Medicare-covered innovation and U.S. disability, 1997–2019: Evidence from healthcare procedure codes and health survey data

• Medical innovation reduced US disability 1997-2019 across 19 of 21 indicators • Mean lag from technology introduction to disability reduction was 11.5 years • Supplemental Security Income recipiency declined 21.3%; inability to work declined 13.7% due to innovation • Procedure code growth linked to conditions via Medicare coverage determinations • Findings inform health technology assessment evaluation windows and research funding policy decisions To assess whether innovation in medical procedures and products—proxied by the expansion of Healthcare Common Procedure Coding System (HCPCS) codes linked to Medicare Coverage Determinations (MCDs)—is associated with declines in disability among U.S. adults from 1997–2019. We link HCPCS codes to ICD condition categories using Local Coverage Determinations and merge these with Medical Expenditure Panel Survey (MEPS) data to construct 21 condition-year disability indicators. We estimate two-way fixed-effects models with distributed lags (0–15 years) at the condition-year level, controlling for prevalence, mean age, educational attainment, and comorbidity counts, with year and condition fixed effects; standard errors are clustered by condition. The analytic file includes ∼1.50 million condition observations from 317,000 people. For 19 of 21 disability indicators, at least some lagged innovation coefficients are negative and statistically significant. The mean time from innovation to measurable disability reduction is 11.5 years, consistent with diffusion and time-to-benefit dynamics. Estimated 1997–2019 disability reductions attributable to prior innovation include: Supplemental Security Income (SSI) recipiency −21.3% (largest effect), with double-digit declines for inability to work, Social Security recipiency, and school limitations; the median reduction across indicators is ∼7%. Growth in Medicare-covered technology—measured via HCPCS/MCD linkages—is associated with meaningful, long-run reductions in multiple dimensions of disability. Findings highlight the importance of accounting for diffusion lags in health technology assessment and suggest value in monitoring coverage-enabled innovation alongside real-world outcomes.