Supply chain failures in pharmaceuticals are rarely framed as a clinical adherence problem — but for patients managing chronic conditions, a shortage isn't just an inconvenience; it's a potential gap in treatment that can cascade into adverse health outcomes. New pharmacoepidemiological evidence from the Veterans Health Administration quantifies exactly how large that gap is and, critically, which factors predict who falls into it.
Drawing on electronic health records spanning 2017 to 2020, researchers analyzed roughly 1.5 million episodes of medication use across 1.3 million unique veterans exposed to one of 29 drugs that experienced a VHA-recognized supply disruption. Using group-based trajectory modeling applied to monthly proportion-of-days-covered values — a validated adherence metric — the team identified four distinct adherence profiles across a 12-month window centered on the shortage event. The largest cluster, representing 69.2% of episodes, maintained high adherence throughout. However, 8.5% of episodes showed a clear shortage-related disruption pattern — adherence declining specifically in the window surrounding the supply event — while another 8.3% exhibited disruption that predated the shortage entirely. Notably, drug-level characteristics, including therapeutic class and the number of active manufacturers, were more predictive of trajectory membership than patient-level demographics.
This finding reframes the shortage problem in an important way: the vulnerability appears to reside more in the drug supply ecosystem than in individual patient behavior or demographics. That has genuine policy and clinical implications. The observation that manufacturer concentration amplifies disruption risk aligns with broader pharmacoeconomic research linking generic market consolidation to shortage frequency and severity. For health systems managing large chronic-disease populations — cardiovascular, metabolic, psychiatric medications being perennial shortage candidates — these trajectory models offer a practical framework for prospective risk stratification when a shortage is announced. The study's retrospective, single-system design limits generalizability to non-VA populations, and causality cannot be established, but the scale of the cohort and the precision of the trajectory approach make this more than incremental — it provides a replicable methodology for shortage impact surveillance.