Skip to main content

ResearchResearch summary

Candida enteritis in US claims: what the codes show

A US insurance-claims study reports code B37.82 and related care patterns; claims do not confirm illness.

In this pathway

Current topic
Candida enteritis in US claims: what the codes showCurrent

Page trust and updates

First published
2026-10-09
Evidence as of
2026-10-09
Last updated
2026-10-09
Authorship
Candipedia Research
What changed
Added a contextual summary of the reviewed 2026 claims analysis; no clinical claim or practical action changed.
Mogana Das Murtey and Patchamuthu Ramasamy, "Saccharomyces cerevisiae, SEM image", licensed under CC BY 3.0

Bottom line

US insurance claims show how a Candida-enteritis code was recorded and what care followed, but they do not confirm that people had Candida-caused intestinal disease. Table 1 reports 3,710 coded people, a denominator of 25,873,638, and 14.3 per 100,000; the Results section reports 25,871,280 no-code comparators. Together, the group counts exceed the Table 1 denominator by 1,352, a discrepancy the paper does not explain.

Why this source matters

This study adds a baseline of coded symptoms, test claims, care settings, and antifungal prescriptions in a large commercial and Medicare claims database. It describes practice patterns around code B37.82, not the prevalence of confirmed Candida enteritis or small-intestinal fungal overgrowth (SIFO).

Study and population

This exploratory retrospective cross-sectional analysis used Merative MarketScan Commercial/Medicare claims. The database covered more than 65 million people during 2017–2024; the study window was July 1, 2017, through September 30, 2024. It included 3,710 people with a B37.82 code that was not listed only on a laboratory or imaging claim and with continuous enrollment from 180 days before through 30 days after the code date. Children and adults were included, with no age restriction reported. Table 1 gives a denominator of 25,873,638 and a rate of 14.3 coded people per 100,000. The Results section reports 25,871,280 continuously enrolled people without the code during the corresponding window around their first healthcare visit. Together, the reported coded and comparator counts exceed the Table 1 denominator by 1,352; the paper does not explain the discrepancy.

What was tested

The analysis compared people with and without the B37.82 claims code. It counted diagnoses, symptoms, and testing claims in the 180 days before the index date, care setting and provider type on that date, and antifungal prescriptions from 7 days before through 30 days after it. The code is the study's measured exposure; it was not independently confirmed as a clinical diagnosis. The study did not assess test accuracy, symptom response, fungal eradication, or comparative treatment effectiveness.

Results

The coded group had a median age of 44 years (IQR 31–56); 2,687 of 3,710 people (72.4%) were recorded as women and 1,023 (27.6%) as men.

Fatigue or malaise was recorded for 1,623 of 3,710 people (43.7%) and abdominal pain for 949 (25.6%) in the previous 180 days. These co-recorded symptoms do not establish a Candida cause.

Claims in that 180-day window included esophagogastroduodenoscopy (EGD) for 223 of 3,710 people (6.0%), fungal stool culture for 284 (7.7%), and Candida antibody testing for 456 (12.3%). A billing claim is not a test result. An EGD claim alone does not show that small-intestinal aspirate culture or histologic confirmation was performed.

Index-date care was outpatient for 3,669 of 3,710 people (98.9%). Provider categories were non-mutually exclusive: family practice/internal medicine was recorded for 1,632 people (44.0%), alternative therapist for 430 (11.6%), and gastroenterologist for 74 (2.0%).

Antifungal prescriptions were recorded for 1,398 of 3,710 people (37.7%) from 7 days before through 30 days after the code date: fluconazole for 871 (23.5%) and nystatin tablets for 725 (19.5%). These prescription counts may overlap and do not show whether treatment helped or caused harm; those outcomes were not reported.

Applicability

The findings describe people represented in this commercial and Medicare claims database and the way B37.82, symptoms, testing, and prescriptions were recorded. They do not estimate the prevalence of clinically confirmed Candida enteritis or SIFO, and they may not represent people with other insurance or no insurance.

Limitations

No clinical examination, pathology, culture, or chart review validated the code. Claims may misclassify disease or omit symptoms. The paper reports no laboratory results, test-accuracy measures, symptom response, microbiologic outcomes, or antifungal harms. EGD frequency is procedure use, not a confirmation rate. The 3,710 coded people plus the separately reported 25,871,280 no-code comparators exceed the Table 1 denominator (25,873,638) by 1,352; the paper does not explain the discrepancy. The reported rate is repeated as published without deriving another denominator.

Contribution to the evidence

Contextual. The study documents how a code was used and what care was recorded in a large insured population. Because it did not validate diagnoses or measure treatment outcomes, it does not replace site-specific disease definitions, condition-specific testing guidance, or aspirate-defined SIFO research.

Practical change

No practical change. A diagnosis code, a billed test, or a prescription in claims cannot by itself confirm Candida infection, show a test result, or demonstrate treatment benefit. The linked evidence map's conclusions and practical action remain unchanged.

Funding and conflicts

The authors reported no specific funding and no conflicts related to this work.

Source

Read the original publication (published online 16 April 2026).

Study type, population, and sample scale

Study type
Exploratory retrospective cross-sectional analysis of US insurance claims
Population
Children and adults with or without an ICD-10-CM B37.82 claims diagnosis in the Merative MarketScan Commercial/Medicare database
Sample scale
Table 1 reports 3,710 coded people, a denominator of 25,873,638, and 14.3 coded people per 100,000; the Results section reports 25,871,280 no-code comparators. Together, the reported coded and comparator counts exceed Table 1's denominator by 1,352; the paper does not explain the discrepancy.
Read the original source

Results

  • Table 1 reports 3,710 people with a B37.82 code and a denominator of 25,873,638, corresponding to 14.3 coded people per 100,000. The Results section reports 25,871,280 people without the code; together, the coded and comparator counts exceed Table 1's denominator by 1,352. The paper does not explain the discrepancy, so no combined denominator is inferred.
  • The coded group had a median age of 44 years (IQR 31–56); 2,687 of 3,710 people (72.4%) were recorded as women and 1,023 (27.6%) as men.
  • In the coded group, fatigue or malaise appeared in 1,623 of 3,710 records (43.7%) and abdominal pain in 949 (25.6%) during the prior 180 days; these are recorded symptoms, not evidence of a shared cause.
  • In the prior 180 days, claims included an EGD procedure for 223 of 3,710 people (6.0%), a fungal stool culture for 284 (7.7%), and a Candida antibody test for 456 (12.3%); claims provide no test results.
  • Index-date care was outpatient for 3,669 of 3,710 people (98.9%). Provider categories were non-mutually exclusive: family practice/internal medicine, 1,632 (44.0%); alternative therapist, 430 (11.6%); gastroenterologist, 74 (2.0%).
  • From 7 days before through 30 days after the code date, antifungal prescriptions were recorded for 1,398 of 3,710 people (37.7%), including fluconazole for 871 (23.5%) and nystatin tablets for 725 (19.5%); response and harms were not reported.

Limitations

The study classified people by presence of ICD-10-CM B37.82 in insurance claims.
There was no clinical examination, pathology, culture, or independent chart validation to establish Candida-caused illness; claims coding can misclassify or miss disease.
Test and procedure claims were counted in the 180 days before the index date.
Test results and accuracy were unavailable. An EGD claim does not show that a small-intestinal aspirate culture or histologic confirmation was performed.
Symptoms and other diagnoses were recorded around the code date.
Co-recording does not show that Candida caused symptoms. Claims may under-code symptoms, and MarketScan does not necessarily represent people with other or no insurance.
The paper reports 3,710 coded people and a Table 1 denominator of 25,873,638, while Results gives 25,871,280 people without the code.
Together, the reported coded and comparator counts exceed the Table 1 denominator by 1,352; the paper does not explain the discrepancy. This summary repeats the reported rate without recalculating it or inferring a combined denominator.

Applies to

  • Recorded diagnosis, symptom, test-claim, care-setting, and prescription patterns among people represented in the MarketScan Commercial/Medicare data during the study window

Does not apply to

  • Prevalence of clinically confirmed Candida enteritis or SIFO, people outside the represented insurance population, test accuracy, symptom causation, or antifungal effectiveness

Contribution to the current conclusion

Contextual

The study describes recorded care patterns around a Candida-enteritis code, but the claims were not clinically validated and do not establish disease, cause, test performance, or treatment benefit.

Related evidence map

Practical change or no change

No practical change. These claims describe coding and care patterns; they do not confirm Candida-caused illness, validate tests, or show antifungal benefit. The linked evidence map's clinical conclusions and practical action remain unchanged.

Funding and conflicts

Funding
No specific funding was reported.
Conflicts
All authors reported no conflicts related to this work.
Paper date
2026-04-16

Sources

View all 1 source