User reports are the evidence people actually consult. Before a trial, before a label, someone reads what others said.
For urinary health specifically, that data is harder to interpret than almost any other category — and the reason is structural rather than a matter of honesty.
The one user-report database worth reading
There is a systematic, public source of user-reported supplement experience: the FDA’s CFSAN Adverse Event Reporting System (CAERS), accessible through the openFDA API[1].
It’s worth naming explicitly because it differs from commerce reviews in ways that matter. Reports are coded to MedDRA terminology, they come from consumers, healthcare practitioners and manufacturers, and the data is published quarterly from January 2004 onward[1][2].
The scale, from a published analysis: 15,430 reports naming at least one suspected dietary supplement between 2004 and 2013 — 66.9% mandatory, 32.2% voluntary[3].
But the first thing to understand is what it does not contain.
CAERS captures adverse events and product complaints. It does not capture whether a product worked.
So for a question like “did this prevent my UTIs from coming back,” this database — the best systematic user-report source available — has no field for the answer.
That’s the structural problem. The systematic data covers harm; the efficacy question lives only in unsystematic sources.
What ‘adverse event’ means
A note on terminology, because the number is easy to misread.
An adverse event is a symptom or condition suspected of being linked to a product. The FDA states this directly: reports do not represent any conclusion by FDA about whether the product actually caused the adverse events[2].
A person taking a supplement who develops a symptom may have developed it regardless. Until causation is established, “adverse event” is the accurate term and “side effect” is not.
The 2% problem
Here is the figure that governs how any of this should be read.
Analysis of CAERS data estimates the reporting rate for supplement-related adverse events in the United States at approximately 2%[4].
Two percent. For every reported event, roughly forty-nine go unreported.
The FDA acknowledges the same in its own documentation: even with mandatory reporting of serious adverse events for dietary supplements, generally only a small fraction of adverse events associated with any product is reported[1].
And the bias runs in a specific direction. Underreporting is worst among voluntary reporters like consumers and physicians[4] — which is to say, the people whose experience you most want to know about.
If systematic reporting captures 2%, informal reviews capture an unknown and differently-biased fraction.
Reports don’t link product to symptom
A limitation that surprised me, and it affects every product-level count.
The FDA states: a report may list several products, as well as several reactions that the consumer experienced. No individual food, dietary supplement, or cosmetic product is connected to any individual reaction[1].
So if someone taking four supplements reports three symptoms, the database records all four products and all three symptoms — without specifying which product was associated with which symptom.
Counting reports mentioning a product therefore counts co-occurrence, not attribution.
There’s also duplication. Multiple people may submit reports for the same event — an affected consumer and the clinician who treated them[5].
And the totals shift. CAERS is constantly updated, so the number of reports for a given product and the content of individual reports may change over time[5].
Why recurrence is the hardest outcome to self-report
Now the specific problem with this category.
Recurrence requires knowing that an episode happened. And as covered in the self-management post, that determination is unreliable without a culture — a hospital study found a 24.9% misdiagnosis rate among admitted patients, with 14.1% actually having asymptomatic bacteriuria.
If a quarter of clinician diagnoses were misclassified against guideline criteria, a self-reported count of episodes is measuring something looser than confirmed infection.
Recurrent UTI fluctuates naturally. Episodes cluster and then don’t. A period of frequent episodes followed by a quieter period is the expected pattern, with or without intervention.
And people start supplements at the worst moment. Nobody buys a urinary health product during a symptom-free stretch. Purchase happens at or near a peak — which means the subsequent period is very likely to be better, regardless of what was purchased.
This is regression to the mean, and it applies with unusual force here.
Put together: a user reporting “no recurrence in three months since starting” is describing something real about their experience. It’s just not evidence about the product, because the same three months would likely have been better anyway.
A bit more detail — on why this differs from other categories. For an outcome measured by instrument — LDL cholesterol, bone density, accommodation amplitude — self-report bias affects the interpretation but not the number. For recurrence, the outcome *is* the self-assessment. There’s no reading to check it against unless a culture was done each time. That collapses the distinction between what happened and what was perceived to have happened.
What controlled data says instead
The contrast is instructive.
The 2023 Cochrane review of cranberry products — 50 studies, 8,857 participants — found reduced risk of symptomatic, culture-confirmed UTI, with the effect present in women with recurrent UTI (RR 0.74, 95% CI 0.55–0.99).
Note the qualifier. Culture-confirmed. The controlled evidence uses an objective endpoint precisely because self-reported episodes wouldn’t be trustworthy.
And the same review found no difference between low, moderate and high PAC doses — a pattern that user reports would be structurally unable to detect, since nobody reports their PAC intake.
How to read reports usefully
They aren’t worthless. They’re useful for different things than people use them for.
Reports are reasonably good at surfacing unexpected symptoms. If many people independently describe the same unusual reaction, that’s a signal worth noting — it’s how post-marketing surveillance works, and as covered in the red yeast rice post, it’s where safety signals came from when trials were too small.
Reports are poor at establishing efficacy, especially for fluctuating conditions with self-assessed outcomes.
Reports are worst for exactly this combination: an intermittent condition, a subjective endpoint, and purchase timed to peak symptoms.
Three questions worth asking of any report about recurrence:
- Was the episode confirmed by culture, or assessed by symptoms?
- How long was the reporting period relative to their usual episode interval?
- What was happening when they started?
None of these are answerable from a typical review. That’s the point.
For readers in Korea
Two notes.
Korea publishes comparable data. The MFDS adverse event reporting statistics are available through the public data portal[6], broken down by product type, symptom, age and sex — the domestic equivalent of CAERS, with the same interpretive cautions.
And urology is directly accessible here, as covered in the self-management post. Since the distinguishing question for recurrence is whether an episode was culture-confirmed, that access is more useful than any volume of reports.
Closing
The honest summary is that user reports and this particular question are badly matched.
CAERS, the systematic source, doesn’t record efficacy at all — and captures an estimated 2% of the events it does record. Informal reviews record efficacy but have no mechanism to distinguish a real effect from regression to the mean.
And recurrence is the outcome where that distinction matters most, because purchase reliably coincides with the worst point in a naturally fluctuating pattern.
None of which means the people writing those reports are wrong about their experience. They’re describing what happened accurately. The gap is between what happened and what caused it — and that gap is exactly what a controlled trial exists to close.
Which is why the Cochrane figure uses culture confirmation, and why the useful question about any recurrence report is whether anyone checked.
Key Terms
- CAERS — the FDA’s adverse event reporting system for foods, supplements and cosmetics.
- Adverse event — a symptom suspected of being linked to a product; not a confirmed side effect.
- MedDRA — the standardised medical terminology used to code reported symptoms.
- Passive surveillance — a system that relies on reports being submitted rather than actively seeking cases.
- Regression to the mean — the tendency for an extreme measurement to be followed by a more typical one.
- Culture-confirmed — verified by growing bacteria from a urine sample, rather than assessed by symptoms.
At a Glance
- CAERS is the systematic public source of user-reported supplement experience, coded to MedDRA
- 15,430 reports named a suspected supplement between 2004 and 2013 — 66.9% mandatory, 32.2% voluntary
- It records adverse events, not efficacy — there is no field for whether a product worked
- Reports do not represent an FDA conclusion about causation
- The estimated reporting rate is approximately 2%
- Reports don’t link individual products to individual reactions — counts reflect co-occurrence
- Recurrence is self-assessed, fluctuates naturally, and purchase coincides with peak symptoms — regression to the mean applies strongly
- Cochrane’s cranberry finding used culture-confirmed episodes precisely because self-report wouldn’t be reliable
※ This article discusses how to interpret reported data and is for general information only. It does not evaluate any specific product, and does not replace medical assessment or treatment. Recurrent urinary symptoms warrant clinical evaluation rather than self-management — fever, flank pain, blood in the urine, or symptoms not improving require prompt medical attention.
References
- openFDA, “Adverse Food Event Overview” (reporting structure; underreporting; products not linked to individual reactions), https://open.fda.gov/apis/food/event
- US FDA, “CFSAN Adverse Event Reporting System (CAERS)” (scope of captured events; MedDRA coding; stated limitations on causation), https://www.fda.gov/food/compliance-enforcement-food/cfsan-adverse-event-reporting-system-caers
- Timbo BB, Chirtel SJ, Ihrie J, et al. “Dietary Supplement Adverse Event Report Data From the FDA Center for Food Safety and Applied Nutrition Adverse Event Reporting System (CAERS), 2004-2013”, Annals of Pharmacotherapy, https://pubmed.ncbi.nlm.nih.gov/29171279/
- “What Should Dietary Supplement Oversight Look Like in the US?”, AMA Journal of Ethics (2% reporting rate estimate; underreporting concentrated among voluntary reporters), https://journalofethics.ama-assn.org/article/what-should-dietary-supplement-oversight-look-us/2022-05
- Federal Register, “Posting Adverse Event Report Data Associated With Conventional Foods, Dietary Supplements, and Cosmetics on the Internet” (duplicate reports; data updated over time), https://www.federalregister.gov/documents/2016/12/07/2016-29277/posting-adverse-event-report-data-associated-with-conventional-foods-dietary-supplements-and
- Ministry of Food and Drug Safety, “Status of Reports on Adverse Events Related to Health Functional Foods,” Public Data Portal, https://www.data.go.kr/data/15099480/fileData.do