information delivered in manageable pieces, attention to emotion before strategy.² What no framework can substitute for is listening, or the kind of relationship that allows a parent to say what they are actually thinking. The discipline is to create that quality of attention regardless of how busy the ward is.

Research consistently shows that parents value clear information, honest communication, and preparation for what may happen near death.³·⁴ Good communication does not remove suffering. It ensures that avoidable confusion is not added to unavoidable grief.

If you leave a conversation feeling it has been completed, it probably has not. The next conversation is usually where real understanding begins.

Subscribe now

What Preparation Actually Means

Preparation is sometimes reduced to a signed form, or a documented decision not to escalate treatment. It is much more than that.

A family considering care at home needs to know what changes in breathing to expect, how discomfort will be managed, whom to call at night, and whether returning to hospital remains possible if things become unmanageable. They need medicines, written instructions, and a number that will actually be answered. They also need something less clinical: permission. Permission to grieve before the death. Permission to disagree with each other. Permission to change their minds.

And they need one thing that is easy to overlook: the clear reassurance that choosing comfort-focused care is not the same as choosing the death. That a decision to stop escalation does not make them responsible for what is happening to their child.

Shared decision-making must not become a transfer of moral responsibility. Clinicians carry the weight of their recommendations alongside the family — not hand it over to them.

Subscribe now

Home Is Not Always the Right Place

There is a quiet orthodoxy gathering around the idea of a good death at home, and I want to push back on it gently.

Some families prefer home. Others feel safer in hospital — close to nurses, oxygen, and immediate help. Preferences change as illness progresses. A child old enough to say something about it may feel differently from the adults around them.

The evidence is more cautious than the narrative. Systematic reviews of preferred place of death in children with life-limiting conditions show that many parents express a preference for home, but the literature is largely retrospective, concentrated in children with cancer, and drawn from settings with specialist palliative care — not from the majority of paediatric units in India where this decision falls entirely on the treating paediatrician.⁵ More recent integrative work finds that these decisions are shaped by the child’s condition, available services, the family’s capacity, and the needs of the household as a whole.⁶

The goal should not be to increase the number of children who die at home. It should be to increase the number of families who receive thoughtful care in a place they have been able to genuinely choose.

Sometimes that is home. Sometimes it is hospital. Neither is a moral achievement. Neither is a failure.

Subscribe now

The Work That Remains After Treatment Changes

When cure is no longer possible, it is easy — I recognise this in myself — to feel there is little left to offer. Families sense this withdrawal even when no one says so aloud.

A change in the goal of treatment is not the end of care. It may require more careful medicine: relief of pain and breathlessness, management of secretions, attention to feeding, support for siblings, clear plans for likely crises. In most district hospitals across India, there is no specialist palliative care team to refer to. The paediatrician, the ward nurses, the community health worker — they carry the coordination. The absence of a formal service does not dissolve the need. It sharpens the responsibility.

Evidence suggests that parents’ support needs in these situations are varied and frequently unmet — particularly around communication, out-of-hours availability, and psychological care.⁷ We should not be surprised by this. We should be honest about it.

We should also be careful how we read families after a death. A composed parent at a follow-up appointment does not mean grief has resolved. Grief unfolds over months and years. A calm final day does not guarantee an uncomplicated bereavement.

What we can claim, and what is enough: the family was informed, accompanied, and given real choices. That is worth doing, even when we cannot measure exactly what it produced.

Subscribe now

Preparation Without Promises

The two families from that week have stayed with me…!

Not because their choices offer a model. Not because I believe every family can be fully prepared for the death of a child, or that preparation guarantees anything about the grief that follows. They stayed with me because of something more specific — the visible difference between families who felt they had been told the truth gradually, honestly, with care, and families who had not. Both still grieve. But the quality of that grief appears different. Whether a controlled trial could capture that difference is a separate question. It is visible in the ward.

Parents do not need us to predict how they will mourn. They do not need to be hurried toward acceptance, or praised for being strong, or made to feel that remaining in hospital means they have somehow failed their child.

They need honesty that leaves room for uncertainty. They need information offered more than once. They need their questions taken seriously — including the ones that medicine cannot answer. And they need the reassurance that care continues even when cure cannot.

We cannot make such a loss orderly. We cannot promise that the grief will be gentler.

But we can ensure that when families reach the hardest part of this journey, they do not arrive there uninformed, unsupported, or alone.

That is less certain than I once hoped it would be. It is also more durable.

Subscribe now

Further Reading

1. Patinadan PV, Tan-Ho G, Choo PY, Ho AHY. Resolving anticipatory grief and enhancing dignity at the end-of life: A systematic review of palliative interventions. Death Stud. 2022;46(2):337-350. doi:10.1080/07481187.2020.1728426

2. Baile WF, Buckman R, Lenzi R, Glober G, Beale EA, Kudelka AP. SPIKES-A six-step protocol for delivering bad news: application to the patient with cancer. Oncologist. 2000;5(4):302-311. doi:10.1634/theoncologist.5-4-302

3. Mack JW, Hilden JM, Watterson J, et al. Parent and physician perspectives on quality of care at the end of life in children with cancer. J Clin Oncol. 2005;23(36):9155-9161. doi:10.1200/JCO.2005.04.010

4. Kassam A, Skiadaresis J, Alexander S, Wolfe J. Differences in end-of-life communication for children with advanced cancer who were referred to a palliative care team. Pediatr Blood Cancer. 2015;62(8):1409-1413. doi:10.1002/pbc.25530

5. Bluebond-Langner M, Beecham E, Candy B, Langner R, Jones L. Preferred place of death for children and young people with life-limiting and life-threatening conditions: a systematic review of the literature and recommendations for future inquiry and policy. Palliat Med. 2013;27(8):705-713. doi:10.1177/0269216313483186

6. Crowe A, Hurley F, Kiernan G, et al. Decision-making regarding place of end-of-life care for children with life-limiting and life-threatening conditions: a systematic integrative review. BMC Palliat Care. 2025;24(1):24. Published 2025 Jan 27. doi:10.1186/s12904-025-01661-1

7. Gill FJ, Hashem Z, Stegmann R, Aoun SM. The support needs of parent caregivers of children with a life-limiting illness and approaches used to meet their needs: A scoping review. Palliat Med. 2021;35(1):76-96. doi:10.1177/0269216320967593

Your PDFs Are Not Research. Your Questions Are….!

How to use NotebookLM to think faster, research deeper, and write a stronger methodology chapter

Most researchers do not suffer from lack of information; but suffer from an excess of badly arranged information.

The average postgraduate student now has PDFs, guidelines, protocols, old theses, systematic reviews, departmental templates, and half-remembered advice from five different seniors. Everything is available. Nothing is organised. The student reads more and understands less. The faculty member collects more articles and writes with less confidence. The methodology chapter becomes a ritual paragraph factory, full of phrases like “a cross-sectional study will be conducted” and “data will be collected using a predesigned proforma,” as if methodology is a clerical formality.

It is not.

Methodology is the spine of research. If it is weak, the study may still collect data, but it will not produce knowledge.

That is where NotebookLM can become useful, not as a magic research machine, but as a disciplined research assistant. Google describes NotebookLM as an AI research and thinking partner that works with the sources you upload.

It supports PDFs, Word files, Google Docs, slides, sheets, images, web URLs, pasted text, YouTube URLs, audio files, ePub files and more. Each source can contain up to 500,000 words or 200 MB, and a standard notebook can include up to 50 sources.

But the tool is only as good as the researcher using it.

NotebookLM will not rescue a badly framed question. It will not create rigor from laziness. It will not know the local realities of your hospital unless you give it those details. It will not replace PubMed, ethics review, statistical judgment, or methodological training. What it can do is make your own uploaded material more searchable, more conversational, and more usable.

That is already a major shift.

For years, researchers treated literature review as a storage problem. Download PDFs. Save them in folders. Highlight lines. Forget where the useful sentence was. Then write a protocol based on memory, pressure, and imitation.

NotebookLM changes the workflow. It allows the researcher to upload the actual source material and ask structured questions from it. Google states that NotebookLM answers are grounded in the uploaded sources and that responses include citations from those sources, allowing the user to check the original context.

That citation feature is the difference between casual AI use and research use.

A generic chatbot gives you fluency. NotebookLM gives you traceability. Fluency is dangerous in research. Traceability is useful.

Subscribe now

The first mistake is asking NotebookLM to “write my methodology”

That is the wrong prompt.

It is also the wrong mindset.

A methodology chapter is not a decorative section added after the objectives. It is an argument. It must show why this design fits this question, why this population is appropriate, why this sampling method is defensible, why this tool measures what it claims to measure, why this analysis plan answers the objective, and why the limitations do not destroy the study.

The prompts in your infographic series correctly move the researcher through stages: foundation, design, sampling, data collection, measurement, rigor, analysis, writing, critique, summary, and blueprint. The uploaded prompt document lists 13 structured research methodology prompts, beginning with “The Methodology Architect” and ending with “The One-Page Methodology Blueprint.”

That sequence matters.

A weak researcher asks: “Write methodology for my topic.”

A better researcher asks: “Given my literature, objectives, setting, design, population, tools, and constraints, what methodology is logically defensible?”

Subscribe now

Start with the Methodology Architect

Before choosing a design, upload your core material: the research problem, objectives, research questions, key literature, local setting, available resources, likely population, and any institutional constraints.

Then use the first prompt:

“Analyze the uploaded literature, research problem, objectives, research questions, and proposed study focus for [topic] and design a strong research methodology for my study. Recommend the most suitable methodological approach, justify why it fits the research problem, explain key methodological decisions, and show how the methodology supports the logic, rigor, and direction of the study.”

This is the correct beginning because it prevents premature method selection.

Many medical studies fail here. People choose cross-sectional design because it is easy. They choose convenience sampling because patients are available. They use a questionnaire because someone has a Google Form. Then they reverse-engineer the objective to fit the method.

That is not research design…!

NotebookLM should be used to expose that laziness early. Ask it to explain whether your proposed approach fits the research problem. Ask it what the design can and cannot answer. Ask it where the logic is weak. Ask it what assumptions you are making.

For example, if your study is on prevalence, a cross-sectional design may fit. If your study is on incidence, temporal sequence, intervention effect, causality, or outcome prediction, a simple cross-sectional design may be inadequate. NotebookLM will not make the final judgment for you, but it can force the issue into the open.

Subscribe now

Then use the Research Design Matchmaker

The second prompt is where many postgraduate protocols can be saved from mediocrity.

“Review the uploaded literature, research problem, objectives, and research questions for [topic] and determine the most appropriate research design for my study. Compare relevant options, explain their assumptions, strengths, limitations, and justify which design best addresses the research problem.”

This is useful because students often do not know that design is not a label. It is a contract between the question and the evidence.

A descriptive cross-sectional study answers a different kind of question from an analytical cross-sectional study. A cohort study answers something else. A case-control study has a different logic. A qualitative phenomenological study is not a weak quantitative study. A mixed-methods study is not “questionnaire plus interview” for decoration. It has a sequence, integration logic, and justification.

NotebookLM can compare these options using your uploaded sources. But it must be forced to compare, not merely comply.

The most important phrase in this prompt is “compare relevant options.”

Never ask AI to approve your first idea. Ask it to attack your first idea.

Share

Build methodological alignment before writing

The third prompt, “The Methodology Logic Builder,” is the most important one in the entire series.

Most methodology chapters fail because the parts do not speak to each other. The objective says one thing. The design does another. The sampling strategy selects a population that cannot answer the question. The tool measures a proxy. The analysis plan is vague. The ethics section is copied from another protocol. The limitations are hidden until the thesis defence.

Use NotebookLM to check whether the whole system is aligned.

Ask:

“Analyze the uploaded literature, research problem, objectives, research questions, and selected research design for [topic] and evaluate whether the methodology is logically aligned. Explain whether the design, data collection strategy, participant selection, and analytical approach work together coherently and identify methodological inconsistencies, weak links, or missing justifications.”

This is where NotebookLM becomes more than a summarizer.

Share Vikram Sakaleshpur Kumar

Sampling is not “whoever comes to OPD”

Sampling is where the action often begins….

Researchers say “random sampling” when they mean “whoever was available.” They say “consecutive sampling” without defining the recruitment period. They say “sample size was calculated” without explaining the assumptions. They include children, mothers, records, blood samples, or isolates without defining eligibility clearly. Then they expect the results to look credible.

They will not…

Use the Sampling Strategy Planner prompt. Ask NotebookLM to identify who should be included, how they should be selected, what sampling approach fits the design, and what biases will remain.

This matters especially in Indian medical college research. Our settings are messy. Patient flow is uneven. Records are incomplete. Follow-up is unpredictable. Referral bias is real. Private medical college data may not represent the district. Government hospital data may not represent private care. Tribal, rural, peri-urban, and tertiary-care populations are not interchangeable.

NotebookLM will not know these realities unless you upload or state them. So add a one-page “setting note” to every research notebook. Write plainly:

“This study will be conducted in a rural private medical college in Shivamogga. The hospital receives pediatric referrals from surrounding Malnad regions. Follow-up may be difficult because families travel from distant villages. Laboratory support is available for basic tests, but advanced testing may need outsourcing.”

That kind of context improves every answer.

Without context, NotebookLM gives generic methodology.

With context, it can help you build a defensible one.

Subscribe now

Data collection needs choreography

A method section should not merely say “data will be collected.”

Who will collect it? When? From whom? Using which tool? At what time point? From records or direct interview? In which language? With consent? Who will verify missing data? How will biological samples be labelled, transported, stored, and processed? What happens if the patient is discharged early? What happens if the lab report is delayed?

The Data Collection Strategist prompt forces this discipline.

It asks NotebookLM to recommend suitable methods such as interviews, surveys, observations, experiments, focus groups, document analysis, questionnaires, or secondary data, and to explain how they support rigor, validity, reliability, credibility, or trustworthiness.

This is where vague protocols become operational protocols.

For clinical research, this step is non-negotiable. A poor data collection plan creates missing data, inconsistent measurement, ethical confusion, and unusable analysis.

NotebookLM can help draft the flow. But the researcher must verify feasibility.

AI does not know that the intern who collects the data may be posted in casualty next week. AI does not know that the OPD register uses a different diagnosis label from the ward case sheet. AI does not know that the freezer has space today but not during inspection season.

You know that. Put it into the notebook….

Subscribe now

Measurement is where false confidence enters

The Instrument and Measurement Builder prompt is essential because researchers often confuse “having a tool” with “having a valid tool.”

A questionnaire is not valid because it looks neat. A checklist is not reliable because it has boxes. A scale is not appropriate because another paper used it. A translated tool is not automatically valid in Kannada because Google Translate produced readable words. A retrospective case sheet is not a reliable source just because it exists.

Ask NotebookLM to identify the best instruments, justify why they fit, and point out weaknesses.

For medical projects, this means asking hard questions:

• Does the tool measure the actual construct?

• Has it been validated in India?

• Is it age-appropriate?

• Is it language-appropriate?

• Does it require training?

• Does it need permission?

• Are laboratory measurements standardized?

• Are machines calibrated?

• Are observers blinded?

• Is inter-rater reliability needed?

• Will the tool produce categorical, ordinal, or continuous data?

• What statistical analysis will that allow?

This is where NotebookLM can save time, but not responsibility…..!

If it suggests a scale, verify the original source. If it mentions validation, check the paper. If it gives a citation, open it. If it cannot show the exact source line, do not trust it.

Research dies quietly in the measurement section.

Subscribe now

Use the Rigor Checker before your guide does

Most students discover flaws too late. During departmental review. During IEC review. During thesis defence. During journal rejection.

Use NotebookLM earlier.

The Methodological Rigor Checker prompt asks it to identify weaknesses related to validity, reliability, trustworthiness, bias, consistency, credibility, transferability, dependability, or confirmability, depending on the methodology.

This is especially useful because rigor differs by design.

A quantitative study needs clarity on internal validity, external validity, measurement reliability, confounding, precision, and statistical assumptions. A qualitative study needs credibility, reflexivity, dependability, transferability, audit trail, and saturation logic. A mixed-methods study needs integration, sequence, and explanation of how one strand informs the other.

Do not let NotebookLM use all these words as decoration.

Ask it to separate them:

“Which rigor criteria apply to my selected design, and which do not?”

This single question prevents a lot of academic nonsense and confusions.

A cross-sectional survey does not need confirmability in the same way a qualitative interview study does. A phenomenological study does not need Cronbach’s alpha unless it is also using a scale. A retrospective record review cannot pretend to have controlled exposure measurement if the records were not designed for research.

Remember, Methodology is not a vocabulary test…!

Share

Data analysis should be planned before data collection

Many researchers collect data first and then go hunting for a statistician.

That is backward.

The Data Analysis Planner prompt should be used before ethics submission, not after data entry.

Ask NotebookLM to review your objectives, variables, design, sampling strategy, and data collection plan. Then ask it to recommend analysis methods. It can help you map objectives to variables and variables to statistical tests. It can also identify when your data will not answer your stated objective.

For example, if your objective is to identify “risk factors,” but your design is descriptive and your sample size is small, NotebookLM should be asked to flag that mismatch. If your outcome is binary, the analysis differs from a continuous outcome. If your exposure is categorical, ordinal, or continuous, that changes the plan. If your data are paired, clustered, repeated, or time-dependent, ordinary tests may not fit.

The phrase to use is:

“Create an objective-wise analysis plan.”

This should include:

• Objective.

• Variables.

• Data type.

• Descriptive statistics.

• Inferential test.

• Effect measure.

• Adjustment plan, if any.

• Missing data plan.

• Assumption checks.

• Interpretation limits.

This is how a protocol becomes defensible.

Share

Only now ask it to write

The Methodology Chapter Builder comes late for a reason.

Writing should follow thinking.

Once the design, sampling, data collection, measurement, rigor, ethics, and analysis have been tested, then NotebookLM can help produce a coherent methodology chapter.

But do not ask for a final draft immediately. Ask for structure first.

• Then ask for a full draft.

• Then ask for a critic.

• Then ask for a summary.

• Then ask for a one-page blueprint.

• This sequence matters because each output has a different function.

• The chapter builder gives detail.

• The storyteller gives coherence.

• The critic gives pressure.

• The summary builder gives compression.

• The one-page blueprint gives presentation readiness.

Used properly, these prompts can take a researcher from scattered ideas to a protocol that can survive departmental scrutiny. The prompt set you shared is built exactly around this progression: foundation, design, logic, sampling, data collection, measurement, rigor, analysis, chapter writing, storytelling, critique, summary, and blueprint.

Subscribe now

The correct workflow for a researcher

Here is the practical sequence.

Create one notebook for one project. Do not mix unrelated projects. Google’s help page states that each notebook is independent and NotebookLM cannot access information across multiple notebooks at the same time.

Upload only the core sources first. Start with 8 to 15 high-value documents: guidelines, systematic reviews, landmark studies, local data, previous protocol, departmental template, and your rough concept note. Do not dump 50 random PDFs on day one.

Add a “project context note.” This should include setting, population, resources, constraints, timelines, available lab support, likely feasibility issues, and what kind of output you need.

1. Ask for a source-wise summary. Check whether NotebookLM has understood the sources correctly.

2. Ask it to compare sources. Look for disagreements, not just common points.

3. Ask it to identify gaps. Then verify those gaps in PubMed or Scopus.

4. Use the methodology prompts in sequence. Do not jump to chapter writing.

5. Ask for a critic after every major output.

6. Ask it to produce a one-page blueprint before you submit to your guide or IEC.

7. Export or copy the useful outputs into your own working document.

8. Then rewrite in your own voice.

9. That last step is not optional.

AI-generated methodology often sounds polished but hollow. Your final protocol must contain local detail, real feasibility, and precise procedural clarity.

Share

To Conclude : A practical prompt stack for medical researchers

Use this stack for almost any protocol.

Start with:

“Summarize each uploaded source in relation to my research topic. Focus only on findings relevant to my research problem, objectives, study design, population, variables, measurement tools, and limitations.”

Then:

“Compare the uploaded sources. Show where they agree, where they disagree, what remains uncertain, and what research gap is defensible.”

Then:

“Based on the uploaded sources and my setting, refine my research problem and suggest clear objectives and research questions. Do not invent claims beyond the sources.”

Then:

“Recommend the most suitable study design. Compare at least three possible designs and explain why the selected design fits best.”

Then:

“Create an objective-wise methodology plan with design, population, sampling, variables, data collection method, tools, analysis method, ethical concerns, and limitations.”

Then:

“Critically attack this methodology. Identify weak logic, feasibility issues, bias, missing variables, unclear definitions, ethical risks, and analysis problems.”

Then:

“Rewrite the methodology as a coherent academic argument suitable for a protocol or thesis submission.”

Then:

“Convert this into a one-page methodology blueprint for presentation to my guide or ethics committee.”

This is enough to change how most students write protocols.

Share

A Bonus set of prompts for general usage…

These 20 general NotebookLM prompts are for thinking, not decoration…!

This broader prompt sheet is also useful, especially for students and faculty who are drowning in source material.

The best prompts in this list are not the fancy ones. They are the disciplined ones….

1. “Instant Summary” helps when you first upload a long document.

2. “Beginner Explanation” helps when a student is new to a topic.

3. “Deep Dive Breakdown” exposes hidden concepts and assumptions.

4. “Compare Sources” is crucial because research begins where sources disagree.

5. “Study Notes Builder” converts reading into usable notes.

6. “Flashcards Generator” and “Quiz Me” are useful for postgraduate learning.

7. “Research Gaps” helps identify what is missing.

8. “Debate Both Sides” is excellent for critical thinking.

9. “Executive Briefing” is useful for busy faculty, ethics committees, and institutional presentations.

10. “Action Plan” converts knowledge into work.

But there is a danger….!! These prompts can make the user feel productive without becoming rigorous….!

Use these outputs as starting points.

Never use them as final academic truth.

Precautions while using NotebookLM, especially the free version…!!

The free version is powerful, but it is not unlimited and not infallible.

First, watch the source limit. Google’s current support page says a notebook can include up to 50 sources, and each source can contain up to 500,000 words or up to 200 MB for uploaded files. Do not waste that space on low-value documents. Upload fewer, better sources.

Second, remember that a source is a static copy. Google notes that an imported source is a static copy of the document you upload or import. If the original Google Doc, website, or guideline changes later, your notebook may not automatically reflect the updated version. For living guidelines and regulatory documents, recheck the source manually.

Third, check citations every time. NotebookLM provides citations, but citations are not a licence to stop thinking. Open the cited passage. Read the surrounding paragraph. Make sure the answer did not overinterpret a line.

Fourth, do not use it as a medical decision-maker. Google’s support page explicitly says not to rely on NotebookLM for medical, legal, financial, or other professional advice, and that such content is informational, not a substitute for qualified professional advice. For medical research, this means you can use it to organize literature and draft logic, but not to decide patient care, final diagnosis, treatment protocols, or ethical risk.

Fifth, do not upload material you have no right to use. Google’s source guidance tells users to avoid uploading documents they do not have rights to, and its support pages also warn users to respect copyright law. This matters for paid books, institutional files, unpublished manuscripts, confidential records, and student theses.

Sixth, be careful with patient data. De-identify everything. Remove names, phone numbers, hospital numbers, addresses, exact dates when not needed, and any combination of details that can identify a person. NotebookLM may be useful, but patient confidentiality is not negotiable.

Seventh, do not assume the mobile app has every feature. Google’s support pages repeatedly note that the NotebookLM mobile app may have limitations for some features, and some features such as Mind Maps are not supported on mobile at the time of the cited help page. Use the web version for serious research work.

Eighth, be aware of usage limits. Google states that limits apply to features such as Mind Map creation and that such limits are subject to change. Treat limits as a reason to work deliberately. Do not burn queries on vague prompts.

Ninth, avoid blind source dumping. If you upload poor literature, predatory papers, outdated guidelines, and irrelevant PDFs, NotebookLM will faithfully help you build a poor argument.

Tenth, remember that NotebookLM is not a database search engine. It works on what you give it. It can help you understand uploaded sources. It cannot guarantee that you have found all relevant evidence. PubMed, Cochrane, Scopus, Embase, clinical trial registries, ICMR guidance, NMC documents, and local policy sources still matter.

Subscribe now

The free version should be used like a disciplined reading room

Free tools reward discipline.

Before opening NotebookLM, write your research question. Write your objectives. Write your setting. Write what you already know. Write what you are confused about.

• Then upload sources.

• Then ask questions.

• The lazy sequence is: upload, ask, copy.

• The correct sequence is: frame, upload, interrogate, verify, rewrite.

This is the difference between using NotebookLM as a shortcut and using it as a research accelerator.

Other Resources from the author: