Nihit Gurram on Why “How Many Medications” Is the Wrong Question in Older-Adult Safety

Medication safety for older adults usually gets measured by a single number: how many prescriptions a patient is taking. It is an easy figure to chart, and a tempting one to act on. For Nihit Gurram, founder of Mosaic Health Solutions, a health informatics graduate student, and an incoming medical student at the Kansas College of Osteopathic Medicine, that number is also the wrong place to start. The count tells you how full the medicine cabinet is. It says almost nothing about which shelves are dangerous.

“The instinct is to treat the pill count as the risk, and it is not,” Gurram says. “Two patients can be on the same number of medications and face completely different odds of ending up on the floor or in the emergency department. What separates them is not how many drugs they take. It is which kinds. That distinction is the entire insight, and most of the system is built to miss it.”

Count Versus Kind

Gurram’s argument turns on a difference clinicians know well but health systems rarely operationalize. A long medication list is a marker of complexity, not a measure of harm. The risk concentrates in particular categories, especially the well-documented group clinicians refer to as fall-risk-increasing drugs, the medications that act on the brain, alertness, and balance. A patient on a short list that includes several of those can be in more danger than a patient on a longer list of comparatively benign ones.

“Counting is easy, so we count,” he says. “But the question that actually predicts harm is qualitative. Which mechanisms are in play, how do they stack, and what does that combination do to an older body. A number cannot answer that. A system that only knows how to count will keep sounding calm right up until the fall.”

When Ordinary Complaints Carry Outsized Risk

Part of what makes the problem stubborn, in Gurram’s telling, is that the higher-risk categories often treat the most ordinary complaints. Trouble sleeping, low mood, anxiety, bladder urgency, these are everyday issues, and the medications that address them are common and easy to prescribe. The very ordinariness is what hides the risk.

“Nobody thinks of a sleep aid as a high-stakes decision,” he notes. “But in an older adult, some of the drugs aimed at the most routine symptoms are exactly the ones that affect balance and cognition. The risk does not announce itself. It arrives wearing the costume of a minor fix, and that is what makes it so easy to wave through.”

Timing Is the Tell That Short Visits Miss

If kind matters more than count, timing is the clue that ties cause to effect, and it is the one a fifteen-minute visit is least equipped to catch. A new symptom that surfaces weeks after a new prescription is often the most important data point in the chart, and the easiest to overlook, because the two events are separated by time and, frequently, by different clinicians.

This is the pattern behind what Gurram describes as the prescribing cascade, where the side effect of one medication gets read as a new condition and treated with a second, which carries its own risks, and so on.

“The calendar holds the answer, but nobody is reading it as a story,” he says. “A dizziness complaint in isolation looks like one thing. A dizziness complaint that began three weeks after a new prescription looks like something else entirely. Connecting those two requires memory the visit does not have time for. That is precisely the kind of work software should be doing in the background.”

Designing for Restraint

The count-versus-kind problem maps directly onto how Gurram believes clinical tools should be built. A system that treats every possible drug interaction as worthy of an alert trains clinicians to ignore all of them. The discipline, he argues, is restraint: surfacing the few risks that genuinely matter, with the context that makes them actionable, and staying quiet about the rest.

“The temptation in software is to prove how much you know by surfacing everything,” he says. “In medication safety, that is the fastest way to become noise. The hard part is not detection. It is deciding what is worth a clinician’s attention inside a visit that is already overbooked. Targeting the small, high-risk group with context beats firing on every theoretical interaction. Restraint is what earns trust.”

A Public-Health Question, Not Only a Clinical One

Gurram is careful to frame the issue at the level of systems and design rather than individual care. He is a technologist and an incoming medical student, not a practicing physician, and he is explicit that none of this is medical guidance. He does not name drugs to avoid or tell anyone to change a regimen. His interest is in how the system frames the question in the first place.

That framing, he believes, is where the leverage is. As populations age and complex regimens become the norm rather than the exception, medication safety stops being only a bedside concern and becomes a public-health one.

“We have been asking how many,” he says. “The more useful question, the one better tools and better systems should be built around, is which ones, and why now. Get the question right and the technology finally has something worth doing.”

As founder of Mosaic Health Solutions, a company building AI-powered clinical decision support focused on medication safety for older adults, Gurram is building toward exactly that premise: that the patients most exposed to medication-related harm deserve tools designed around the risks that actually matter, not the ones that are simply easiest to count.