“Daddy, come in the bathroom and see this before I flush it,” is definitely a sentence that most parents dread.
Usually, it means one of two things. Either something has gone terribly wrong, or your child has decided that whatever is in the toilet is important enough to require a second opinion.
Strangely, medicine may be moving in the same direction. Medicine is often remarkably bad at getting information from people. We stick needles into screaming children, ask very sick oncology patients to wait for lumbar punctures, and tell people with broken bones to hold perfectly still while we take pictures of them. We even manage to let patients use the restroom immediately after check-in, then hand them a specimen cup five minutes later and explain that they are not leaving until they fill it. At that point, unfortunately, diving back into the toilet is generally frowned upon.
So, it is probably not surprising that medicine has a soft spot for information that is easy to acquire. If nobody has to schedule anything, drive anywhere, remember anything, or be stabbed with anything, we are going to collect more of it. That is what makes the toilet interesting.
We are planning to study a device being developed by a company that is trying to acquire some of the easiest information available: urine and feces. That sentence makes the technology sound less medical than it actually is. There are already plenty of consumer products that count steps, calculate sleep scores, and practically throw you a parade every time you stand up. Some are genuinely useful. Some are a little closer to performance theater. Most live in their own strange universe of proprietary measurements.
The smart toilet is currently marketed as a general-wellness product. It tracks stool patterns, hydration, urinary flow, and bathroom habits; its hydration estimate is based on urine color and concentration rather than a conventional laboratory osmolality result. That distinction is exactly why we became interested in studying the device. We are planning to compare its hydration-related measurements with conventional laboratory methods, including osmolality. Once a consumer device starts producing a number that resembles something we already measure in the laboratory, we know how to ask whether it is any good.
And it may not stop there. Home devices can already generate rhythm data, glucose measurements, oxygen saturation, and infectious-disease results. Toilets may eventually detect findings analogous to conventional urinalysis or stool testing. Once a device starts reporting things that have established medical counterparts, calling it “wellness” begins to feel a little silly. Blood in stool is not a readiness score.
The attraction is obvious. A beautifully performed laboratory measurement obtained twice a year may ultimately influence care less than a somewhat noisier measurement collected automatically 365 times a year. Ease and quality are not the same thing, but ease matters enormously. The easiest information gets collected more often, seen more often, and eventually used more often. At some point the question becomes less about what the device is called and more about what everybody is doing with the number.
When a graph starts to feel like truth
There is also something persuasive about repetition. A stray laboratory value is easy to distrust. A graph with hundreds of points feels different. There is a kind of momentum to it. A single unusual number invites questions. Three hundred measurements in a smooth trend line start to look like a story.
A 2026 meta-analysis of 182 studies and more than 31,000 participants found that repeated information tends to be judged as more truthful than novel information, a phenomenon psychologists call the illusory truth effect. The studies were not about laboratory trend lines, but the analogy is difficult to ignore.
The problem is that repetition does not have much to do with accuracy. You can be wrong very consistently. Laboratory medicine has entire textbooks devoted to this idea, although the examples are usually less personal.
Laboratorians spend their careers learning that respectable-looking values can still be wrong. The conditions can be inappropriate. The method can be poor. Quality control can fail. Reagent lots change. Matrix effects appear. Software changes. Interferences show up at exactly the wrong time. A result is not necessarily true just because it arrives on your phone in an elegant font, framed by attractive borders and a beautifully shaded graph. Meanwhile, the same information inside an electronic medical record will probably appear in a gray box somewhere between an alert nobody read and a medication somebody discontinued three years ago.
But that presentation is going to matter. These values will come with elegant interfaces, compelling graphs, green zones, red zones, AI summaries, dietary suggestions, exercise recommendations, supplement suggestions, and enough supporting material to make the whole package feel impressively medical. Some app will eventually inform a patient that their slow progression toward vitality is moving in the wrong direction because of a mattress-alignment problem. Or perhaps their mitochondria are upset.
That sounds ridiculous, but the interface will not. The interface will look fantastic.
A patient may walk into clinic with years of telemetry on sleep, activity, heart rhythm, glucose, urine color, oxygen saturation, hydration, and three proprietary scores nobody in the room can define without opening the app. There will be a graph with a reassuring green area and a terrifying red area. If the CBC is stone-cold normal, is the patient tired because they are anemic, because they slept badly, because their wearable thinks their recovery is 41, or because they have spent the last six months being told that 41 is bad?
Right now we mostly leave patients to sort that out themselves. But maybe now's the time to change that policy. Once patients and clinicians begin acting on these numbers, somebody has to decide how much trust they deserve.
Maybe diagnostics need a formulary
Eventually, there will be multiple devices claiming to measure roughly the same medically relevant things, and they will not all be equally good. One may compare beautifully with conventional testing. Another may work well until you get to the extremes. Another may generate a proprietary hydration or kidney-health score that sounds suspiciously like a laboratory result without actually being interchangeable with one. Many may be commercially available long before a local health system has decided what to make of them. Most will have very nice apps.
Do we really want to decide what we think about each of these one patient at a time?
This is where we think laboratory medicine should borrow an idea from pharmacy. FDA approval does not mean every hospital automatically stocks every approved drug. Pharmacy and therapeutics committees still decide which medications the health system prefers, which it restricts, which alternatives are acceptable, and what it is actually prepared to support. The formulary is not simply a shorter list. It is a local decision about what the institution understands, prefers, and is prepared to support.
There is also a warning in the pharmacy experience. A systematic review of 59 studies found that restrictions usually reduced drug use, but some pharmacy savings were offset by higher medical costs, and the review warned of unintended consequences for patients and payers. Restriction by itself is not the point. The value is making a deliberate decision and being willing to revisit it.
Maybe diagnostics needs something similar. A health system could say that one external device is reasonable for trend analysis, another is useful but requires confirmation, and a third is reliable enough for a defined clinical use. And another is basically astrology with a wonderful user interface.
We did not expect laboratory medicine to eventually need something resembling a preferred-toilet list, but here we are.
That does not mean the laboratory should own every device. Nor should we want to. Clinicians have a say. Patient advocates may have a say. Regulators and professional organizations certainly have a role. But I think it is reasonable to expect the laboratory to be asked three fairly basic questions: Does this thing measure what it says it measures? Does it agree with the test we already use? And how the heck does it work anyway?
Understanding how something works should probably be a prerequisite for formal adoption by a health system. Then come the less glamorous questions. How does it fail? How reliable is it after six months? Or after five years? What if it gets dirty? What happens in sunlight? Does humidity matter? What happens after a firmware update? Is version 17 measuring the same thing version 14 measured? If a patient replaces the device next month, can the new numbers really be trended against the old ones?
Those questions are annoying when the device is measuring something trivial. They become consequential when the number starts changing medical management.
None of this means the laboratory needs to own the device. The manufacturer still has to prove what it built, regulators still have to set minimum expectations, and clinicians still have to decide what a result means for the patient. But the laboratory has a very specific job in the middle of all of that: deciding whether the measurement deserves to be trusted. We do not need to own every sensor. We do need to understand the measurement.
The boundary should follow the measurement
The future laboratory may not be a department full of analyzers. It may be a network of measurements generated across hospitals, clinics, homes, phones, watches, and increasingly ordinary objects. Laboratory medicine’s role may therefore shift from controlling every instrument to determining which measurements deserve clinical trust.
It used to be that the location of the laboratory mattered a lot. If the analyzer was downstairs, everybody knew whose problem it was. Point-of-care testing complicated that, but we adapted. Instruments moved to the ICU, the emergency department, the bedside, and the clinic, and laboratory standards followed them.
Now the testing is moving again. It may not be physically inside the health system at all. It may be in someone’s bedroom, bathroom, pocket, or on their wrist. At that point geography stops being very useful for defining what belongs to laboratory medicine.
A better boundary may be the claim being made.
If a device says that you slept badly, maybe that stays in the world of health and wellness. If it reports an established medical measurement, detects a finding that normally drives clinical evaluation, or produces a number intended to substitute for something we already measure, that is different. The standards of laboratory medicine probably ought to follow the measurement wherever it goes.
We may literally end up comparing a number generated next to somebody’s toilet paper with a conventional laboratory instrument. That sounds ridiculous now. But if the number proves useful, patients start watching it, clinicians start acting on it, and health systems start importing it into the medical record, the ridiculous part will disappear very quickly.
The toilet will still be in the bathroom. The laboratory does not need to follow every device home. But our standards may need to follow the measurement wherever it goes.
Newsletters
Receive the latest pathologist news, personalities, education, and career development – weekly to your inbox.
