One Patient, Fifty Silos: How Disease-Specific Research Is Failing America's Most Complex Patients

Consider a patient presenting to a primary care physician in suburban Ohio: 67 years old, managing type 2 diabetes, heart failure with reduced ejection fraction, stage 3 chronic kidney disease, and moderate depression. She takes eleven medications. Her cardiologist has optimized her beta-blocker. Her endocrinologist has adjusted her SGLT2 inhibitor. Her nephrologist has flagged a contraindication her psychiatrist was not informed of. Each specialist is operating from a robust, evidence-grounded guideline. And yet no single guideline — no single clinical trial that generated the evidence underpinning those guidelines — was designed with her in mind.

This is not an exceptional case. According to data from the Centers for Disease Control and Prevention, approximately 42% of American adults live with two or more chronic conditions. Among adults over 65, that figure approaches 70%. Yet the randomized controlled trials that populate our most authoritative clinical guidelines routinely exclude patients with significant comorbidities. The result is a research enterprise that has grown increasingly sophisticated in its understanding of individual diseases while remaining structurally blind to the patients who carry several of them simultaneously.

The Exclusion by Design Problem

The tendency to exclude multimorbid patients from clinical trials is not arbitrary — it is, in many respects, methodologically rational. Patients with multiple conditions introduce confounding variables that complicate causal inference. They are more likely to experience adverse events, more likely to discontinue participation, and more likely to be taking concomitant medications that interact with the investigational compound. Excluding them makes trials cleaner. It makes regulatory submissions more defensible. It makes the path to approval shorter.

But methodological convenience is not the same as clinical utility. When the FDA approves a hypertension drug based on a trial population with a mean age of 54 and no significant comorbidities, and that drug subsequently enters the prescribing patterns of a population where the median hypertension patient is older and carries three additional diagnoses, the evidentiary bridge between trial and practice becomes structurally unsound. Clinicians are left extrapolating — often without acknowledging that they are doing so.

A 2019 analysis published in the Journal of General Internal Medicine found that across a sample of landmark cardiovascular trials, patients with three or more chronic conditions represented fewer than 10% of enrolled participants, despite comprising the majority of real-world patients treated with the resulting therapies. The evidence base, in other words, was constructed on a foundation that excluded the population most likely to use it.

Specialty Silos and the Guideline Problem

The compartmentalization of clinical research reflects and reinforces the compartmentalization of clinical practice. Cardiology trials are designed by cardiologists, funded by cardiovascular-focused agencies, published in cardiovascular journals, and reviewed by cardiovascular guideline committees. The same logic applies across nephrology, endocrinology, psychiatry, oncology, and every other specialty domain. Each discipline develops its own evidentiary ecosystem — internally coherent, externally isolated.

The consequences surface most acutely in clinical guidelines. When the American Diabetes Association recommends a particular pharmacological approach and the American College of Cardiology issues a partially overlapping recommendation for the same drug class in a different indication, the guidance documents rarely speak to each other. They do not address what to do when a patient qualifies for both protocols simultaneously, when adherence to one may compromise outcomes in the other, or when the combined medication burden creates polypharmacy risks that neither guideline considered.

This is not a failure of individual guideline committees, which are typically operating at the frontier of their specialty's available evidence. It is a failure of research architecture — a structural problem that no specialty-based committee is positioned to solve unilaterally.

Polypharmacy as an Evidence Vacuum

Polypharmacy — generally defined as the concurrent use of five or more medications — is one of the most clinically significant and least rigorously studied phenomena in US healthcare. Adverse drug-drug interactions, cumulative anticholinergic burden, competing metabolic pathways, and medication adherence fatigue all compound in patients carrying complex regimens. Yet the evidence base for managing polypharmacy is remarkably thin relative to the scale of the problem.

The National Institute on Aging and several academic medical centers have called for dedicated polypharmacy trials, but funding structures continue to favor disease-specific research. The National Institutes of Health, which remains the largest single source of biomedical research funding in the United States, allocates the majority of its extramural budget through disease-specific institutes and centers. A research proposal targeting multimorbidity as a primary focus must navigate institutional mechanisms that were not designed to accommodate cross-cutting questions.

The result is that clinicians managing patients on complex regimens are often making decisions in an evidence vacuum. They rely on clinical experience, pharmacist consultation, and rule-of-thumb heuristics developed outside of formal research contexts. These are not worthless inputs — experienced clinicians develop genuine expertise — but they do not constitute a substitute for a rigorous, generalizable evidence base.

Emerging Frameworks and Their Limitations

There is growing recognition within the research community that the single-disease trial paradigm requires structural supplementation. Pragmatic trial designs, which enroll broader and more representative populations, have gained traction as a partial corrective. The NIH's National Center for Advancing Translational Sciences has invested in platforms designed to facilitate cross-disease research collaboration. The Patient-Centered Outcomes Research Institute (PCORI) has explicitly prioritized multimorbidity and complex patient populations in its funding priorities.

Real-world evidence, derived from electronic health records and insurance claims data, offers another avenue for studying comorbidity patterns at scale. Unlike randomized trials, observational data captures the full heterogeneity of the patient population, including those with multiple chronic conditions who would have been excluded from the controlled setting. However, real-world evidence carries its own methodological limitations — confounding, selection bias, and data quality constraints — that make it an imperfect substitute for prospective experimental research.

Adaptive platform trials represent a more ambitious structural innovation. By testing multiple interventions across multiple conditions within a single overarching protocol, these designs create opportunities to examine interaction effects that traditional parallel-group trials cannot accommodate. But they remain resource-intensive, logistically complex, and slow to achieve broad adoption across the research enterprise.

A Research System Misaligned With Its Patient Population

The fundamental tension is this: the patients who most urgently require an integrated evidence base are the patients the research system is least equipped to study. Multimorbid individuals are expensive to enroll, difficult to retain, and analytically inconvenient. The incentive structures governing academic research — publication in high-impact specialty journals, grant success through disease-specific institutes, career advancement within disciplinary silos — do not reward the cross-cutting investigative work that these patients require.

Correcting this misalignment demands more than methodological innovation. It requires deliberate changes to funding priorities, journal scope, guideline development processes, and the institutional reward structures that shape researcher behavior. It requires treating multimorbidity not as a complicating factor to be excluded from research design, but as a primary clinical reality to be studied with the same rigor applied to any single disease.

Until that shift occurs, the clinician facing a patient with five diagnoses and eleven medications will continue to work at the intersection of five separate evidence bases — none of which was built for the person sitting across from them.