NHANES, the dataset this project was originally built around, could not do the one thing I ultimately wanted it to do: tell an individual person what to change. It could describe populations and show associations, but it could not take one person’s situation and turn that information into a clear recommendation for what they should actually try next.

That post ended with a promise that the next one would explain what replaced it.

A lot has happened since then. The project moved from research into a working system, and it finally has a name: GutStream.

This post is the full arc of that transition: what replaced NHANES, how the recommendation engine decides what to suggest, how the system determines whether something actually worked, what happens when it doesn’t, and how all of this fits with the clinicians who will eventually use the data.

I’m writing this for anyone who wants to understand how GutStream was built before trusting it, whether that’s someone dealing with gut symptoms or a clinician treating them.

The difference from what NHANES could have powered comes down to three ideas:

• Individual instead of population average

• Experimental instead of observational

• Safety-gated instead of blind

Part I: The Recommendation Engine

Who it’s for

GutStream will offer a free recommendation tool for people looking for evidence-graded guidance on what to try for their symptoms. The goal isn’t to give someone a list of 30 things they could change. It’s to give them one reasonable place to start, measure what happens, and decide what to do next based on the result.

Two decisions changed everything

The first major change was expanding the outcome set.

The project originally focused on brain fog, mental clarity, energy, and anxiety. Those were interesting outcomes, especially because people with digestive problems frequently report them, but the more I worked on it, the more obvious the limitation became. Someone dealing with gut problems doesn’t experience their symptoms in isolated categories: brain fog can show up alongside bloating, anxiety alongside abdominal pain, and fatigue alongside diarrhea.

Measuring the mental side while ignoring the gut side was only measuring half of the problem. So the system expanded to ten symptoms across both gut and mental health. That one decision made almost everything downstream harder, but it also made the product much closer to what I actually wanted to build.

The second major change was the evidence itself.

NHANES was useful for understanding populations, but it was never going to be enough for recommendations. Take coffee and focus. NHANES might show that people who drink more coffee report better focus, but that doesn’t tell us why. Maybe coffee improves focus, maybe people who are already more focused drink more coffee, or maybe another factor drives both.

A controlled trial gets closer to the question we actually care about. Researchers change one input, compare it with a control, and observe what happens to the outcome. That gives us an estimate of the effect rather than simply an association, which is much closer to what a recommendation system needs.

So the evidence base was rebuilt around experimental literature, study by study.

There is an important distinction here, though. Using trials does not mean every recommendation is proven causal for an individual person. A trial estimates an average effect across a group. When GutStream applies that evidence to one person, it is still testing whether that person responds. The rigor comes from the quality of the evidence and from measuring what happens next, not from pretending that a population-level trial proves what will happen to one individual.

Where the inputs come from

Before GutStream can make a recommendation, it needs to understand where someone is starting. The onboarding takes roughly ten minutes and begins by asking the top symptoms the person wants to focus on. From there, the intake adapts, with someone only receiving the questions that are relevant to their situation.

The point isn’t to collect every possible piece of information. It’s to collect enough to understand the person’s starting point, so the engine isn’t guessing about things like headroom, fit, or safety. Two people can therefore go through different numbers of questions depending on what is actually true for them.

That same principle carries into the safety system. A hard stop only works if the system knows the information needed to trigger it. If someone with compromised kidney function should not receive a particular recommendation, the intake needs to have asked about that risk in the first place

How recommendations are selected

Every possible recommendation goes through two filters before it can even compete for a spot.

1. Safety filter: if the system cannot confirm that an option is safe for that person, it is removed. Missing information counts as a no.

2. Fit filter: this asks whether the recommendation makes practical sense for that person’s situation. Here, missing information counts as a yes.

Those defaults are intentionally different. For safety, uncertainty should make the system more conservative. For fit, uncertainty shouldn’t unnecessarily prevent someone from trying a harmless option.

Once something survives both filters, it gets ranked using three factors:

• Evidence: how strongly the research supports the intervention

• Confidence: how confident we should be in that evidence

• Headroom: how much room this particular person has left to improve

Headroom is especially important. If someone already sleeps eight hours a night, sleep might have excellent evidence behind it, but there isn’t much left to improve. Two people can therefore receive completely different recommendations from the exact same evidence base simply because they are starting from different places.

Why I built a deterministic engine instead of a trained model

For a data scientist building a recommendation system in 2026, the obvious move would be to train a model. I decided on the opposite. GutStream’s recommendation engine is deterministic: the same person, with the same information, gets the same recommendation, and the system can show exactly why.

There are three reasons for that:

1. Auditability. When a system is making recommendations about someone’s body, “the model decided” isn’t a good enough explanation. Every recommendation needs to trace back to a rule, a piece of evidence, and information the person actually provided.

2. Safety. Hard safety guarantees require hard rules. A trained model can perform extremely well on average and still fail unpredictably at the edges, and people with chronic conditions are often exactly where those edge cases matter most.

3. The data isn’t good enough to justify the complexity. My original NHANES analysis explained roughly 13 to 22 percent of the variation in the outcomes I cared about. Training a sophisticated model on that data might produce impressive-looking predictions, but it would create a level of precision the underlying data doesn’t support. That would be false precision with good branding.

What actually shows up

Open the plan and you don’t get a wall of recommendations. You get one clear place to start for each symptom. Two additional recommendations sit behind the first one, held back intentionally rather than dumped on someone all at once.

The reason is simple: change one thing at a time, and you can actually tell whether it mattered. If someone changes three things and feels better, they don’t know which change helped. So each top recommendation runs for two weeks, long enough to reach a meaningful verdict.

If it works, it stays. If it doesn’t improve the symptom, the next recommendation moves into place and gets its own two-week trial. After trying the top three evidence-backed approaches and six total weeks, if outcomes still haven’t meaningfully moved through the highest-evidence options available, the system doesn’t just keep generating another suggestion. It tells the truth.

At that point, free lifestyle changes have done what they can, and the appropriate next step is a professional. Everything that wasn’t selected remains visible too. Nothing is hidden. Those options are simply labeled differently and sorted by evidence strength rather than personal fit. Some also require a practitioner to act on them, and those are clearly identified.

A good system needs to know where its own advice ends.

The safety layer

Most consumer health tools handle safety with a sentence at the bottom of the page: consult your doctor. I wanted GutStream’s safety system to be structural instead, because a disclaimer after a recommendation doesn’t actually prevent a bad recommendation from being made.

There are three main parts:

Hard stops. Some interventions are reasonable for one person and inappropriate for another. L-glutamine is one example: there is evidence around certain gut symptoms, but it can be inappropriate for someone with compromised kidney function. That’s not something the system should mention in a disclaimer after making the recommendation. The recommendation should never be generated in the first place.

Ongoing checkpoints. A system that measures mood and cognition will eventually be used by someone who is going through a genuinely difficult time. That means safety can’t be handled by one screening question during onboarding. GutStream checks in at intervals over time, with an immediate flag if someone’s initial answers indicate severe symptoms. Crisis resources remain visible on the platform at all times.

No fake gates. This one matters just as much. The system never claims to enforce a protection it cannot actually verify. If there is a safety gap I haven’t solved yet, it is documented internally as unenforced. I’d rather have an imperfect system that admits what it cannot do than a polished system that pretends it has solved something it hasn’t.

What happens when someone has multiple symptoms?

A recommendation engine that works for a textbook patient with one clean symptom isn’t very useful in the real world. Actual people have several things going on at once, so GutStream tracks up to two symptoms simultaneously, to preserve clean attribution, with each one running on its own independent track.

Bloating doesn’t borrow priority from brain fog, and brain fog doesn’t get ranked higher just because another symptom is worse. Each symptom is evaluated using its own evidence, confidence, and headroom.

The two-symptom limit is intentional. Every additional symptom spreads the available signal thinner and makes the daily check-in more burdensome. At some point, collecting more information starts making the system worse rather than better at identifying the true catalyst for change.

Two keeps the data meaningful while keeping the daily interaction short enough that someone will actually still be doing it a month later.

Potential nutrient deficiencies are handled separately. “You may be low in something” is a different kind of recommendation from “this behavior may help your symptoms,” so those recommendations are scored differently, based primarily on the probability of a deficiency rather than behavioral headroom.

One key rule applies across the entire system: Nothing that plausibly helps one symptom gets through if it plausibly worsens the other. The whole person comes before any individual score.

Part II: Measurement and Verdicts

An engine that recommends and never checks itself is an opinion with a database. Deciding what to suggest is half the system. The other half is knowing whether it worked, and what happens next because of it.

How outcomes get measured

Construct-aligned measures, not borrowed instruments. Daily check-ins capture three things:

1. Severity: How bad the symptom was

2. Frequency/Duration: How often or how long it lasted

3. Interference: How much it got in the way of daily life

The wording deliberately mirrors validated research instruments so the data stays interpretable to a clinician, but it does not administer those instruments directly. A questionnaire validated for monthly recall is no longer validated once you ask it every day. Borrowing the phrasing preserves comparability. Borrowing the instrument itself would have been a claim the design can’t support.

Burden matters. A system that exhausts people only produces data about the kind of people who tolerate exhausting systems. Daily check-ins are tiered to capacity, and on a rough day the whole interaction collapses to a single tap.

Episodes are logged as events, not averaged away. Brain fog does not arrive as a smooth weekly average. It arrives in episodes. A brutal Tuesday and a wrecked Wednesday can disappear inside a calm-looking weekly score.

Reaching a verdict

This is the part I care most about, and the part that separates this from every tracking app I studied. Trackers log. This reaches a verdict on whether a change worked.

Adherence is measured separately from outcome. If a change was barely followed, the honest conclusion is “this was not really tested,” not “this failed.” There is also an explicit too-early state. Most apps choose not to say it, because uncertainty does not retain users. Lastly, worsening requires more than one signal. A single bad reading never triggers a reversal.

Each outcome is judged on its own scale. Symptoms measured on incommensurable scales are not blended into one composite number, because a composite is only as meaningful as the assumption that the scales are comparable. They are not.

Verdicts are never tuned to feel encouraging. At no point in this build did “will this feel motivating” factor into what the system concludes. This last rule was the most expensive decision in the project. Encouraging verdicts retain users. Honest ones do not always. I took the retention cost, because the entire value of a verdict depends on whether it can be believed.

What a verdict changes

A verdict that doesn’t change anything is just a number. Here, it decides what happens to the plan itself.

If something is working, it stays. It does not get swapped out to make room for the next idea just because time has passed. The next recommendation gets added alongside it instead, so progress that’s real doesn’t get undone in the name of moving forward.

If something isn’t working, it gets replaced, not left in the plan out of momentum.

Every lever added this way tracks its own timeline from the day it started, so when three things are eventually running at once, success and failure never blur into one undifferentiated bundle. There’s still a ceiling, so a working plan can’t quietly turn into a pile of everything that’s ever helped a little.

Each change traces to a logged record: what started, when, against which outcome, and what the verdict was, so the history is auditable rather than remembered.

Part III: Built With Clinicians in Mind

Most of what happens to a patient happens where no practitioner can see it, in the weeks between appointments. That absence shaped decisions throughout this build, not just the ones that show up on a clinician’s screen.

It’s why check-ins happen every day instead of only at appointments. A patient’s best recollection of a rough month is not data, it’s a guess dressed up as a report, and that guess is what most clinicians currently work from. Daily logging replaces the guess with an actual day-by-day account of what happened, which is the only way “between visits” stops being a blind spot.

But daily only works if daily is fast. A check-in that takes real effort gets skipped, and a skipped check-in is worse than no system at all, it’s a gap that looks like data. So the patient side had to stay under ninety seconds, something closer to a glance than a form. The harder problem was making that same minute and a half add up to something a practitioner could actually use, not just a pile of taps. Both sides had to be true at once, or neither one was worth building.

Two rules govern what the clinician actually sees and controls. First, the clinician prescribes and the system measures. When a practitioner directs a specific change, the engine tracks it faithfully and never overrules the human. Second, sustained non-improvement at good adherence is a referral signal, not a retry loop. When someone follows a plan closely and still does not improve, confirmed over time rather than off one data point, the honest move is to say so and route them to get professional assistance, not keep proposing new levers to pull.

From a Data Perspective

The NHANES version of this project was comfortable. Clean, defensible, and genuinely interesting to work through. But comfortable was never the point. Somewhere in that comfort I had to admit what I actually wanted: not a good dataset, a system that helped someone.

What I wanted turned out to have two halves, both from the same place. I know exactly what it feels like to be handed general advice for a specific body. Real experimentation, trying something, tracking it honestly, and adjusting has been the only thing that’s ever worked for me. I wanted to build the tool I wish I’d had: pulling from the best available evidence, handing someone a real place to start, at no cost, because the ability to try shouldn’t be something you have to pay for.

The other half came from the other side of those same appointments. I’ve sat across from clinicians doing their best with almost nothing to work from, because a patient’s memory of the last month is an impression, not data. I wanted to give GI clinicians and dietitians what I never saw them have: a real, day-by-day account of what a patient’s body was actually doing.

Those two halves are why this stopped being a data project and became something that has to earn trust twice, once from the person using it daily and once from the person reading their chart. Hardest and most rewarding thing I’ve built, and a real part of what made it possible in this timeframe was learning to work alongside AI systems. Not just to use them, but to learn where they genuinely speed engineering up and where they quietly cost you if you’re not paying attention.

A few things I didn’t expect to learn:

The instinct to reach for a smarter model is usually a distraction from a harder question: can you defend the answer at all. Determinism isn’t a safety compromise, it’s what lets you stand behind an answer instead of hoping it’s right.

Measurement design isn’t a supporting detail, it’s the ceiling on what you’re even capable of discovering. How you ask a question, how often, at what resolution, quietly decides what truths are reachable later. Get that wrong early and no clever analysis downstream fixes it.

GutStream launches in a few weeks. Next post: what it costs, how the free side works, what it gives clinicians that a practice management tool doesn’t, and how to get in.

That number gets treated like data. It goes in the chart and shapes what happens next, but it is one of the few numbers in medicine with no objective reference behind it. Sleep has hours. Weight has pounds. Even bowel habits have a scale with pictures next to it. A six means something different to every person who says it, and sometimes to the same person on different days.

That isn’t just frustrating in an exam room. It slowed the science too. A symptom that cannot be compared between people is hard to study, and persistent abdominal pain sat in the “we do not know” pile far longer than it should have.

The Balloon Test

What moved the research was a balloon.

Researchers needed a way to measure pain using the exact same stimulus in every person. In a study of 86 people with IBS and 25 healthy controls, researchers inflated a small balloon in the rectum at controlled pressures and recorded when each person reported pain. Pain began at 30.4 mmHg in the IBS group and 44.5 mmHg in the healthy controls (Gastroenterology). Same balloon, same pressure, different point at which the sensation registered as pain.

That difference eventually became part of a much larger idea. In 2017, the International Association for the Study of Pain adopted nociplastic pain as a third category, alongside pain caused by tissue damage and pain caused by nerve injury (The Lancet). It describes pain caused by the nervous system amplifying otherwise normal signals. Chronic visceral pain sits inside it.

One question researchers still debate is direction. Does heightened sensitivity cause chronic pain, or do years of chronic pain create heightened sensitivity? Evidence supports both possibilities, and nobody has fully untangled them.

Fortunately, whether hypersensitivity is the cause or the consequence doesn't change much when it comes to treatment. The goal is the same: reduce how strongly those signals are amplified. A clean scope was never evidence that nothing was wrong. It was evidence that the problem was not the kind of thing a scope can see.

What You Can Try First

If pain is being amplified rather than driven by ongoing tissue damage, the obvious question becomes: can anything reduce that amplification without a prescription? Three approaches have the strongest evidence.

1. Self-guided cognitive behavioral therapy is the one most people skip because it sounds like it needs a therapist. In many cases, it doesn’t. A 2024 network meta-analysis of 42 trials and 5,220 patients found minimal-contact or self-guided CBT cut the risk of abdominal pain not improving by 29%, RR 0.71 (Gastroenterology), and it performed no differently than the versions delivered face to face.

2. Gut-directed hypnotherapy has its own substantial body of research. A 2025 meta-analysis of 12 trials and 1,158 patients found it improved abdominal pain specifically compared with standard care, with the largest gains when delivered in groups (Neurogastroenterology & Motility). Smaller effect than CBT’s, reported differently too, a standardized mean difference rather than a risk ratio, so treat it as directionally consistent rather than head-to-head comparable.

3. Peppermint oil is different. It doesn't change how the brain processes pain signals. Instead, it relaxes intestinal smooth muscle directly. A 2022 analysis of 10 trials and 1,030 patients found it cut the risk of abdominal pain not improving by 24%, RR 0.76 (Alimentary Pharmacology & Therapeutics). Side effects were mostly heartburn, and enteric-coated capsules exist specifically to cut that down.

Second Line

If non-prescription approaches aren't enough, the treatment with the strongest evidence may surprise you.

Tricyclic antidepressants change how pain signals are processed within the nervous system. That is why they have been used for decades in nerve pain with nothing to do with digestion, at doses well below what treats depression. Pooling seven trials and 708 patients, a 2025 analysis found tricyclics cut the risk of abdominal pain persisting by roughly 31%, compared with placebo (Lancet Gastroenterology & Hepatology). Among gut-brain neuromodulators studied for IBS pain, tricyclics have the strongest and most consistent evidence.

SNRIs look better on that chart, and they are a different antidepressant class that also looks promising. But two trials and 94 people is a thin foundation, which is what the width of that bar is telling you. Right now tricyclics earn the recommendation because they have the most trial consistency behind them, not necessarily because they’re the best fit for every person in that chart.

The evidence for the other classes is thin because almost nobody has tested them properly for pain, not because they don't work. The researchers who ran these numbers say as much themselves, calling for more trials of SNRIs and the newer classes specifically. As better trials accumulate, treatment will likely become less about choosing one "best" drug and more about matching the right mechanism to the right patient.

Takeaways

None of this tells you which lever is yours. It tells you which levers exist, and how well each one has actually held up when tested.

• A clean scope isn’t proof nothing is wrong. People with IBS registered pain at 30.4 mmHg in one study, healthy people at 44.5. Same balloon, same pressure, different threshold.

• Three non-prescription approaches have the strongest evidence. Self-guided CBT and gut-directed hypnotherapy target pain amplification, while peppermint oil takes a different approach by relaxing the gut itself.

• Tricyclic antidepressants remain the leading prescription option, cutting the risk of persistent abdominal pain by about 31% across studies.

As always, promising results in a study don't guarantee they'll work for you. Research can tell you what works on average. Only careful trial and tracking can tell you what works for you.

Most of that attention lands on the same explanation: these foods are engineered, calorie-dense, and easy to overeat. For general health, that’s a real part of the story. For the gut, it’s largely beside the point. The mechanism doing the most specific damage there isn’t about how many calories a food delivers or how it’s flavored. It's about emulsifiers, the detergent-like additives that hold a product together on the shelf, and the mucus barrier in your gut they were never meant to touch.

This post covers that barrier, the controlled human trial that tested what happens when it's exposed to one common emulsifier, and the population-level pattern that shows up across 116,000 people tracked for nearly a decade.

The Barrier Under Attack

Your intestinal lining is not in direct contact with your gut bacteria. Between the two sits a mucus layer, and its entire job is to hold those bacteria at a physical distance from the cells lining your intestine. That separation is one of the primary ways the intestine protects itself from the trillions of microbes living inside it. Anything that weakens that layer is compromising one of the gut's primary defenses.

Emulsifiers function like detergents. They keep oil and water from separating, which is why they appear in ice cream, salad dressing, packaged bread, and almost every sauce sold in a jar. Researchers wondered whether detergent-like molecules repeatedly passing over that mucus layer could slowly erode the barrier designed to keep bacteria at a safe distance.

In 2015 they tested it. Mice given relatively low concentrations of two common emulsifiers developed low-grade inflammation and metabolic syndrome. Furthermore, mice already predisposed to colitis developed robust colitis. In both cases, their bacteria moved in closer to the intestinal wall than they should have been able to (Nature).

What made the study especially convincing came next. They transferred the altered microbiota into germ-free mice, animals raised with no bacteria of their own, and those mice developed the same inflammation. That proved causality and passed a bar most nutrition research never clears, further strengthening the case against emulsifiers and the necessity to learn more about their impact.

Then Came the Human Trial

Mice are not people, and a finding this clean in a rodent model has an obligation to survive a human test before anyone acts on it.

Six years later, researchers tested the question in humans. A double-blind controlled feeding study put 16 healthy adults on a diet containing no emulsifiers at all. Seven of them had 15 grams a day of carboxymethylcellulose added to otherwise identical food. Neither the participants nor the researchers evaluating them knew who was in which group (Gastroenterology).

After eleven days, the group receiving the emulsifier reported more abdominal discomfort after meals. Their microbiome composition shifted and lost diversity. Their stool chemistry changed as well, showing reductions in short-chain fatty acids, the compounds gut bacteria produce from fiber that feed the cells lining the colon and help keep inflammation in check.

Two of the seven even showed exactly what the mouse finding predicted: bacteria encroaching into the mucus layer meant to stay clear. While the study was small, it showed that the mechanism observed in mice could also be detected in people under controlled conditions.

Population-Level Risk Factors

Mice showed the mechanism. The double-blind trial showed it happens in humans. The last question is whether those findings translate into disease risk across an actual population.

Researchers followed 116,087 adults in 21 countries for close to a decade, tracking who developed inflammatory bowel disease and how their diet predicted it. The people eating the most ultra-processed food carried the highest risk, and it rose in line with their consumption (BMJ).

What sharpens this finding is which specific foods were driving it. Soft drinks, refined sweetened foods, salty snacks, and processed meat each carried their own independent association with IBD risk, on top of the overall dose-response pattern above. Meanwhile, white meat, red meat, dairy, starch, and fruit, vegetables, and legumes showed no association at all, in the same cohort, using the same methods. It wasn't food volume, and it wasn't a particular macronutrient. It was a specific slice of the ultra-processed category.

Five or more servings a day sounds like an extreme diet. For a lot of people it isn't. On average, Americans get around 55% of their daily calories from ultra processed foods (CDC). Packaged cereal, flavored yogurt, deli meat, a soda, something from a box for dessert, and the count is there without anyone consciously trying to eat "processed." For many Americans, the highest-risk category is closer to an ordinary day than they might assume.

Takeaways

The case against ultra-processed food usually gets argued on calories and unintentional overeating. For the gut specifically, the sharper problem is physical, and it holds up at every scale researchers have tested it: a mouse model, a controlled human trial, and a population followed for a decade.

• The mice proved it’s causal, not just correlated. Emulsifiers act as detergents on the gut’s mucus barrier, and transferring the altered microbiota alone into germ-free mice reproduced the same disease.

• Eleven days was enough to see it in real people. A blinded human trial found one common emulsifier lowered microbiome diversity and caused measurable mucus-layer encroachment.

• At population scale, the risk was undeniable. Across over 100,000 adults tracked for nearly a decade, IBD risk climbed with intake from specific categories while whole foods showed none.

The research doesn’t suggest that every packaged food is harmful or that every emulsifier should be avoided. Instead, it points to a more useful way of thinking about ultra-processed foods. Your gut doesn't know whether a food came from a factory. It responds to the molecules that reach its lining. For emulsifiers, evidence from animal studies, controlled human trials, and large population studies consistently point to the same underlying mechanism.

As the science evolves, the conversation is becoming less about whether a food is “processed” and more about which specific ingredients alter the gut environment. That shift matters because it replaces a broad label with something people can actually understand, measure, and make decisions around.

This post breaks down the most research-backed ways to handle bloating. I wanted to stick to natural strategies that are cost free, so pills, powders, or supplements are excluded. Each tip comes with an honest read on how good the evidence really is, so you know where to start and what to expect.

Low-FODMAP

If you try only one thing, make it this one as it has the most evidence behind it. A low-FODMAP diet temporarily removes a group of fermentable carbohydrates found in foods like onion, garlic, wheat, beans, and certain fruits that the small intestine struggles to absorb. For bloating in particular, it stands as the clear best dietary option. A 2025 network meta-analysis compared eleven different diets, and only low-FODMAP did better than a normal diet at reducing bloating, cutting the chance of it sticking around by almost half.

The reason it works is straightforward. Because those carbohydrates are poorly absorbed in the small intestine, they pass into the colon, where gut bacteria ferment them and produce gas. A separate intervention trial measured exactly this and found that a high-FODMAP diet produced about four times the breath hydrogen of a low one (JGH). The study was small, but the result is hard to misread: more of these carbohydrates means measurably more gas.

There is an important nuance to remember. Low-FODMAP was designed as a short-term test, not a way of eating forever. The full approach has three phases:

1. A few weeks of strict restriction.

2. A careful reintroduction to learn which carbohydrates are actually your triggers.

3. A relaxed diet that avoids only those.

Staying in the restriction phase permanently has a real downside. In one randomized trial, just four weeks of it noticeably lowered beneficial Bifidobacteria, because the same carbohydrates that feed gas-producing bacteria also feed the helpful ones. (The Journal of Nutrition). The aim is to use the restriction only long enough to find your real triggers, then deliberately widen the diet back out, because keeping it narrow for too long trades one problem for another.

Move, Especially After Eating

The next approach changes nothing about your food, and the evidence for it is stronger than most people expect. In a randomized control trial of 102 people with IBS, those coached to move more improved their symptom scores about ten times as much as those who stayed at their usual activity level (AJG). The researchers went so far as to call physical activity a primary treatment, not an afterthought.

Bloating responds to this for a specific reason. People who bloat more are often not producing more gas than anyone else, they are just worse at clearing it, and movement helps it pass through the system. In a smaller study, gas was retained at 45 percent during rest but only 24 percent during light pedaling, with less discomfort.

Source: American Journal of Gastroenterology

In practice, that means a 10 to 15 minute walk after meals, when your system is busiest with digestion. It won’t fix everything, but it does something no food list can, which is help the gas you do produce actually move through and out.

Diaphragmatic Breathing

As surprising as it may sound, some bloating is not about gas at all but rather about muscle. When the gut fills after a meal, the diaphragm should lift and the abdominal wall should tighten, keeping the belly flat. In many people who bloat, that reflex reverses. The diaphragm drops, the wall goes slack, and the stomach pushes out with barely any extra gas behind it. This is referred to as abdominophrenic dyssynergia, and it’s why a flat morning stomach can balloon by night.

The fix is to retrain it with diaphragmatic breathing, slow breaths that restore normal coordination between the diaphragm and abdominal wall. In one randomized, placebo-controlled trial, patients with visible bloating who learned to correct this pattern saw their distension drop by about two thirds, while the placebo group saw no change (Gastroenterology). For good reason, the AGA now recommends retraining this muscle pattern.

Takeaways

There is no shortage of pills and powders marketed for bloating, and some do have evidence behind them. What gets lost in the noise is how much you can manage yourself, for nothing, with strategies that target the actual mechanisms.

• Low-FODMAP has the strongest evidence of any natural approach, but don’t treat it as a forever diet. Restrict to find your triggers, then widen back out.

• Movement is proven for IBS and helps clear trapped gas. A 10-to-15 minute walk after meals is the simplest version, timed for when digestion peaks.

• Diaphragmatic breathing retrains the muscle reflex behind visible bloating, supported by a placebo-controlled trial and major guidelines.

The thread running through all three is that bloating rarely has one cause, so it rarely has one fix. It sits at the intersection of diet, motility, the gut microbiome, and the nervous system. As a result of its complexity, the levers that matter most are deeply individual.

That is why the move that matters most is not picking the “best” remedy but testing these against your own body. Bloating can seem random, but it runs on mechanisms you can actually influence, and the most effective tools here are free of charge. Test each approach, track what changes, and build from the ones that hold up. The goal is not to follow a list, but to learn how your own gut works. Most people who try this find there is more they can do about it than they expected.

One of those effects is brain fog. Difficulty concentrating, slower thinking, forgetfulness, and mental fatigue are commonly reported by people with IBD, despite receiving far less attention than the physical symptoms of the disease. Yet a growing body of research suggests the relationship between IBD and cognition may be much stronger than previously thought.

This post explores the specific ways in which cognitive function is impacted by IBD, how brain fog can persist even after the disease appears to be under control, and why that connection is only now beginning to get the attention it deserves.

It’s Not in Your Head

For something so easy to dismiss, brain fog holds up surprisingly well under direct testing.

In 2021, researchers pooled eleven studies that put IBD patients and healthy controls through the same standardized cognitive tasks. The combined analysis, drawn from a search of over 8,000 articles, found consistent deficits in attention, executive function, and working memory.

Source: Hopkins et al. (2021), Journal of the Academy of Consultation-Liaison Psychiatry

Each bar shows how far IBD patients fell behind healthy controls on cognitive testing, with working memory showing the largest gap of the three. These are not obscure laboratory measures. They are the same mental processes behind staying on task at work, following a conversation, and finishing a thought before it slips away. These results held even after excluding elderly patients and those in active flare.

What sharpens this is what did not show up. On the MMSE, the quick cognitive screen most physicians reach for when a patient reports trouble thinking, the same patients scored completely normal. The deficits were real. The standard screen just could not see them. A clean result on the most common cognitive tool in clinical settings is not the same as a clean bill of cognitive health.

The deficits are measurable, they replicate across labs, and they appear even in patients whose disease is actively under control. That last part is where the story gets especially complicated.

The Remission Trap

The reasonable assumption is that when the gut settles, so does everything downstream of it. Inflammation resolves, symptoms fade, and the brain clears. For a meaningful share of people with IBD, that final step never comes.

A study by Breslin et al. followed Crohn's patients already in clinical remission, tracking both their cognitive performance and their bloodwork over six months. Their attention stayed measurably worse than healthy controls at both visits. And while their disease was officially quiet, their blood told a different story. IL-6, a marker the body uses to signal ongoing inflammation, remained elevated. Tryptophan, which gets converted into serotonin, was being diverted down an inflammatory pathway instead. The gut looked stable but the inflammatory program running beneath it had not caught up.

Source: Brain Inflammation Collaborative

That rerouting process has a name. The kynurenine pathway already handles most of the body's tryptophan, but it splits in two. One branch produces a protective compound, the other a neurotoxin. Inflammation tips that balance toward the toxic side, and those byproducts can cross into the brain. It is one of the clearest routes by which inflammation in the gut turns into a measurable effect in the head, and it runs even without an active flare.

This is the part that reframes the symptom. Brain fog in IBD is not a vague accompaniment to the disease, or the residue of a few bad nights of sleep. It is the same inflammatory process that drives the gut, reaching tissue the standard tests were never built to see.

The disease did not stop at the intestine. It just stopped being measured there.

The Lens is Finally Widening

For most of the time IBD has been studied, cognition was not part of the conversation. The disease was defined by what it did to the gut and measured by what showed up on a scope. Brain fog was something patients mentioned and clinicians had no framework to address. You cannot track what you cannot measure, and for decades, no one was measuring it.

That is starting to change. Brain fog did not have a validated, IBD-specific measurement tool until 2024 (Knowles & Dickinson). It sounds like a technicality, but it is the step everything depends on. Its arrival marks the point where cognition starts being treated as part of the disease rather than a footnote to it.

The shift matters because of what the mechanism tells us. If the fog is driven by inflammation, then it is responsive to the things that influence it in the first place: keeping the disease well controlled, correcting real nutrient deficiencies, protecting sleep, staying active. None of these have been tested head-to-head against brain fog, and anyone promising a clean fix is overselling. But the fog has identifiable inputs, which is a very different situation than a symptom with no handle at all.

The future here is not a sharper scope or a breakthrough drug. It is people paying attention to how they think and function over time, and connecting it to what they are actually doing, until the fog stops being a mystery and starts being something with visible cause and effect.

Takeaways

IBD is not typically thought of as a condition with cognitive consequences. Most of the existing research connects it to depression and anxiety, while its direct impact on the brain has been studied far less. Emerging work suggests that overlap runs deeper than the field assumed.

• Brain fog is real and measurable. On direct cognitive testing, IBD patients lag healthy controls on attention, executive function, and working memory, even when routine screens look normal.

• It outlasts the flare. Cognitive deficits persist in patients whose disease is in remission, driven by the same inflammatory process that reaches the brain through pathways like kynurenine.

• It went unmeasured for decades. Brain fog did not have a validated, IBD-specific measurement tool until 2024, which is exactly why it has been so easy to overlook and so hard to act on.

What happens next comes down to a shift that is already quietly underway. Fatigue spent years being brushed off in IBD until there was finally a way to measure it, and almost overnight it became something doctors took seriously. Brain fog is reaching that same turning point now.

And as it does, what counts as feeling better starts to widen. Not just a calmer gut, but a clearer head, the part of the disease you carry into every conversation and every workday. The science spent years a step behind what patients already knew. It is finally catching up.

1. Context: Who someone is

2. Drivers: What someone does

3. Outcomes: What someone experiences

All of it is drawn from NHANES, the large, nationally representative U.S. health survey, all of it pointing toward the same goal. Tell the system who you are and how you live, and it tells you what's most worth changing. Before any scores were generated, the dataset had to be reduced and validated. This post covers that process, the findings, and what they mean for where FPI goes from here.

The Modeling Approach

The first question following preprocessing was straightforward. Of the 424 variables that remained, which ones did the model actually need?

To answer this, SHAP (SHapley Additive exPlanations) was applied to a gradient-boosted tree model (LightGBM), measuring how much each feature contributed to each prediction. True black-box models have their place, but a system making health recommendations needs to value interpretability just as much as accuracy. This approach makes it possible to see what is driving the outcomes rather than treating the model as a sealed machine.

With that, iterative pruning began with the goal of reaching a tighter feature set without sacrificing much predictive power. Every variable faced the same two tests, scored on normalized SHAP so each number reads as a share of the model's total importance:

• Average importance: mean SHAP share across outcomes ≥ 0.3%.

• Peak importance: SHAP share on its strongest single outcome ≥ 1.0%.

A variable survived if it cleared either test. Falling short of both meant it was dropped. The cut ran in iterative passes, retraining and rescoring after each round, since removing one set of variables changes how the remaining ones rank. It continued until a full pass removed nothing.

After three iterations, the feature set was reduced from 424 variables to 77. The final set split into 31 context variables, 31 behaviors, and 15 biomarkers, with an even count between the two main groups. The more important question was how that split translated into predictive weight.

The Strongest Predictors

Sorting the remaining variables by importance and splitting them by either context or behavior gives the clearest picture of what influences the model.

The variables carrying the most weight describe who a person is, not what they do. Prescription count, age, employment status, and food security all outrank the strongest behavioral variable in the model. Although context and behavior contain the same number of features, context accounts for 55% of total model importance versus 33% for behavior. Sleep and activity show up consistently, but the strongest signal of how someone feels is the environment they are operating in.

While context carries the most weight, it falls outside the scope of recommendations or guidance that can be given. Behaviors remain the most direct way to see how day-to-day inputs move each outcome, shown in the chart below.

Across all four outcomes, many of the findings were predictable. Sleep timing was more important than sleep length, falling in line with recent circadian research. Heavy drinkers and smokers consistently reported lower outcomes in focus, anxiety, and mood. Moderate activity held up as an important staple for functional health.

These patterns are worth knowing. But a population correlation is closer to a weather average than a weather forecast. It describes what tends to happen across a large sample, not what happens when a specific person decides to change a specific behavior. That is a key distinction of cross-sectional data, and an important caveat to keep in mind when looking at the results.

The Limitations of NHANES

That caveat is exactly what the numbers show. Across the four outcomes, the full model explained only 13% to 22% of variance between individuals, with mood on the high end and focus at the bottom. That ceiling is expected for a single cross-sectional snapshot of a heterogeneous population.

The limitations are not surprising when you consider what NHANES actually captures. Every participant enters the dataset at a single point in time, so the model sees where people landed, not the weeks and months of sleep, stress, and habits that got them there. Two people can arrive at the same outcome through entirely different pathways, and a one-time survey cannot tell them apart. Most of the unexplained variation lives in that missing time dimension.

Human performance is shaped not just by who someone is and what they do, but by how those factors interact over time. That does not make NHANES less valuable. It simply defines its role. While well suited for identifying population-level patterns and highlighting factors worth investigating, truly personalized insight requires following individuals over time.

From a Data Perspective

Early on, this system was designed with the hypothesis that NHANES could function beneath the recommendation engine, estimating the marginal impact of behavioral changes on outcomes. But two things made it clear that wasn’t viable:

1. Limited predictive power, with only 13%–22% of variance explained across outcomes.

2. The lack of true behavioral depth in a single cross-sectional survey.

The biggest surprise was one I had not set out to model. Behaviors and context explained each outcome poorly, but the outcomes were reliably tied to one another with the same relationships holding up across several different cuts of the data. Mood, anxiety, energy, and focus behaved less like four separate targets and more like one connected system. That reframed what the data is for. Rather than predicting outcomes, NHANES will serve two specific functions as FPI develops.

The first is connecting the outcomes to one another. Most research on sleep, activity, and the rest measures its effect on a singular outcome. This is where those correlations do real work.

Each relationship becomes a coupling, a fixed conversion factor between two outcomes, set deliberately below the raw correlation so it understates the effect. This accounts for differences between patterns observed across a population and what holds within any single person. The result is that for whichever outcome is selected, the system can reliably draw on a far wider body of research than studies that measured that one outcome directly.

The second is deciding who to compare a person to. The variables that carried the most weight were context, the parts of someone’s situation they cannot simply change, like their age or the conditions they already manage. Those make poor advice but excellent grouping tools. They show which traits matter most for grouping people into meaningful peer sets, so FPI can eventually personalize guidance based on others with genuinely similar contexts, rather than a faceless national average.

From here, FPI Update #10 turns to the recommendation engine, the part that takes this body of research, extends it through the couplings, and turns it into ranked guidance. I appreciate your continued support as FPI moves toward a usable and reliable system.

Compare that to how we measure the heart, or sleep, or glucose. Each has decades of tracking infrastructure behind it, tools designed to close the feedback loop between a behavior and its biological effect. The gut has almost none of that. The most common GI disorders are diagnosed by exclusion, meaning clinicians work through what something is not before arriving at what it might be.

For patients, that process often takes years. For anyone trying to understand or improve gut function, it means operating with very little signal. That is the data problem this post will unpack, highlighting three key reasons why it persists.

1. Designed for Measurable Disease

When most people think about medical testing, they assume that if a problem exists, a test can measure it. In gastroenterology, that assumption breaks down often.

GI medicine is exceptionally good at finding structural damage. Colonoscopies detect polyps and inflammation. Imaging catches obstructions and tumors. Blood work flags infection and immune activity. These tools answer one question: is something visibly wrong?

The trouble is that the most common digestive disorders are not structural. IBS and the other disorders of gut-brain interaction involve motility, nerve signaling, visceral sensitivity, and immune regulation, not visible tissue damage. A patient can have severe symptoms while every scan, scope, and blood panel comes back clean.

The two colors in that chart capture the core issue. A small set of gut conditions like inflammatory bowel disease and celiac disease have clear confirmatory tests. The far more common functional disorders do not. Without a definitive marker, diagnosis becomes a process of elimination. Clinicians rule out measurable conditions one by one until what remains is labeled functional. For patients, that path is often slow, fragmented, and frustratingly indirect.

2. The Sensation Gap

The absence of functional measurement creates a second issue: when objective data is limited, subjective experience becomes the primary signal.

Few biological systems respond to behavior as quickly as the gut. Dietary changes, stress, sleep disruption, alcohol, or even social context can trigger noticeable changes in digestion within hours. The problem is that subjective experience is often a poor proxy for what is happening at the biological level.

Recent research highlights this challenge. In a 2024 study in the Journal of Clinical Medicine, researchers tested how well those impressions align with objective measurement. They used intestinal ultrasound to assess what was physically in patients' bowels and compared it to what those same patients reported.

The two correlated poorly. Patients with very different symptom profiles often showed substantial overlap on the objective measure, and the relationship between what patients reported and what the scan found was relatively weak. None of that makes the symptoms any less real. It simply illustrates how difficult it is to infer underlying biology from sensation alone. For many patients, the experience itself becomes both the signal and the evidence.

3. The Snapshot Problem

Even if we solved the first two problems, a third challenge remains. The gut is not a static system. It responds continuously to behavioral inputs, from what someone ate to the environment they live in. As life changes, so does the gut. The challenge is that most of our measurement tools were not built to capture the daily realities of our microbiome.

Research suggests that this variability is far greater than many people realize. In a 2021 study in Nature Communications, researchers sampled the same individuals’ gut microbiomes every day for six weeks. For 78% of bacterial types, within-person variation over time exceeded variation between different people. Some bacteria shifted by as much as 100-fold between consecutive days. That is not a gradual drift, but rather a system that can look completely different from one day to the next.

Other health domains solved this problem through continuous monitoring. Glucose management changed when finger-stick measurements gave way to continuous glucose monitors. Fitness changed when occasional weigh-ins gave way to wearables generating thousands of observations per day. The gut has not yet undergone a similar transition. Most digestive decisions are still being made from isolated snapshots.

Takeaways

The gut is the hardest organ to track because everything about it resists measurement. Diagnostic tools miss the most common conditions, the feedback is largely subjective, and the system never holds still long enough to read cleanly. Each problem makes the next even harder to solve, and most digestive health decisions are still being made partially in the dark.

• The standard tests were built to find structural damage, but the conditions affecting the most people are largely functional.

• Without an objective readout, people fall back on sensation, which tracks poorly with what is happening at the biological level.

• Even a good measurement is only a snapshot of a system that can shift drastically across days.

The future of gut measurement looks less like a sharper test and more like a longer one. A single reading cannot survive a system that swings this much, but a dense stream of readings from the same person, gathered over weeks and months, turns that noise into a personal baseline. That is the shift many other domains of health have already made.

From there, the inputs matter as much as the outputs. Pairing a long record of how someone feels and functions with what they were eating, sleeping, and managing is what turns scattered data points into cause and effect. That is the data the gut has never had, and assembling it is the difference between guesswork and understanding.

The intuition of this layer is part of what makes it so interesting to build. People already have strong priors about exercise, sleep, and substance use. The model's job is not to confirm what everyone already suspects, but to find where the signal actually exists and how strong it is compared to everything else in the system.

Physical Activity

NHANES grouped physical activity into two primary types:

• Moderate activity: causes light sweating or a moderate increase in breathing and heart rate. Brisk walking, casual cycling, recreational swimming, and light yoga.

• Vigorous activity: causes heavy sweating or a large increase in breathing and heart rate. Running, HIIT, competitive sports, and heavy resistance training.

Sedentary time was also collected independently, capturing how much of the day someone spends sitting regardless of how much they exercise. The visual below shows how both activity types associate with outcomes across all five dimensions.

More movement, better outcomes. The relationship holds across all five dimensions and the gap between the most and least active groups is one of the more consistent findings across the entire dataset. The person getting regular activity in is pulling away from the person who is not, and the data shows it across energy, mood, cognition, and beyond.

What is more interesting is what happens when you separate the two activity types. Moderate activity held its own against vigorous across mental health outcomes and in some cases edged ahead. For mood and energy in particular, a long walk appears to do at least as much as a hard run or lift.

Sleep

NHANES captured sleep data across several dimensions: average hours, weekday and weekend hours separately, bedtime and wake time for both, and a derived measure of social jetlag. That last variable reflects the shift people make between their weekday and weekend schedules, the kind of internal clock disruption that tends to fly under the radar but has a real effect over time.

Sleep duration shows one of the cleaner patterns. Higher mood scorers cluster tightly around seven hours while lower mood scorers spread further into both tails, undersleeping and oversleeping at higher rates. The middle tertile sits in between, but the separation at the extremes is where the real story lives.

What is more surprising than hours alone is that timing appears to be just as impactful. Weekday bedtime and weekend wake time both showed meaningful associations with anxiety and mood outcomes, independent of total duration. Someone sleeping seven hours on a shifted schedule looks different in the data than someone sleeping seven hours on a consistent one.

Duration gets most of the attention. But what the data suggests is that when and how consistently someone sleeps may matter just as much.

Substance Use

Substance use was measured through alcohol, tobacco, and caffeine-related variables. Alcohol produced the clearest signal of all three. It was tracked across quantity, frequency, binge behavior, and a chronic use flag derived from sustained heavy consumption patterns. The visual below compares non-drinkers against chronic users across the four mental health outcomes the FPI is built to improve.

Across all four outcomes, chronic users score consistently and meaningfully below non-drinkers. If you find yourself running low on energy, struggling to focus or concentrate, or just feeling a little off mentally, it is worth asking whether drinking patterns are part of the picture.

Something worth noting is that cognition and mood show slightly negative scores for non-drinkers too, sitting just below the population average despite being the “healthy” group. That feels counterintuitive but reflects a well-documented pattern called the sick quitter effect, where people who abstain entirely are more likely to have stopped due to illness or medications than to have simply never had the habit.

Examples like this are precisely why the context layer exists. Before any behavioral signal gets evaluated, the model first accounts for age, medical history, and the broader circumstances someone is operating within. A non-drinker managing multiple chronic conditions is not the same baseline as a non-drinker in good health, and without that separation the data will be misleading.

From a Data Perspective

This post closes out Phase 2. Every domain across context, drivers, and outcomes has now been preprocessed and organized into the model-ready dataset. The real question preprocessing is meant to answer is not what data is available, but what data is actually worth using.

One step worth noting is the application of survey weights. NHANES is a complex survey designed to represent the broader U.S. population rather than a simple random sample, and the final dataset carries four survey design variables as pipeline infrastructure. Without them the model would be drawing conclusions from a sample rather than the population it is meant to represent.

From here the series moves into modeling and production across the next several updates:

• Update #9: Pruning, where the variable set gets trimmed to what the data can actually support.

• Update #10: Stratification, where comparison groups get formally defined before any behavioral signal is evaluated.

• Update #11: App design and layout, where the model output starts becoming something a person can actually use.

The preprocessing phase has been slower than a typical series update cadence, and intentionally so for two reasons. The first is that a system designed to operate at this level of complexity requires genuine depth in each domain before any of it gets modeled. It is better to understand each domain before the model runs than to untangle bad assumptions after it does.

The second is that documenting each domain publicly while actively building it forced a level of clarity that pure development work rarely demands. When you have to explain why a variable belongs in the model or why one was excluded, you either know the answer or you find out quickly that you don’t. That discipline carried into the modeling work and made the decisions that followed cleaner and more defensible.

The pace picks up from here and the next few updates will reflect it. Grateful as always for the support along the way.

One of them also has diabetes.

IBD management often limits high-fiber, hard-to-digest foods and flags raw vegetables, whole grains, and legumes as common triggers. Diabetes management often leans on those same foods to stabilize blood sugar, improve satiety, and support metabolic health. For the person managing both, the guidance does not just fall short. It points in two directions at once.

This is not a rare case. Conflicts like this are common, under-discussed, and quietly absorbed by the millions of patients living inside them. Neither doctor is wrong, and neither guideline is inherently flawed. They are simply products of a system designed to treat one thing at a time, in a world where that is rarely how illness actually works.

The Single-Condition Illusion

Most people with a chronic condition are not managing just one. They are navigating several at once, receiving separate guidance for each, and trying to make it all work in the same body.

Source: CDC

Over 51% of American adults are managing two or more chronic conditions. Among older age groups, the rate of managing two or more conditions nearly matches the rate of having any condition at all. At that scale, isolated treatment protocols are not just incomplete. They are the wrong tool entirely.

The research meant to close that gap has its own problem. A PLOS One study reviewing 161 clinical trials found that in more than half of them, researchers did not record whether their participants had other conditions alongside the one being studied. Not whether those conditions changed the results. Just whether they were there at all. In many trials, patients with additional conditions were screened out before the study even began.

The problem is not only that guidelines are condition-specific. It is that the research underneath those guidelines often removes the very complexity that defines real chronic illness. Patients are left trying to reconcile advice that was never fully tested for lives like theirs. A system designed around the exception has spent decades treating it like the rule.

Where Patients Started Going

As chronic disease prevalence grew and patient dissatisfaction compounded, a parallel market was quietly building. Patients who felt unseen by conventional care began seeking out a model that treated the whole person rather than the presenting condition. Functional medicine stepped into that opening.

Source: Grand View Research

The Complementary and Alternative Medicine Market is projected to grow in the U.S. from $66.7 billion in 2026 to $375.5 billion by 2033, with a compound annual growth rate of nearly 28%. That is not a niche trend finding its moment. It is a reallocation of healthcare spending at a scale that reflects a deeper shift in the kind of care patients are looking for.

NIH research tracking complementary health use found adoption nearly doubled from 19.2% in 2002 to 36.7% in 2022, driven by patients seeking something conventional care was not offering. The reasons were consistent across every survey:

• Root cause care over symptom management

• Guidance personalized to their conditions and circumstances

• Lifestyle treated as seriously as lab results

Functional medicine became the most organized expression of that demand. Longer appointments, individualized protocols, and a definition of health that extended beyond what a blood panel could capture. For patients who had spent years being told their labs looked fine while feeling anything but, that reframing did not just feel different. It felt like the first honest conversation they had been offered.

A Closer Look: The Cleveland Clinic Study

Functional medicine is still a relatively young field, which makes the research behind it worth examining carefully. The most rigorous test of the model to date came out of the Cleveland Clinic in 2019, published in JAMA Network Open.

Researchers compared 1,595 patients treated at their Center for Functional Medicine against 5,657 patients in standard primary care. They measured outcomes across energy, pain, digestion, physical function, and emotional wellbeing using PROMIS, an NIH-validated assessment tool.

The results were surprising to some. Functional medicine patients were over 40% more likely to hit a clinically meaningful improvement threshold than their primary care counterparts at six months.

No single study settles a debate this complex, but this one was particularly hard to dismiss. Large sample size, NIH-validated outcome measures, a real primary care comparison group, and results that held up over time. It is the kind of institutional, peer-reviewed evidence that functional medicine has often been criticized for lacking. Then Cleveland Clinic provided it.

The Accessibility Issue

While functional medicine has shown tremendous promise early on, cost remains the largest barrier the field has yet to solve. This is the natural consequence of delivering highly personalized care through a system built for volume. Longer appointments and individualized protocols are exactly what makes functional medicine effective and exactly what makes it expensive.

Initial consultations typically run $300 to $500, with lab panels and diagnostic testing easily pushing total first-year costs into the thousands. Insurance rarely covers any of it, for a litany of reasons worth their own conversation. This means the people carrying the heaviest chronic disease burden are often times the ones least positioned to absorb what insurance will not.

Solving the access problem requires rethinking how personalized guidance gets delivered in the first place. The insight at the center of functional medicine, that recommendations should be built around who someone actually is, does not have to be gated behind an expensive clinical visit. The right data about a person and enough population-level research to find people who genuinely resemble them across age, medical history, and lifestyle context can surface the behavioral changes most likely to matter for that specific individual. This is guidance derived from real outcome patterns in real people with real complexity.

Lab results, blood panels, and microbiome testing do not lose their value in that model. They just stop being the price of admission. A tool that works without them is accessible to everyone. A tool that can incorporate them when available becomes more precise for the people who have them. Start with what someone can provide and get more specific as more information comes in. That is how you serve the full spectrum of people who need this kind of guidance, not just the ones who can already afford to know their biomarkers.

Takeaways

The numbers behind functional medicine's rise tell a story about unmet need more than they tell a story about alternative health. Here is what the research actually shows.

• Over half of American adults manage two or more chronic conditions simultaneously, yet the clinical guidelines meant to help them were built around one condition at a time.

• Complementary health adoption nearly doubled over two decades, driven by patients seeking care that conventional medicine could not keep up with.

• The Cleveland Clinic's 2019 study found functional medicine patients were over 40% more likely to reach meaningful improvement than primary care patients at six months.

• At $300 to $500 per visit with lab costs that can easily exceed thousands, personalized care remains a luxury for the patients carrying the heaviest burden.

Functional medicine identified the right problem: chronic care needs to account for the full person, not just the diagnosis in front of them. The work that remains is making that insight available to everyone, not just the patients who can afford to seek it out. The demand has been there for a long time. What is still being built is a way to meet it.

That is what the context layer is for.

Context refers to the parts of a person’s profile that are fixed, slow-moving, or not treated as direct behavioral levers, but still meaningfully shape who they should be compared with. It is the baseline foundation that comes before behavior.

For this stage of the FPI, the context layer is organized around three core areas that are supported by research literature and available within the NHANES scope: basic information, medical history, and socioeconomic status. Together, these categories help define the comparison group before the system moves into the drivers people can actually change.

Basic Information

The first group is the most straightforward. It is mainly split between demographics such as age and gender, and body context parameters such as height and weight. These variables are simple to interpret, easy for a user to answer, and explain meaningful differences in how functional outcomes show up across people.

Age is the clearest example. It does not determine someone’s functional health, but it changes how the rest of the profile should be read. A response that looks unusual in one age group may be much more common in another, and a fair comparison requires accounting for that before interpreting deeper health or behavioral patterns.

When you look across an entire population, small shifts like these can still reflect real baseline separation between groups. That separation also tends to compound once behavioral drivers enter the picture, since age shapes how the body responds to the same inputs differently.

Weight is a more nuanced parameter. It is not the target outcome of the FPI, and treating it that way would pull the system away from functional health. At the same time, weight still carries contextual information that may matter for how people experience energy, physical demand, and overall health. It is one of those variables that can be easy to overfocus on in health products, but also wrong to ignore entirely.

These kinds of basic inputs are low-effort and high-value signals. They are not the most complex part of the system, but they help set up cleaner stratification later on by making sure people are compared against groups that better reflect their starting point.

Medical History

The second group moves into territory that requires a bit more care. While basic information helps define someone’s general profile, medical history adds a stronger signal of what their body may already be managing. Someone with a cardiovascular background and someone dealing primarily with gut health issues have different nutritional needs, different risk profiles, and different bodies to design guidance around.

NHANES captures medical history in two ways. The survey collects overall prescription medication count alongside a detailed inventory of specific diagnosed conditions. Grouping participants by total medication and condition burden creates a reliable proxy for overall medical load and is essential to the first layer of stratification. From there, the condition-level data is preserved for future analysis to allow the model to go further, placing individuals within more specific clusters that better reflect the unique biological context shaping their outcomes.

The visual below takes a broad look at how medical conditions are distributed across the U.S. population.

More than half of American adults live with at least one chronic condition, and a meaningful share are managing several simultaneously. That spread is what makes this dataset so well suited for the kind of layered analysis the FPI is built around. There is enough variation across the medical burden spectrum to build reliable comparison groups at multiple levels, not just at the extremes.

At a broader level, this section ties directly into the core thesis of the FPI and much of what Digestive Data has been building toward. People with different chronic conditions and medical backgrounds should be receiving different behavioral guidance to navigate their health.

Socioeconomic

The third area is the one most commonly left out of health tools entirely. Socioeconomic context does not fit neatly into a behavioral framework, so it often gets treated as outside the scope of the model. In reality, it impacts someone’s access to quality food, time available in a day, the realistic ability to prioritize sleep or exercise. These are not gaps in behavioral discipline. These are real-life constraints.

This layer includes variables most people would immediately recognize: insurance coverage, hours worked, income level, and food security status. The last one most directly pertains to the constraints mentioned above and tends to capture what the others collectively produce, which is why it is the focus of the visual below.

For every outcome, the pattern is consistent. Food secure participants score higher on overall health, energy, mood, focus, and stress management, with several gaps exceeding a full point. These are not marginal differences at a population-level. They are a result of the circumstances people are operating within daily, and a model that ignores them will consistently misread the people most shaped by them.

Socioeconomic status is included in the context layer for the same reason medical history is. Not because the FPI is designed to change it, but because pretending it is not there produces a less honest model. The goal has always been to meet people where they actually are, and that requires knowing something about the conditions they are operating within before evaluating anything else.

Biomarkers as Optional Enrichment

A previous update covered the biological backbone of the FPI, where lab markers were grouped into physiological axes like inflammatory, metabolic, hepatic, kidney, lipid, and hematologic domains. Those markers still matter, but they serve a slightly different role than the context variables discussed here.

In the app, biomarkers will function as context enrichment rather than the starting point. Lab panels can improve the quality of someone’s profile when available, but they are also expensive, unevenly accessible, and not something every user can be expected to provide. The first version of the FPI should still be useful without requiring bloodwork.

The FPI is being built around functional outcomes first: what people feel, experience, and want to improve in daily life. Biomarkers can strengthen the profile and add biological depth over time, but they should support the functional model rather than replace it.

From a Data Perspective

Context variables are an essential layer of the FPI and are fundamentally handled differently than the rest of the model. They serve as confounders, effect modifiers, and stratifiers, which means their job is to define who someone is being compared against before any behavioral signal is evaluated.

Because of that role, most of these variables are passed through with minimal transformation. Raw NHANES coding is preserved where possible, sentinel values are cleaned to NaN, and sparsity decisions are made on a case by case basis depending on whether a variable can meaningfully contribute to a comparison group. Some low count variables were dropped. Others, like specific liver condition subtypes, were kept at this stage despite sparse coverage because the biological context they carry is too relevant to exclude.

The next post wraps up preprocessing entirely. From there the work shifts into actual model building, where many of these variables will be pruned down to what the data can reliably support. The process has been thorough by design, but the model that comes out the other side will be leaner than what went in. As always, I appreciate your patience and support with this build.

But the science underneath those experiences is real, and it is more structurally complex than it often gets credit for. A growing number of researchers are now using AI to model exactly what is happening between your gut and your brain, and the early findings point toward something the wellness world has been circling for years without the data to back it up.

The AI Black Box Problem

Most explanations of the gut-brain axis start and end with the vagus nerve. Stress hits, your gut reacts. That’s true, but it’s also incomplete.

The communication runs in both directions at once, across neural, hormonal, and immune pathways simultaneously. The gut isn’t just responding to the brain. It’s actively shaping it through the enteric nervous system, a network of over 100 million neurons operating in parallel with your central nervous system.

That bidirectionality is what makes it so hard to study. Traditional research is built for linear systems: one variable changes, one outcome follows. But the gut-brain axis doesn’t behave like that. You can’t isolate a single input without affecting everything else, and there’s no stable baseline to measure against because the system is always adapting.

Source: The AI Black Box

Researchers often refer to this as a black box problem. You can observe the inputs. You can observe the outputs. But the mechanism connecting them remains largely invisible.

How Explainability Changes the Equation

This is where AI, specifically explainable AI, becomes useful. Most models in health output predictions without context. Explainable AI goes a step further by identifying why a prediction was made, identifying the specific biological features driving an outcome.

Source: IBM: How Does Explainable AI Work?

In a system as interconnected as the gut-brain axis, that is an essential distinction. Instead of just predicting a change in mood or cognition, these models can begin to map which microbial patterns, metabolites, or inflammatory signals are most strongly associated with those shifts. It’s not a complete picture yet, but it’s the first time the middle layer of this system is becoming partially visible. And early results are surprisingly consistent.

What the Early Data Shows

The research is still early, but across different populations and methodologies, the microbiome is consistently associated with both mood and cognitive performance. Here is what three recent studies found.

• A 2024 study in Psychiatry and Clinical Neurosciences used machine learning to correlate specific gut bacterial families with structural brain changes. The bacteria they flagged were linked to the exact brain regions tied to depression and attention deficit (PCN).

• A team at Fudan University trained a model on microbiome data from patients across stages of Alzheimer’s disease, reaching 80% accuracy in identifying disease stage using gut data alone. The microbes driving those predictions were tied to energy, mood regulation, and neurotransmitter production (Nature).

• Researchers are now building “gut-brain digital twins,” combining real-time behavioral and biological data to model how gut state influences cognitive and emotional outcomes in real time (University of Chicago).

Different approaches and conditions. Same directional finding. The microbiome isn’t sending occasional distress signals. It’s continuously shaping how you think, feel, and perform.

Source: Nature

The image above shows how changes in the gut microbiome move through the body. Signals like metabolites and neurotransmitters travel through systems such as the immune system and vagus nerve, eventually influencing mood, cognition, and physical performance. What AI is starting to map is how these internal pathways connect gut changes to real-world outcomes.

So if we’re able to begin tracing the system, what is actually holding this back from being useful in everyday life?

The Real Bottleneck Isn’t Science

The limiting factors right now are cost and relevance. Comprehensive microbiome testing runs $120 to $400 out of pocket, is rarely covered by insurance, and returns broad compositional data that has no clear connection to the functional outcomes most people actually care about.

Source: Cost Comparison Chart

The research that matters for energy, focus, and cognitive performance is pointing less toward broad diversity and more toward specific bacterial families. These are often directly tied to neurotransmitter production, metabolite output, and inflammatory signaling. A personalized panel built around those targets would be:

1. Cheaper

2. More actionable

3. Directly tied to what people experience

What is also missing from most of this conversation is the framing. Microbiome composition is not a lever you can pull. In reality, it’s a biological baseline that influences how your body responds to everything else, but not something to directly optimize in isolation. You don’t improve your microbiome directly. You improve the behaviors that shape it over time.

That is where the real potential lies. Not another thing to optimize, but an enrichment layer that helps stratify people more accurately and understand which behaviors actually move the needle for which biological profiles. The tools built to use it that way, behavioral foundation first, biological context second, are the ones best positioned to lead in proactive health solutions when affordable, outcome-specific testing finally makes that possible at scale.

Takeaways

AI has earned a fair amount of skepticism in health, and most of it is justified. Tools built on pattern recognition are not diagnosticians. They cannot replace the judgment of a clinician or the nuance of a patient-provider relationship. But what they can do is narrow the distance between population-level research and individual application, informing more personalized health plans from the treatments someone is most likely to respond to, to the behaviors that matter most for their specific biological profile.

Here are a few of the key points from this article:

• The gut-brain axis is not a one-way stress response. It is a continuous bidirectional network running across neural, hormonal, and immune pathways simultaneously. What happens in one system shapes what is possible in the other.

• The machine learning research is still early but the pattern is hard to ignore: your microbiome has a measurable and consistent influence on how you think, feel, and function.

• The gap between that research and personal application is mostly a cost and relevance problem, not a scientific one. The tools closing that distance fastest are the ones building behavioral infrastructure first and designing it to absorb biological context as it becomes accessible.

Some of the most important questions in health went unanswered for decades not because the science was wrong but because the data and supporting technology was not ready. The gut-brain axis is one of them. That’s changing now. Not because the biology is new, but because we finally have the tools to start making sense of it.

Chronic disease often gets discussed in clinical terms. But what most people with a persistent condition actually want is simpler and harder to measure. To feel better. To have more energy. To think more clearly and carry less anxiety. The core foundation of the FPI is that many patients experience similar functional challenges across a wide array of unique chronic conditions. Different diseases, often similar daily burden. That shared functional cost is where a population-based framework can find personalized behavioral direction even when the underlying condition differs.

If you have been reading these updates, you may have noticed the name change from DSI to FPI. While my health story was specific to digestive health, this tool was always meant for anyone managing chronic disease. Since NHANES is far richer in nutrition and lifestyle data than GI-specific measures, the system has more depth at the behavioral level than anywhere else. Functional performance is what this project is actually measuring, and the name now reflects that.

The Raw Candidate Pool

Before any filtering, the preprocessing script assembled every NHANES variable that could plausibly capture how poor health is experienced in daily life. These variables span depression screeners, functional limitation surveys, self-rated general health, body composition, a physical balance exam, and respiratory modules. There were dozens of candidates, and many of them were engineered into composites before any selection decisions were made.

This step was primarily about preserving all possible signal before condensing and making cuts. Anxiety severity was built by combining how often someone experiences it with how intense it is when it occurs. Cognitive limitation was blended from two different survey instruments. Fall burden was folded from three separate measures.

The objective was to build everything correctly before deciding what to use, leaving 33 possible model outputs after initial preprocessing. While only a handful will actually end up being used after filtering, variables must be handled carefully as those collapsed carelessly at this stage cannot be recovered.

Three Core Decisions in the Filtering Process

Ultimately, the end goal for this product is an app where someone can select from a handful of outcomes and get personalized guidance based off who they are. With this in mind, the outcomes need to be narrowed down to those that are both representative of real-world symptoms and have meaningful variance in responses.

The filtering process involved many small decisions and a few larger ones. These three shaped the outcome layer most.

1. Body composition was reclassified as context. Weight, BMI, and waist circumference matter to daily function, but they sit at a different level of the system than the outcomes the FPI targets. They reflect years of accumulated behavior, genetics, and medical history as much as anything happening right now. Using them as outcomes would let the model predict body frame rather than behavioral change, which strays from the core premise of the FPI.

2. Physical limitation was removed as a primary outcome. Over 80% of the 18-64 population reported no physical limitation, producing a floor effect with no usable variance. A variable where nearly everyone scores at zero has nothing to model. The chronic disease experience absolutely includes physical function, but forcing signal out of a floor effect will produce useless predictions.

3. Certain terms were collapsed into composites. One example is the PHQ-9 appetite item which asks about "poor appetite or overeating". While this question is clinically relevant to depression as it directly covers dietary disruption, it cannot be modeled on its own. This is because the same score can mean two opposite behaviors, with no coherent direction for a recommendation to point. As a result, the information is preserved through a depression composite rather than abandoned entirely.

The Five Core Outcomes

After filtering, five outcomes remained. These fundamentally capture different aspects of what people experience, have varying interactions with daily behaviors, and have a rich variance in response necessary for modeling.

1. Overall Health. Drawn from a single NHANES question asking participants to rate their general health on a five-point scale. While the most general, it is one of the strongest predictors of future hospitalization in the research literature and has near-full population coverage.

2. Energy Levels. Comes from one PHQ-9 item capturing how often someone has felt tired or had little energy over the past two weeks. Scored and inverted so higher reflects more energy. Fatigue is also one of the most universally reported symptoms across chronic conditions.

3. Stress & Anxiety. Built from two functional limitation survey items capturing how often someone experiences anxiety and how severe it is. The two items are z-scored and averaged into a single composite. This is kept separate from mental resilience because the two respond to meaningfully different behavioral inputs.

4. Focus & Cognition. Blends a functional limitation item on difficulty concentrating with a PHQ-9 cognition item, averaged when both are available and falling back to one when only one is present.

5. Mental Resilience. Uses the full PHQ-9 depression score sum across nine symptom dimensions, ranging from zero to twenty-seven. Scored and inverted so that higher reflects greater resilience. The total is more sensitive to gradual behavioral change than any individual item.

The distributions above reflect each outcome across the full 18-64 working-age population. Mental Resilience in particular is visibly right-skewed, which is expected in a generally healthy working-age population. Non-normality across these outcomes was handled carefully, and the gradient boosting models used in the FPI are robust to it by design.

The larger takeaway is what these five outcomes represent: the pillars that every behavioral recommendation in the system will be built on.

Do These Five Outcomes Overlap?

Naturally, five outcomes measuring different dimensions of someone’s health are going to share some common ground. The matrix below shows exactly how much.

Pairwise correlations across the working-age population range from weak to relatively strong, spanning 0.22 to 0.66 depending on the pairing. This reflects the biological reality that a person struggling with their overall health is more likely to also struggle with energy, depression, and cognition.

The more important point is behavioral. Each of these outcomes responds differently to the same drivers. Sleep timing associates more strongly with anxiety than other outcomes. Vitamins and supplements affect energy more heavily than cognition. Five separate models will surface those distinctions. One composite score would average them away, producing recommendations that are broader and less useful to the people relying on them.

From a Data Perspective

Defining these five outcomes required balancing two things that do not always align: what people actually experience living with chronic disease every day, and what a large population dataset can reliably measure and model. The goal was never to capture every dimension of poor health, but to identify the outcomes that are both meaningful to the people using this system and supported by enough signal in the data to generate trustworthy, personalized recommendations.

The next two posts close out Phase 2. One covers the remaining behavioral driver domains: sleep, physical activity, and substance use. The other covers the context layer that will anchor the comparison groups: demographics, medical history, and socioeconomic status.

The end of preprocessing is close. Thank you for your patience with this build.

The science behind the gut-brain axis tells a more connected story. Circadian rhythms, microbial balance, and the hormones that regulate both run on shared infrastructure. Disruption in one rarely stays contained to the other.

How Your Gut Health Impacts Sleep

The common assumption is that melatonin is a brain hormone. The pineal gland releases it, the brain regulates it, and sleep follows. That picture is accurate, but it starts in the wrong place.

Over 90 percent of the body’s serotonin is synthesized in the gut, with gut bacteria serving as major producers (AAAS). That matters because serotonin is the direct precursor to melatonin. Before the pineal gland releases anything, the gut has already played a central role in whether the raw material is there to work with.

The chain starts with tryptophan, an amino acid that gut bacteria help convert into serotonin, which then feeds melatonin production. Specific gut bacteria enhance sleep through their role in serotonin and GABA production, two of the primary chemical signals the brain uses to initiate and sustain sleep (NIH). When those bacterial populations are depleted or out of balance, the effects travel up that chain.

Source: IntegrativeVision

The gut-brain axis shapes circadian rhythms and the brain's response to melatonin, and research has found that this extends to measurable differences in sleep quality tied directly to microbiome diversity (NIH). Lower bacterial diversity is consistently linked to worse outcomes. The gut is not simply reacting to how well someone slept the night before, but rather shaping the conditions that make sleep possible.

What Poor Sleep Does to the Gut Microbiome

The relationship runs the other direction just as clearly, and in some ways more immediately.

One of the most overlooked ways poor sleep impacts health is through what it does to the microbiome. Poor sleep depletes the beneficial bacteria responsible for regulating how the body processes food and manages hunger. For appetite, the effects are significant on both ends.

Sleep restriction has been associated with an 18 percent decrease in leptin, the hormone that signals fullness, alongside a 28 percent increase in ghrelin, the hormone that triggers hunger (Annals of Internal Medicine). What feels like a simple craving after a bad night of sleep is, at least in part, a microbial and hormonal chain reaction set off well before breakfast.

Sleep timing matters too, not just sleep length. The gut runs on a biological clock of its own, and when sleep patterns are erratic, that clock loses its rhythm. Sleep deprivation disturbs the balance of gut bacteria, leading to shifts in microbial composition and function (ScienceDirect). The bacterial populations that keep the gut lining strong and inflammation low tend to be the first ones to take a hit.

The result is a system that becomes progressively more reactive. Symptoms that might have been manageable on a good night of sleep feel amplified after a bad one, not because of anything that changed in the diet, but because the gut itself is operating in a more compromised state.

GERD and the Cycle That Sustains Itself

GERD affects roughly 14 percent of adults globally at least once a week, making it one of the most common chronic digestive conditions (NIH). Most people think of it as a diet problem. Eat the wrong thing before bed, wake up with heartburn. The fix seems obvious. But the research paints a more complicated picture, and it is one of the clearest documented examples of what happens when the sleep-gut loop is ignored.

A meta-analysis of 22 studies found that insomnia, sleep disturbance, or short sleep duration significantly increased the risk of GERD, with odds ratios ranging from 1.98 to 2.66 depending on the sleep variable examined (PeerJ).

In plain terms, poor sleep roughly doubles the odds of developing GERD symptoms. Sleep deprivation heightens the esophagus's sensitivity to acid, meaning the same amount of reflux that would go unnoticed on a good night registers as pain after a bad one. This means the gut becomes more reactive, not because more acid is being produced, but because the threshold for feeling it drops.

Many GERD patients report nighttime awakenings from heartburn, and that lost sleep makes the esophagus more sensitive the next night. This creates a frustrating cycle where poor sleep worsens reflux symptoms, and worsening symptoms make restful sleep harder to come by.

GERD is a useful example precisely because the data is so clear in both directions. But the pattern it describes is not unique to reflux. The same loop shows up wherever chronic symptoms and disrupted sleep coexist. The condition becomes harder to manage, sleep becomes harder to protect, and the two problems quietly compound.

Takeaways

The gut and the brain are not operating independently. They share chemistry, timing, and a communication network that runs continuously in both directions. What happens in one system shapes what is possible in the other.

• The gut produces over 90 percent of the body's serotonin, the direct precursor to melatonin, giving it real chemical influence over the conditions that make sleep possible.

• Poor sleep depletes the microbiome responsible for appetite regulation and gut barrier integrity, driving hunger hormone shifts that are largely physiological.

• Sleep problems roughly double the odds of developing GERD symptoms, and GERD disrupts sleep in return, creating a vicious cycle for recovery.

Most conventional advice treats these as separate problems with separate solutions. That isolation is part of why so many people manage one without ever fully resolving the other. Biology does not make the same distinction.

The body does not compartmentalize the way traditional guidance does. Poor sleep degrades the gut's ability to function, and a gut that is struggling makes reliable sleep harder to sustain. The circular loop is not a metaphor. It shows up in the data, and it shows up in how people feel day to day without ever knowing why.

Nutrition in NHANES goes considerably further than most datasets allow. Over 300 raw variables across dietary intake, supplement behavior, and individual food items were collected to provide a complete picture of someone's nutritional profile. The goal was not to keep them in their original form, but to engineer a more deliberate set of signals that reflect how food and eating actually show up in daily life.

This update covers the seven subcategories within the nutritional layer, the decisions made to build each one, and what they capture together.

Nutrient Intake

The most fundamental data within this layer comes from dietary intake. Across two days of recall, participants were tracked across roughly 70 specific variables capturing absolute consumption. These covered everything from macronutrients and fiber to individual vitamins, minerals, and fatty acid subtypes.

To make this information useful, intake was averaged across both recall days and standardized for modeling. That process also produced a measure of dietary consistency across days, which serves as a confidence signal in the model. Participants whose intake was more stable across both recall days have a more reliable relationship between what they reported eating and the outcomes FPI is designed to predict.

Below is a look at how several of those nutrient distributions actually sit across the population.

After standardization, most variables distributed close to normal. The one notable pattern was directional: macronutrients tended toward a slight right skew while vitamins and minerals often leaned left. For macros, the average is inflated by a subset eating significantly more than most, making typical intake look higher than it is. On the contrary, most people have enough vitamins and minerals in their diet but those who fall short tend to fall well short.

Theobromine is a good example of how this can break down for less common nutrients. Its distribution is visibly bimodal, with two distinct peaks driven almost entirely by whether someone consumed chocolate or cocoa on a given day. For nutrients this specific, the shape of the population distribution is not about gradual dietary variation. It is about whether one or two particular foods showed up at all.

Density & Dietary Ratios

While raw intake captures what entered the system, it doesn’t account for whether that intake was appropriate relative to how much someone was eating overall. Calculating density measurements was essential for making meaningful comparisons across people with very different caloric baselines.

To address this, each nutrient was normalized per 1,000 kilocalories consumed. A person eating 3,000 calories a day and a person eating 1,600 may report similar raw totals, but their nutritional efficiency looks very different once intake is scaled to consumption. Each density variable was derived independently and added alongside its raw counterpart, giving the system two distinct lenses on the same nutrient.

Beyond individual density variables, a set of ratios between key nutrients was created to capture broader dietary patterns rather than any single measurement in isolation. Combinations like protein-to-fiber and carb-to-sugar reveal things neither variable shows alone. Someone could have perfectly adequate protein intake, but if their fiber is consistently low relative to it, that imbalance may matter more to their outcomes than any single nutrient in isolation.

Together, density variables and ratios move the nutritional layer beyond simple accounting. They capture not just what someone consumed, but how balanced and efficient that consumption actually was.

Food Groups

The first two sections focused on sheer and relative intake. This one looks at the specific foods behind those numbers.

Five groups were selected based on their established relevance to chronic disease and systemic health. Each food item reported across both recall days was mapped to a group using USDA food codes, aggregated into daily gram totals, and averaged across days to produce a consistent measure for each participant.

• Fruits and vegetables: apples, berries, citrus, broccoli, spinach, carrots, tomatoes, potatoes, and fruit juices

• Legumes and beans: beans, peas, lentils, soy products, and meat alternatives

• Whole grains: oatmeal and low-sugar ready-to-eat cereals

• Red and processed meat: beef, pork, lamb, organ meats, bacon, sausages, and cold cuts

• Ultra-processed foods: chips, cookies, pastries, candy, soft drinks, energy drinks, pizza, and fast food sandwiches

What stands out immediately is how unevenly these groups are represented across what we eat.

Ultra-processed foods appeared in 92% of diets, a stark reflection of how deeply convenience has shaped the way most people eat. While fruits and vegetables held strong, other health-supportive categories like whole grains and legumes rarely made it to the dinner table. The contrast across groups is exactly the kind of variability that will allow the model to find meaningful differences between profiles.

These five categories are among the most actionable variables in the system. Unlike nutrients which require interpretation, food groups map directly to decisions people make every day. The aim is to understand where someone's food patterns diverge from healthier profiles and use that to surface the dietary adjustments most likely to matter for people like them.

Meal Time & Sourcing

Most nutritional datasets capture what people eat. NHANES also captured when they ate, how long they were eating across the day, and where their food came from. That behavioral layer adds a dimension raw nutrient data cannot provide on its own.

Timing variables included the time of first and last eating occasion, total eating window, number of eating occasions, and meal occasion proportions across breakfast, lunch, dinner, and snack. Research has increasingly shown that when someone eats matters to metabolic health and daily function independently of what they consume, making these variables a necessary addition to the nutritional picture. The chart below captures what that looks like at the population scale.

The distance between first and last meal is particularly informative. It can reflect anything from intermittent fasting to late-night snacking habits that slow down overnight recovery. Identifying where someone falls within that range changes how their nutritional profile should be read and what recommendations are actually relevant to them.

Each food item was also tagged by where it came from, spanning grocery stores, fast food, restaurants, and food assistance programs. That sourcing signal, paired with timing, shifts this sublayer from a nutritional snapshot into something closer to a full picture of how someone actually eats day to day.

Supplement Use

To round out the nutritional layer, supplement use was tracked separately from dietary intake. Since the body does not absorb nutrients the same way from a pill as it does from food, that distinction is worth preserving for future modeling.

Prevalence varied widely across supplements. Vitamin D was the most commonly reported at 45%, while several others like lutein, phosphorus, and potassium sat below 15%. That sparsity created a modeling challenge. For nutrients used by fewer than ten percent of the population, continuous variables were encoded as binary flags instead to capture use versus non-use rather than amount.

More than half of participants reported taking at least one supplement in the past 30 days, and among those who did, the average user was taking nearly four. That pattern suggests a population that is actively, and in some cases aggressively, trying to manage their nutritional intake outside of diet alone.

Capturing supplements alongside dietary intake means the model has access to total nutritional exposure rather than just what showed up on someone’s plate.

From a Data Perspective

The nutritional layer is the most variable-dense sublayer in the FPI, and the one that required the most deliberate engineering. After preprocessing, it produced 183 variables, nearly half of all possible features in the model. That density reflects how much signal NHANES captured, and is a large part of why this sublayer warranted its own dedicated post.

Some of these variable decisions came from well established research and common nutritional practices, such as nutrient density measurements and food group classification. Others capture things that are often overlooked in standard nutritional analysis, like meal timing, dietary consistency, and food sourcing. Together they aim to create a grounded foundation while leaving room for more nuanced recommendations as the system develops.

The next update rounds out the driver layer by covering the remaining three domains: substance use, sleep, and physical activity. Within the next few posts, the foundation will be complete and the system will move from being built to being run. As always, I appreciate the patience and support as the FPI continues to take shape.

The economic conversation around digestive health tends to focus on the obvious: hospitalizations, surgeries, medications, specialist visits. While those costs are real and large, they only represent the portion of the problem that made it into a medical record. The larger share never gets counted at all.

When function declines gradually and discomfort becomes the background noise of a workday, nothing gets flagged. No diagnosis is made. No cost gets attributed. The person just performs at a fraction of their capacity, day after day, while the system registers nothing unusual.

That is the hidden cost. And it is far larger than what the healthcare data shows.

A $111 Billion Undercount

Gastrointestinal diseases are among the most costly and widespread conditions in the United States. In 2021, the NIH reported that GI issues were responsible for:

• Over $111 billion in healthcare expenditures.

• 14.5 million emergency department visits.

• Nearly 3 million hospital admissions.

Yet those numbers still only capture part of the picture. They reflect the cases that crossed a diagnostic threshold. The majority of people experiencing frequent GI symptoms never make it into those counts at all.

Source: UNC School of Medicine, Rome IV FBD Study

Of the 60 percent of US adults who report frequent GI symptoms in any given three-month window, only about 25 percent meet formal criteria for a functional GI disorder. The rest have no diagnosis, no clinical pathway, and no framework for understanding what is happening to them. They are not in the data. They are just managing.

The $111 billion captures what the system had to respond to. It ignores the larger population managing symptoms below that line, showing up every day without ever being formally diagnosed. One of the clearest places that gap presents itself is at work.

IBS as a Case Study: Present But Not There

The standard way to measure how illness affects work is to count days missed. It is a clean metric, easy to track, and easy to report. It’s also the wrong one.

Absenteeism captures the moments when someone does not show up. It says nothing about the far more common experience of showing up and operating at a fraction of normal capacity. That is presenteeism, and in the context of chronic illness, it is consistently the larger cost. IBS offers one of the clearest case studies available.

Source: American Journal of Medical Care

Among employees with IBS, the cost of showing up impaired is roughly eleven times larger than the cost of not showing up at all. The most surprising part? That ratio is not unique to IBS. The same research showed comparable impairment levels show up across GERD and other chronic conditions, suggesting the pattern is less about calling in sick and more about what happens when chronic symptoms become a background condition of the workday.

Moreover, workers with IBS-D report a statistically significant difference in overall productivity loss compared to healthy controls, amounting to an estimated $2,486 more in annual indirect costs per person (NIH). These indirect costs cover the economic value of lost or impaired output, including the hours worked at reduced capacity, the tasks that took longer, and the decisions made through fatigue.

These costs may not appear in a billing record, but they show up in the quality of work produced. For people managing chronic GI symptoms, that cost compounds in ways that go beyond productivity alone. What those symptoms do to mental health, and what mental health does back to the gut, is a cycle that the economic data has only begun to account for.

The Mental Health Tax

Productivity loss captures what chronic symptoms cost from an employer standpoint, but it only tells part of the story. Mental health is one of the most significant drivers of workplace performance, and for people managing chronic conditions, the two are rarely separate.

Sources: Anxiety Global Estimate, Depression Global Estimate, IBD & Mental Health, Diabetes & Anxiety, Diabetes & Depression, Chronic Pain & Mental Health

The conditions on the chart above share almost nothing in common. They affect different organs, have entirely different treatments, and cause different lived experiences. Yet across all three conditions, anxiety and depression prevalence runs anywhere from five to ten times higher than the general population baseline. The common thread is not the diagnosis. It is the experience of living with chronic burden.

It’s worth noting that these figures come from clinical populations, so people managing milder symptoms outside the healthcare system may experience lower rates. The evidence suggests a gradual relationship, where mental health tracks symptom burden across the full range rather than only at the severe end.

For many conditions, this relationship tends to run bidirectionally. Research in IBD has found that having anxiety or depression roughly doubles the odds of adverse disease outcomes, including flare, steroid resistance, and the need for therapy escalation (NIH). The research on this specific mechanism is largely concentrated in IBD. But the pattern it describes, physical symptoms feeding psychological ones and vice versa, is something people living with chronic conditions tend to arrive at on their own. Mental health burden is not only a response to living with a difficult condition. It is, at least in part, a physiological output of the condition itself.

The Diagnostic Gap

The costs described in this article do not begin at diagnosis. They begin at symptom onset. For those with persistent GI symptoms, those two moments are often separated by years.

As mentioned earlier, many people managing chronic GI conditions like IBS are never formally recognized at all. But even for those who are eventually diagnosed, more than 40 percent waited at least five years for it (NIH). In the meantime, many cycle through unnecessary tests, procedures, and a healthcare system that keeps looking for something it was never designed to find.

While IBS provides one of the clearest examples of the gap between what people feel and what gets formally diagnosed, most chronic conditions operate in a similar way. That delay is not a bug. Our healthcare system was built to detect disease, not to measure function. IBS has no biomarker, no imaging finding, no lab value that confirms it. Conditions like this are only diagnosed through exclusion, leaving many patients in the dark for years on how to handle their symptoms.

A person can be clinically normal on paper and still be losing meaningful functional capacity every day. Their productivity is affected, their mental health is affected, their quality of life is affected. None of it appears in a billing record until something crosses a threshold severe enough to generate one. People with fewer resources, less access to specialist care, and less capacity to navigate a fragmented system wait longer and absorb more of the true cost. The diagnostic gap is not just a medical problem. It is an economic one, and it compounds differently depending on where you start.

Takeaways

Digestive health has a visibility problem. The costs are real, they are large, and most of them never make it into the data.

• The economic burden of GI conditions extends far beyond what healthcare data captures. The majority of costs never make it into a medical record.

• Presenteeism, not absenteeism, is the primary mechanism of workplace loss. Showing up impaired is consistently more expensive than not showing up at all.

• Anxiety and depression have a bidirectional relationship with chronic disease. Physical burden worsens mental health, and poor mental health worsens physical outcomes.

• A diagnosis can take years to arrive. For many people, the cost of managing symptoms starts long before it does.

The economic case for early identification is already well established in the research. What has been slower to develop is the infrastructure to act on it. There is no clinical pathway, no standardized framework, and no clear way to connect what people are experiencing to what their data might already be showing. That gap is both a problem and an opportunity that the healthcare system has largely left on the table.

FPI is being built with that in mind. Digestive health does not operate independently of sleep, activity, mental health, or metabolic function, and a tool that treats it as if it does will miss most of what actually matters. The aim is a framework oriented around functional outcomes, one that meets people where they are, whether they have a formal diagnosis or just a persistent sense that something is off.

The FPI is built on the premise that the biological picture and the lived experience of chronic disease need to be understood together. While symptoms and lab results may not always align, biomarkers provide an essential biological foundation for understanding how well the system is being tolerated at a structural level. They reflect physiological strain that may not yet rise to visible disease, but still shapes daily function over time.

This update focuses on one part of that picture: what biomarkers reveal when organized as a coordinated system rather than read in isolation.

From Raw Labs to Structured Axes

NHANES collected 93 biomarker variables for each participant, spanning inflammatory markers, metabolic panels, blood counts, renal indicators, liver enzymes, and lipid measures. While extensive, raw laboratory data requires careful preparation before it can produce anything meaningful.

The cleaning approach followed three core steps:

• Remove structural duplication. Highly correlated variables representing the same physiological signal in different units or formats were identified using pairwise correlations above 0.995. One representative marker was retained to remove redundancy while preserving each signal.

• Consolidate repeated measures. Blood pressure readings for systolic, diastolic, and pulse were collapsed into mean values to avoid future model overweighting.

• Normalize and standardize distributions. Skewed metabolic markers were log-transformed where necessary, and all variables were standardized to allow comparison across systems operating on different scales.

By the end of this process, 60 usable biomarkers remained. After removing duplicates and standardizing the remaining variables, the next step was feature engineering.

Rather than grouping markers by the laboratory panels in which they were collected, they were reorganized by the biological systems they most directly influence. ALT and AST, for example, were grouped within a hepatic axis because they move together as indicators of liver strain. Albumin and globulin were placed within a malabsorption context due to their link to protein balance and inflammatory stress.

This shift moves the focus from how tests are administered to what systems they reflect. Individual biomarker relationships remain intact, but they now sit within coordinated physiological dimensions instead of isolated lab panels. All 60 variables were grouped into eight system-level axes, shown below.

Across the eight dimensions, distribution patterns are largely stable, with only modest skew in most cases. The metabolic axis remains more right-skewed, a feature that was further corrected and will be incorporated into downstream modeling decisions.

Interaxis Correlation

With the eight axes defined, the next step was to determine whether they function as distinct physiological dimensions. If the structure is valid, the axes should not simply mirror one another. Some overlap is natural in biology, but persistently high correlations would indicate redundancy.

To evaluate this, pairwise correlations were calculated across all eight axes at the population level.

Across the matrix, cross-axis correlations fall between 0 and 0.30, with most near zero. This supports the interpretation that the axes represent largely independent physiological dimensions and confirms successful separation of biological domains.

Because these calculations are based on large population-level data, effect sizes are expected to be modest. Heterogeneity dampens correlations. As a result, a low-to-moderate coefficient in this context can still indicate meaningful biological linkage in more narrowly defined subgroups.

Taken together, the structure holds. The axes remain largely independent, with only mild associations between certain domains. Those modest links may become more informative under deeper or subgroup-specific analysis.

Spotlight: Malabsorption & Hematologic Burden

The strongest cross-domain signal appears between the Malabsorption and Hematologic axes at r = 0.30.

On its surface, a correlation of 0.30 may seem modest. But across a broad, unadjusted population, a coefficient of this size points to moderately coordinated physiology. In this case, the link makes biological sense. The intestine governs absorption of iron, B12, and protein substrates required for red blood cell production. When nutrient absorption shifts, hematologic markers tend to follow. The relationship is not deterministic, but it reflects a coherent pathway connecting digestive function to systemic blood dynamics.

Practically, this matters because inefficient absorption does not stay confined to the gut. It can surface as downstream vulnerability in blood-based markers before overt deficiency is diagnosed. Identifying these bridges early provides a clearer sense of where physiological stress may be accumulating and where attention may be warranted next.

This pairing illustrates the broader aim of the FPI framework. Instead of evaluating systems in isolation, it looks for coordinated strain across domains. Further analysis will examine which specific markers within each axis are most responsible for these alignments and under what contexts they strengthen.

Inside the Axes

Having verified independence across axes, attention now shifts to relationships within each axis.

Rather than reviewing all eight axes in detail, this section focuses on two of the most relevant domains as a spotlight: the Hepatic axis and the Malabsorption axis. The goal is not to be exhaustive, but to examine whether the internal structure within these domains reflects real biological relationships.

Within the Hepatic axis, ALT and AST are strongly correlated at 0.76, which is expected for markers of hepatocellular injury. GGT also shows moderate alignment with these enzymes, linking liver cell stress with biliary and metabolic strain. Together, these relationships validate established and well-researched liver physiology.

The Malabsorption axis shows similarly coherent structure. Albumin correlates with calcium (0.51), consistent with the role of albumin in binding circulating calcium. Total protein and globulin are also strongly linked (0.66), reflecting their shared contribution to overall protein balance. These internal relationships are clinically grounded and reinforce that the axis captures meaningful information.

The key is structural balance. The system separates domains of burden while preserving the internal complexity that makes biological interpretation possible. Structure is introduced without flattening the nuance that allows real signals to emerge.

From a Data Perspective

Working with biomarkers at this scale requires restraint as much as analytical judgment. It is easy to bake in assumptions, overinterpret correlations, or collapse complexity into a single composite score. The goal here is different. The objective is to introduce structure without distorting physiology.

The biomarker layer is the first context subgroup completed within the FPI. It provides the structural backbone necessary to map biology onto functional outcomes in later phases. The growing system map below visualizes how these domains fit within the broader framework across context, drivers, and outcomes, along with key subgroups and representative variables.

The next update will move into diet structure within the driver layer, capturing variation across protein intake, fiber density, sodium exposure, eating patterns, and food sourcing behaviors. With the biological backbone established, the focus now shifts toward building out remaining subgroups and analyzing the relationships that connect them.

Protein has become one of the latest fixations in health and wellness. What was once framed as a basic nutritional requirement is now marketed as a performance strategy, not just for athletes chasing strength, but for everyday people pursuing a healthier life. A quick look at Google search trends shows how that interest has expanded, especially over the past five years, well beyond the fitness crowd.

Despite the momentum, many people still feel unsure how protein applies to their specific goals, lifestyle, and health status. Research is extensive, yet results vary across populations and study designs, highlighting how personalized protein intake can be.

This post explores how protein became a cultural fixation, what “high protein” actually means, why findings conflict, and how clearer segmentation can sharpen dietary guidance.

The Protein Economy

To understand why skepticism persists, we have to look at how protein entered the mainstream.

Once considered foundational, protein is now promotional. Restaurants spotlight it. Brands center products around it. Consumers gravitate toward it. The data reflects that shift.

From 2014 to 2021, protein callouts on U.S. restaurant menus rose gradually from 5.9% to 11.5% before sharply accelerating in 2022 . The steep increase was largely a result of fast-casual chains leaned into customizable protein options and paid upgrade. By 2024 that number reached 28.4%, and CNBC reports it is projected to surpass 40% by 2029.

So why has protein become such a powerful menu strategy?

The answer is largely economic. Meats, poultry, and fish are among the most expensive items on the plate, while grains and fats provide the cheapest calories. When a nutrient is both in demand and inherently higher-cost, it creates natural room for margin. Restaurants can frame protein as premium and justify upgrades.

And the strategy appears to be working. The same CNBC article from earlier reports the share of customers who say they “love” high-protein foods climbed from 24% to 33% in just three years. Promotion fuels demand, and growing demand fuels even more promotion.

While advertising can shape perception, it can’t set biological needs. As restaurants race to add more protein, a quieter question remains: how much do you actually need?

How Much Protein Is Actually Enough?

One complication of the current research is that the definition of “high protein” varies depending on the source. Below are several well-researched, yet contradicting recommendations:

• Harvard Health suggests up to 0.9 grams per pound per day.

• Consumer Reports recommends between 0.5 to 0.7 grams per pound per day.

• Cleveland Clinic cites 0.36 grams per pound per day as the general rule.

That spread allows for drastically different interpretations. For a 150-pound individual, recommended intake ranges from 54 grams to 135 grams per day. And that range expands even further when age, gender, and activity level are layered in.

Given how personal these targets are, how do we know if we are actually getting what we need? Surprisingly, national surveys suggest most adults are not falling short.

NHANES data show men consume roughly 0.48 grams per pound per day on average, while women consume around 0.41 grams per pound per day. For most individuals, their intake lies within the broad range of recommendations, though it may not always align with their specific goals.

Much of the complication stems from how these guidelines were defined. Some aim to prevent deficiency, others to optimize performance or support aging. Different populations and different objectives create a wide range of what counts as “sufficient”.

That variability affects research as well. When studies define “high protein” at different consumption levels, outcomes will naturally vary. Differences in design, baseline intake, and participant selection shape what is actually being tested.

Weight Loss and the Protein Illusion

One of the clearest examples of how shifting definitions create conflicting conclusions is in weight loss research.

Healthline highlights short-term metabolic research indicating that higher-protein diets can support weight loss. Protein increases satiety, helps preserve lean mass during calorie restriction, and may support metabolic rate when paired with strength training.

But long-term data tell a more complicated story.

A secondary analysis of the PREDIMED trial published in Clinical Nutrition found that higher protein intake was associated with greater weight gain when it replaced carbohydrates. The study focused on a Mediterranean population at high cardiovascular risk, and also discovered a connection between higher consumption of animal protein and an increased risk of all-cause mortality.

Both are labeled “high protein.” Both rely on real data. Yet they examine different populations, different time frames, and different dietary contexts. One focuses on short-term metabolic mechanisms under controlled conditions. The other evaluates long-term outcomes in a high-risk cohort.

The contradiction is less about whether protein “works” and more about what protein replaces, who is consuming it, and how long the exposure lasts. This does not mean one study is inaccurate. It means they were measuring different outcomes.

Fiber as an Important Caveat

Fiber can act as a confounding variable in high-protein diets, influencing how those diets ultimately affect weight, digestion, and metabolic health. Research from CHN highlights that adequate fiber supports gut health, improves satiety, and helps nutrient absorption. When protein intake increases alongside fiber, weight loss efforts tend to be more sustainable and metabolically stable.

Factors like fiber are often omitted control variables in high-protein research. When key variables are left out of the model, results can shift and appear contradictory. In weight loss studies especially, failing to adjust for fiber can skew conclusions about the effect of protein. What looks like disagreement is often incomplete modeling.

From Generalization to Personalization

Most high-protein studies examine narrow populations: young athletes, overweight adults in calorie deficits, aging individuals at risk of muscle decline, or people with preexisting cardiovascular concerns. These are not interchangeable profiles.

Yet recommendations are often generalized across them. The amount of protein that supports muscle retention in a 25-year-old lifter may not be appropriate for someone sedentary with elevated cholesterol. What preserves lean mass in a calorie deficit may look very different from what supports digestive comfort at maintenance. The simplicity of “add more protein” works for marketing. Biology is less uniform.

Before jumping to suggestions, a more useful starting point would be asking better questions:

• Is your goal muscle gain, weight loss, or digestive stability?

• What is your average daily fiber intake?

• Do you have a history of cardiovascular disease?

• Are you experiencing bloating after you eat?

And the list goes on. Meaningful patterns emerge when individuals are grouped by shared baseline traits rather than collapsed into a single average. When protein is evaluated within comparable profiles and clear objectives, the guidance becomes sharper. The evidence does not need louder conclusions. It needs more personalized framing.

Takeaways

Protein isn’t a magic fix or a hidden threat. It’s context-dependent. The research behind high-protein diets is substantial, but it is layered, conditional, and often shaped by underlying variables. Once you zoom out, the contradictions begin to look less like conflict and more like incomplete framing.

• Consumer interest has been amplified by marketing incentives. Protein is profitable, premium, and easy to equate with better health.

• Definitions of “high protein” vary widely, and most Americans already fall somewhere within the broad recommended range of 0.36 to 0.9 grams per pound per day.

• Short-term metabolic benefits can coexist with complex long-term findings. Weight loss illustrates how context shapes results more than sheer intake.

• Protein is conditional. Fiber intake, cardiovascular history, digestive tolerance, and many other factors shape how higher intake is experienced.

Protein does not need louder endorsements or stronger claims. It needs clearer evaluation. Finding the right amount requires examining evidence that reflects people like you, considering your baseline intake of other nutrients, medical history, and involving a clinician when appropriate.

The future of nutrition will be less about blanket recommendations and more about defined profiles. As data grows more granular, advice will increasingly reflect systems rather than single nutrients. The more structured the analysis, the more valuable the guidance.

Most people assume that generating data-driven insights starts with exploratory analysis or data cleaning. But the reality of a system as complex as chronic disease is that the data arrives complicated, fragmented, and far from perfectly packaged.

In the previous FPI update, I laid out the three core layers that define the framework behind the Functional Performance Index: context, drivers, and outcomes. That framework describes how daily function should be interpreted, but the variables do not magically place themselves into clean categories.

The real process of understanding the data begins earlier. It starts with deciding how each signal should be classified and where clear boundaries need to exist.

This post focuses on that preprocessing step. Decoding raw variable names, adding metadata that governs how each variable should behave inside the system, and organizing related parameters. Taken together, these steps allow daily function to be interpreted as a system rather than a collection of isolated measures.

Decoding Raw Variable Names

Large health datasets rarely arrive in plain language. In NHANES, most variables are identified by short codes and abbreviations created for data collection rather than interpretation. On their own, they offer little clarity about what a variable captures or how it should be understood.

Preprocessing began by reviewing the documentation for each data file and writing out clear descriptions for every variable. That process quickly revealed a pattern. Many variables look nearly identical at first glance, particularly in their wording, despite being designed to capture different aspects of health.

A frequent example is the difference between questions framed as “have you ever had” and those that ask “do you still have.” Although the wording appears subtle, the distinction is meaningful. The latter identifies individuals who are currently living with an active condition, which is critical for separating historical diagnosis from present burden.

Thyroid related variables illustrate this clearly. Both reference the same condition, but one reflects a past diagnosis while the other captures current burden. This separation creates a natural comparison between individuals with a history of disease and those who no longer experience ongoing impact, making it possible to study factors associated with improvement.

After understanding these distinctions, the next step is making them explicit through allowed values and metadata.

Defining Allowed Values

Once variables are decoded, classification does not stop at assigning them to a layer. Each parameter also needs rules that define how it should behave inside the system.

This is where allowed values come in. Rather than leaving interpretation open ended, each variable is paired with a small set of metadata fields that constrain how it can be used. These rules ensure consistency across the system and prevent similar looking signals from being treated the same by default.

The table above shows the allowed values used to define how variables behave inside the FPI. Each is assigned values across several dimensions:

• Layer. Indicates whether the variable belongs to context, drivers, or outcomes. This defines the role the signal plays within the system.

• Type. Further specifies what kind of signal it is within its layer. For example, context variables are separated into categories like demographics or medical history, drivers into domains like diet or activity, and outcomes into energy, mood, stress, cognition, physical capacity, or overall health.

• Actionability. Describes whether the signal can be directly changed, indirectly influenced, or should be treated as non actionable. This prevents fixed characteristics from being treated as levers for change.

• Temporal behavior. Defines how the signal changes over time. Variables may be static, slow moving, fast moving, or episodic, which determines the time horizon over which changes should be interpreted.

• Use flag. Specifies whether a variable should be included in modeling, included conditionally, or excluded entirely. This adds an explicit gate for downstream use.

These allowed values are not descriptive labels. They act as constraints that tell the system how seriously to take a signal, how often it can change, and when it should be used. With these rules in place, variables behave consistently across analyses.

The next step is applying this structure to real variables and showing how similar signals are intentionally treated differently in practice.

Applying Metadata Rules

Once allowed values are defined, they can be applied to real variables in a way that preserves meaning. The screenshot below shows how similar variables are classified using allowed values, ensuring each contributes to the system as intended.

Context variables capture historical and structural factors such as prior doctor visits, previously diagnosed conditions, and socioeconomic stress. While difficult to change, they are essential for defining baseline comparison profiles and enabling meaningful personalization.

Driver variables represent signals that can change over time, including nutritional intake, active medications, dietary structure, sleep patterns, and activity levels. These function as levers rather than background conditions and form the basis for recommendations within similar baseline profiles.

Outcome variables reflect where poor daily function is most apparent. Some capture direct symptom burden while others represent broader functional impact across energy, mood, cognition, stress, and physical capacity. Together, these outcomes link behavioral inputs to the dimensions of daily life that matter most.

This structure supports consistent interpretation and preserves meaningful differences across the system, creating a necessary foundation for delivering the kind of personalized recommendations the FPI aims to provide.

From a Data Perspective

This phase of the FPI closely mirrors what I learned working with large-scale data at PNC Bank. When data scales, it never arrives neatly organized. Variables overlap, concepts are represented multiple ways, and similar signals appear across different sources with slightly different definitions.

In those environments, the challenge is not running models. It is deciding how data should be understood. Which variables describe the same idea. Which ones capture different states. And which distinctions matter enough to preserve.

People often say that data scientists spend 60 to 80 percent of their time cleaning and preparing data. While not always the most exciting step, taking the time to define structure before anything is modeled is what prevents small assumptions from compounding into systematic errors.

As the FPI continues to take shape, the breadth of the data suggests a clear next step. Rather than expanding across every subsection at once, the focus will likely shift toward building depth within individual domains and refining how signals interact within those spaces. That process will guide how the system moves from structure into insight.

Digestion is one of the systems most sensitive to those shifts. It depends on steady circulation, nervous system balance, and consistent signaling to function smoothly.

This post focuses on how exercise impacts digestion, how different types of exercise affect digestion in different ways, and how exercise supports related systems that indirectly influence digestive health.

How Exercise Impacts Digestion

Exercise influences digestion by changing how the gut moves, how well it is supplied with blood, and how the body allocates energy.

One of the most direct effects is on gut motility. The digestive tract relies on coordinated muscle contractions to move food and waste through the system. Regular movement helps maintain this coordination, while physical inactivity allows gut muscles to become less efficient. Exercise strengthens these contractions, helping the gut empty more completely and reducing stagnation that can disrupt digestive comfort.

Exercise also improves circulation to the digestive tract. Increased blood flow during physical activity supports the gut lining and helps maintain a healthy balance of bacteria. Better perfusion strengthens digestive tissues and supports overall gut function.

Metabolic regulation plays a role as well. When energy availability is low, the body prioritizes vital organs and reduces resources allocated to digestion. Regular exercise supports metabolic efficiency, helping digestion remain active and better regulated.

Finally, exercise indirectly supports digestion through sleep and recovery. Improved sleep quality allows the gut time to repair, regulate immune activity, and restore balance.

Source: Cleveland Clinic

Together, these mechanisms show that exercise shapes digestion by influencing movement, circulation, energy availability, and recovery. These effects create the conditions that determine how digestion functions day to day.

Different Types of Exercise, Different Digestive Effects

Exercise does not influence digestion in a single way. Its effects depend on type, intensity, duration, context, and system’s current capacity to absorb activity.

Across NIH Findings, moderate-intensity movement is consistently associated with improved digestive function. Activities like walking, cycling, and yoga support gut motility, reduce inflammation, and strengthen gut barrier integrity, especially when recovery and hydration are adequate. At this level, exercise is typically beneficial for digestion across conditions such as IBS, IBD, and GERD.

In contrast, high-intensity or prolonged exercise introduces a physiological tradeoff. During intense exertion, blood flow is redirected away from the gut toward working muscles. This temporarily suppresses digestion and increases intestinal permeability, which can lead to nausea, cramping, diarrhea, or abdominal pain, particularly in individuals with lower digestive stability.

Digestive responses to exercise are context-dependent, not dictated by whether movement is “good” or “bad.” Activities ranging from yoga to heavy lifting offer different benefits, yet all can support digestive health when matched to the individual. Understanding your body and goals allows you to set realistic targets that strengthen digestion without setting yourself back.

The Broader Systems Exercise Supports

The benefits of exercise extend beyond the gut itself. By improving mental health, brain function, and disease resilience, movement creates conditions that allow digestion to function more consistently over time.

Mental Health

A study published by UCLA Health found that adults average 3.4 poor mental health days per month. Among those who exercised, that number was over 40% lower. The findings suggest exercise changes brain function in ways that reduce depression and anxiety, not just perceived stress.

Brain Health & Aging

Even small amounts of movement appear to protect the brain. A Johns Hopkins–led study found that 35 minutes of moderate to vigorous activity per week was linked to a 41% lower risk of dementia over four years. Higher activity levels were associated with progressively greater risk reductions, including in frail older adults.

Cancer Risk Reduction

Across NIH Cancer Research, exercise shows a consistent protective pattern. Meta-analyses indicate that physically active individuals have roughly a 10–25% lower risk of several major cancers, including breast, colon, endometrial, kidney, stomach, esophageal, and bladder cancer. These benefits are thought to stem from reduced inflammation, improved metabolic regulation, and better immune function.

Takeaways

Exercise influences digestion in ways that extend beyond fitness or performance alone. Its relationship with digestion is reciprocal: digestion shapes capacity for movement, and movement reshapes the conditions that allow digestion to function reliably over time.

• Digestion responds best to regulated movement. Exercise supports gut function by improving motility, blood flow, and energy balance.

• Different intensities carry different tradeoffs. Moderate exercise is generally supportive, and higher-intensity training can offer meaningful benefits as long as recovery, nutrition, and overall capacity are respected.

• The benefits extend beyond the gut. Improvements in mental health, brain resilience, and disease risk reduction create upstream conditions that support digestive stability over time.

Taken together, this reinforces that exercise works best when it reflects the individual. Goals can shift, routines can vary, and priorities can change, as long as movement remains part of the routine.

Fitness tends to stick when people put their own spin on training and focus on what they can sustain. In the long run, showing up matters more than pushing harder. Over time, that consistency builds a level of confidence and trust in the body that often carries into other areas of life.

This post focuses on the developing framework underneath that idea. Not results yet, and not recommendations, but the logic that connects context, drivers, and outcomes into a single functional signal.

Chronic disease is often discussed as a collection of isolated symptoms or habits. The FPI approaches it as a system. This update breaks down how that system is organized and why each layer matters.

Why Chronic Disease Needs Layers

Chronic disease is frequently framed as one problem, even though it is shaped by many interacting factors. Symptoms, behaviors, lab values, and personal circumstances all get mixed together, making it difficult to tell what is actually driving poor daily function versus what is simply being experienced.

The Functional Performance Index separates health into three layers to keep those signals distinct:

• Context defines the baseline conditions a person brings with them, such as demographics, medical history, and medications. These factors shape interpretation but are not targets for change.

• Drivers represent modifiable behaviors and physiological processes, including diet, activity, sleep, and metabolic markers. These are the levers that influence daily function.

• Outcomes capture how health is experienced day to day, reflecting impact across energy, mood, stress, cognition, physical capacity, and overall health.

Each layer answers a different question. Context explains who someone is. Drivers explain what may be influencing change. Outcomes reflect what that change feels like.

Keeping these layers separate allows patterns to be interpreted more clearly and compared more meaningfully across similar profiles. Together, they span over 1,000 parameters across context, drivers, and outcomes. Defining these layers is the first step toward understanding complex interactions without oversimplifying them.

Context: Who You Are Before Anything Changes

Context defines the baseline conditions a person brings with them. It describes the constraints and background factors that shape how health is expressed and interpreted across chronic disease and daily function.

This includes factors like demographics, body composition, medical history, medications, and access-related variables such as income or food security. These elements influence outcomes, but they are not levers to pull.

Two people can follow the same diet, have similar symptoms, or show comparable lab values and still experience very different outcomes because their context is different. Age, prior disease, prescriptions, and baseline physiology all change how signals should be read.

In the FPI, context is used to create appropriate comparison groups. Instead of asking how someone compares to the population at large, the system asks how they compare to others with similar baseline profiles.

Context does not explain what to change. It explains how to interpret what is already there.

Once that baseline is established, the next layer focuses on what may actually be pushing the system toward or away from poor daily function.

Drivers: What Actually Moves the System

Drivers are the factors that actively push daily function toward better or worse outcomes. Unlike context, these are not fixed. They represent behaviors and physiological processes that can change over time and meaningfully influence how someone feels and performs.

In the FPI, drivers fall into two broad categories: behavioral and metabolic.

Behavioral drivers capture patterns like diet composition, eating regularity, physical activity, sleep tendencies, and substance use. These inputs shape energy, recovery capacity, and day-to-day variability. Small changes here often accumulate into large effects over time.

Metabolic drivers reflect how the body is responding internally. This includes markers such as blood pressure, cholesterol profiles, glucose control, inflammation, iron status, and kidney-related measures. These signals help explain whether behaviors are being tolerated, compensated for, or contributing to downstream strain.

In the FPI, drivers are treated as levers. By comparing individuals to similar baseline profiles, the system can identify which drivers are most strongly associated with improved or worsened outcomes in that specific context.

The final layer reflects how context and drivers combine to shape how someone feels and functions every day.

Outcomes: What Poor Daily Function Feels Like

Outcomes capture how health is experienced in daily life. They reflect the downstream impact of context and drivers rather than underlying mechanisms.

Within the FPI, outcomes are interpreted as signals. They capture what the system feels like once multiple influences interact over time. The six core outcomes are:

• Energy and Vitality. Captures how much fuel someone has to get through the day, and how consistently that energy holds.

• Mood and Mental Wellbeing. Reflects emotional stability and the psychological burden of managing chronic health day to day.

• Stress and Anxiety. Tracks the cognitive and physiological stress load shaped by behavioral patterns and health status.

• Focus and Cognition. Captures concentration, mental clarity, and the ability to think and perform under normal daily demands.

• Physical Capacity. Reflects physical limitation, mobility, and the ability to function without pain or fatigue as a barrier.

• Overall Health. A composite signal of how well someone feels across all dimensions relative to similar profiles.

Together these six outcomes reflect how well or poorly the body is functioning across the dimensions that matter most for daily life and long-term performance. Outcomes are compared across similar profiles to identify where function is most impaired and which drivers are most likely to shift things in a better direction.

Outcomes do not explain poor daily function. They show where it is most apparent and where change is likely to matter most.

Putting the Layers Together

The Functional Performance Index is built on a simple premise: daily function cannot be captured by a single metric.

Context establishes the baseline a person starts from. Drivers introduce change within that baseline. Outcomes reflect how those forces are experienced in daily life. Each layer serves a different purpose and plays a unique role in personalizing behavioral guidance.

Keeping these layers separate prevents common misinterpretations. Symptoms are not assumed to explain themselves. Behaviors are not evaluated without reference to baseline physiology. Comparisons are drawn from similar profiles rather than the population at large.

The goal is not to reduce chronic disease to simplicity, but to structure its complexity in a way that preserves meaning. Better daily function emerges from understanding how context, drivers, and outcomes interact over time.

This layered approach allows the FPI to translate population-level data into insights that remain interpretable at the individual level.

From a Data Perspective

The FPI is being developed using data drawn from 37 distinct datasets, together tracking over 1,000 parameters across context, drivers, and outcomes.

This scale makes it possible to move beyond broad averages and look at chronic disease and daily function in more specific situations. Instead of asking what tends to matter in general, the system can ask what has tended to matter for people with similar starting points, constraints, and health backgrounds.

The output is not a medical diagnosis, but a clearer sense of where attention is most likely to be useful. In practice, this means someone can better understand how their functional outcomes compare to similar profiles, and which drivers are most commonly associated with improvement in those contexts.

While this approach remains flexible, defining the framework upfront is a critical step in any data science or product development process. It sets realistic expectations for what the system can support and creates a foundation for thoughtful iteration rather than reactive adjustment.

The aim is to build a tool that helps people make sense of their daily function with more context and less guesswork. One that organizes complexity in a way that feels practical, interpretable, and honest about what the data can support as the system continues to evolve.