🔎︎
Health-related social needs (HRSN) data now sit at the center of managed care operations, informing referrals, risk adjustment, population health analytics, and tailored payer benefits. A qualitative study published May 1, 2026, in JAMA Health Forum suggests that the underlying data may be sufficiently biased to compromise broader applications.
Researchers led by Joshua Vest, PhD, MPH, of Indiana University, interviewed 20 patients and 20 healthcare professionals across multiple health systems in Indiana and Florida between January and May 2025. The settings ranged from federally qualified health centers and a multihospital safety-net system to a large academic medical center and an affiliated Veterans Affairs hospital.
Four bias categories emerged. Sampling bias arose because organizations applied screening unevenly, with some clinicians acknowledging that they screened fewer patients in affluent settings or skipped patients based on appearance. Detection bias occurred when clinicians did not document HRSNs they identified or did not access social work notes already in the chart. Nonresponse bias was driven by stigma around finances, perceived power dynamics, fear of consequences, such as elderly patients worried about losing autonomy at home, and uncertainty about whether disclosure would yield help. Misclassification bias surfaced in differing definitions of food insecurity, housing instability, and transportation barriers between patients and clinicians.
For managed care leaders, the implications run in two directions. HRSN data have a high positive predictive value, supporting individual-level decisions such as referrals. At the aggregated level, however, sampling and nonresponse biases mean population statistics likely underestimate true need, particularly among Spanish-speaking patients and older adults. The authors note that the recent CMS removal of HRSN-related quality metrics weakens incentives for systematic collection. Standardized instruments, technology-enabled workflows, and consistent Z-code documentation could narrow the gap.
