Simpson's Index

Two people can have the exact same list of bacteria in their gut and still have completely different guts. In one, no single organism takes up much room. In the other, one species has swallowed half the community and everything else is scraping by.

This score puts a number on that difference. It is a research-style stool microbiome measurement rather than a diagnostic one, so it will not tell you that you have a disease. What it can do is show whether the community in the sample is broadly shared or being run by a few dominant organisms, and whether that changes when you do something.

Richness, Evenness, and Why the Formula Matters

Diversity has two parts. Richness is how many kinds of bacteria are present. Evenness is how equally the population is split among them. This score folds both in, but it leans hard on evenness, so a gut where one organism dominates scores as low diversity even when the full species list is long.

Three different formulas travel under the same name, and they move in opposite directions. The raw version, often written as D, is a dominance measure. It runs from near zero up to 1, and a higher number means one species is taking over. The two common transformations, 1 minus D and 1 divided by D, are diversity measures, where higher means a more evenly shared community.

You can watch this cause confusion in the published work. Children with persistent unexplained diarrhea had a median score nearly triple that of healthy controls, 0.55 against 0.20, and that was reported as worse gut health. Adults with acute myeloid leukemia, a blood cancer, saw their median fall by roughly a sixth during treatment, from 0.628 to 0.521, and that was also reported as worse gut health. Both readings are correct. They are different formulas.

So before you read your own result, find out which version your lab reports. A number above 1 is an inverse Simpson value and has no fixed upper ceiling across all possible samples. A number between 0 and 1 could be either the dominance form or the diversity form, and you cannot tell which from the number alone. The score is never negative under any of the three.

When the Gut Community Collapses

The clearest signal from this measurement comes from people whose gut has been hit hard. In adults with acute myeloid leukemia treated with intensive chemotherapy, gut diversity fell sharply during episodes of fever with almost no white blood cells left to fight it. The number of distinct bacterial types dropped at the same time, and Enterococcus, an opportunistic organism, expanded to fill the space. Bone marrow recovery did not bring the diversity back.

That matters because the level of diversity during treatment tracks with what happens next. In a 534-person study of autologous stem cell transplant, where your own blood stem cells are returned after chemotherapy, the 240 people with stool samples around the time their new blood cells took hold had better outcomes when diversity was above the group median. Their risk of dying or having their disease progress was about half as high.

Other forms of collapse show the same pattern. Active Crohn's disease and active inflammation of a surgically built intestinal pouch both show steep drops in diversity compared with the same conditions in remission. In people with COVID-19, more severe illness tracked with lower diversity, though the trend fell short of statistical significance in a small group. Faecalibacterium was clearly depleted in those people. It is one of the main producers of butyrate. Butyrate is the fuel your colon lining runs on.

Sources: Khan et al. (stem cell transplant), Rattanathammethee et al. (leukemia), Alcázar et al. (childhood obesity).

The settings where this number carries the most weight are the ones where something drastic has happened to the gut. If you are heading into cancer treatment or have just come through it, a diversity reading is one way to see what the treatment did to your gut ecology, and there is outcome data behind it. In everyday life, the signal is far weaker.

Metabolic Health Has a Signal, but It Is Blunt

Among 191 children aged 8 to 14 with obesity, those classified as metabolically unhealthy had lower gut diversity than those who were metabolically healthy at the same weight. They also had less Akkermansia, a bacterium that lives in the mucus layer lining the gut wall.

Adult population data point in the same general direction, but not strongly enough to make this a metabolic test. In 2,166 adults from Dutch population studies, higher overall gut bacterial diversity was associated with lower insulin resistance and lower type 2 diabetes prevalence after adjustment for age, sex, lifestyle, BMI, and some medication use. That study scored diversity with a related measure rather than this exact index, so read it as support for the general idea and not for this specific number. It also found that specific butyrate-producing bacteria carried useful signal. If you want to understand your metabolic health, fasting insulin and a glucose tolerance test will tell you far more than this number will, and you can order both.

Where Higher Diversity Can Mislead

More diversity is not automatically better. In a small cohort of adults that included people living with HIV, those with fatty liver disease had higher Shannon and Simpson diversity than those without it. Endometriosis data are less clean: a 2025 pooled analysis found Simpson differences across four studies, but heterogeneity was high, and earlier gut-focused analyses did not find a significant gut Simpson difference.

This stops being a contradiction once you stop treating the score as a grade. It measures the shape of a community, not its quality. A gut can become more even because a healthy range of organisms is thriving, or because the species that normally dominate a healthy colon have been pushed aside by a wider spread of organisms that do not belong there. The number cannot tell those two apart.

That is the central limit of the measurement, and it is why the score works better as a change detector within one person than as a verdict on whether your gut is good or bad. In pooled analyses of fatty liver disease, this index often fails to separate cases from controls at all, which is what you would expect from a metric that is blind to which organisms are doing the dominating.

Why a Single Reading Can Fool You

Start with how much this number moves on its own. Daily stool sampling in healthy adults over six weeks shows substantial swing in evenness without a new disease or treatment. In a year-long Swedish population study, about a quarter of all the variation in gut composition was within individuals rather than between them. A single reading is a snapshot of a community that reorganizes week to week.

• The formula your lab used: the dominance form and the diversity form move in opposite directions and are both labeled Simpson. Compare two results only if they came from the same lab reporting the same version.
• The sequencing pipeline: how deeply the lab sequenced, which stretch of the bacterial gene it targeted, and how it grouped the reads all shift the score. Two labs can score the same sample differently.
• Recent antibiotics: one dose is enough to drop diversity within days, and in some people it stays down for months. A reading taken in that window tells you about the antibiotics, not about you.
• What abnormal even means here: in one evaluation of commercial panels, about 16 in 100 healthy reference people were classified as dysbiotic. There is no validated normal range for this score.

Collection method matters, but less than biology if you use the kit as directed. Stabilizing cards and several standard collection kits often track reasonably with immediate freezing for diversity measures. A month in RNAlater did not significantly shift Simpson in pooled preservation data. But some methods, especially improvised room-temperature storage or ethanol handling, can move parts of the microbiome profile, so do not compare results across collection systems casually.

Tracking Your Trend

One reading is weak. A sequence of readings tells you more, because you become your own control and the pipeline stays constant. Get a baseline when nothing unusual is happening: no antibiotics in the past three months, no acute illness, no recent travel.

If you are making a real dietary change, retest at three months rather than three weeks. In the bean trial, the rise in inverse Simpson diversity showed up at eight weeks and faded once people went back to their usual diet, which is roughly the timescale to expect from food.

After antibiotics, the only comparison worth making is against your own pre-antibiotic reading. Diversity often rebounds within about two months, but not in everyone, and some people are still below where they started at six months. Without a prior number you cannot tell recovery from a new normal. Once you have two or three readings, annual retesting is usually enough unless something changes.

What an Out-of-Pattern Result Should Make You Do

A low or falling score is a prompt to look elsewhere, not a diagnosis. Retest first, because the week-to-week swing is large enough to produce a low reading by itself.

Then pair it with a stool inflammation marker. Calprotectin tells you whether your intestinal lining is actually inflamed. Low diversity with a normal calprotectin and no symptoms is an ecological observation you can watch. Low diversity alongside a raised calprotectin, or with blood in the stool, unexplained weight loss, or persistent diarrhea, is a reason to get a gastroenterologist involved and to consider colonoscopy rather than more sequencing.

Look at what is filling the space. A diversity score that has dropped usually means something expanded, and the named-organism portion of your panel will tell you which. A bloom of a single opportunistic organism reads very differently from a general thinning of the community, and only one of those points toward infection.

Finally, check the functional output. Short-chain fatty acid measurements and the abundance of the main butyrate producers tell you whether the community is still doing its job. A gut can lose evenness and keep working. That combination, diversity down but function intact, is far more reassuring than the diversity score alone would suggest.