tetano
Editor, Senior Moderator
JAMA Netw Open
. 2026 Jun 1;9(6):e2617766.
doi: 10.1001/jamanetworkopen.2026.17766.
Chronic Conditions and Mortality in Medicare Beneficiaries Before and After the COVID-19 Pandemic
R J Waken[SUP] 1 2 [/SUP], Fengxian Wang[SUP] 2 3 [/SUP], Jose F Figueroa[SUP] 4 [/SUP], Rishi K Wadhera[SUP] 5 6 [/SUP], Khavya Avula[SUP] 2 3 [/SUP], E John Orav[SUP] 7 8 [/SUP], Anne Mobley Butler[SUP] 9 [/SUP], Karen E Joynt Maddox[SUP] 2 3 [/SUP]
Affiliations
Importance: There were well-documented changes in health care use during the COVID-19 pandemic. Little is known about whether there were any associated decreases in claims-based comorbidity ascertainment that might have relevance to health services and policy research.
Objective: To quantify differences in claims-based comorbidity assessment in Medicare beneficiaries pre- vs post-COVID-19.
Design, setting, and participants: This cross-sectional study analyzed data from the 30 Chronic Conditions Warehouse 1- or 2-year lookback claims algorithms to ascertain each comorbidity with exact date ranges for all fee-for-service (FFS) and Medicare Advantage (MA) beneficiaries. Data were analyzed from April 2025 to April 2026.
Exposures: The first quarter (Q1) of 2019 and 2022 were evaluated as pre- and post-COVID-19, respectively.
Main outcomes and measures: The main outcomes were comorbidities per beneficiary and association between each comorbidity and mortality in 2019 vs 2022. The changes in prevalence of each comorbidity between 2019 and 2022 were analyzed and fit models within disjoint population subgroups were combined via multilevel meta-analysis models to determine whether each comorbidity's association with mortality changed over time.
Results: This study included 59 514 042 beneficiaries in 2019 (32 351 732 females [54.4%]; 50 814 834 aged 65 years or older [85.4%]) and 63 202 599 beneficiaries in 2022 (34 377 560 females [54.4%]; 55 197 435 aged 65 years or older [87.3%]). The mean number of coded comorbidities per beneficiary decreased from 3.85 to 3.62 for FFS and 4.54 to 4.39 for MA (-0.15) between 2019 and 2022. In FFS, 19 comorbidities (63.3%) decreased, 2 (6.7%) were unchanged, and 9 (30%) increased. In MA, 14 comorbidities (46.7%) decreased, 7 (23.3%) were unchanged, and 9 (30%) increased. In multivariable analyses pooled across FFS and MA, 11 comorbidities (36.7%) were more positively associated with mortality, and 4 (13.3%) were more negatively associated with mortality in Q1 2022 than in Q1 2019.
Conclusions and relevance: Coded levels of many comorbidities in Medicare were lower postpandemic than prepandemic, and the association of these codes with mortality changed. This suggests that across the COVID-19 pandemic comorbidity capture was affected by utilization changes, and the association between comorbidities and mortality changed as a result. Analyses that include immediately pre- or post-COVID-19 data and condition inferences on membership in utilization-based disease groups, or use claims-based risk adjustment, may be subject to bias.
. 2026 Jun 1;9(6):e2617766.
doi: 10.1001/jamanetworkopen.2026.17766.
Chronic Conditions and Mortality in Medicare Beneficiaries Before and After the COVID-19 Pandemic
R J Waken[SUP] 1 2 [/SUP], Fengxian Wang[SUP] 2 3 [/SUP], Jose F Figueroa[SUP] 4 [/SUP], Rishi K Wadhera[SUP] 5 6 [/SUP], Khavya Avula[SUP] 2 3 [/SUP], E John Orav[SUP] 7 8 [/SUP], Anne Mobley Butler[SUP] 9 [/SUP], Karen E Joynt Maddox[SUP] 2 3 [/SUP]
Affiliations
- PMID: 42284057
- PMCID: PMC13263784
- DOI: 10.1001/jamanetworkopen.2026.17766
Importance: There were well-documented changes in health care use during the COVID-19 pandemic. Little is known about whether there were any associated decreases in claims-based comorbidity ascertainment that might have relevance to health services and policy research.
Objective: To quantify differences in claims-based comorbidity assessment in Medicare beneficiaries pre- vs post-COVID-19.
Design, setting, and participants: This cross-sectional study analyzed data from the 30 Chronic Conditions Warehouse 1- or 2-year lookback claims algorithms to ascertain each comorbidity with exact date ranges for all fee-for-service (FFS) and Medicare Advantage (MA) beneficiaries. Data were analyzed from April 2025 to April 2026.
Exposures: The first quarter (Q1) of 2019 and 2022 were evaluated as pre- and post-COVID-19, respectively.
Main outcomes and measures: The main outcomes were comorbidities per beneficiary and association between each comorbidity and mortality in 2019 vs 2022. The changes in prevalence of each comorbidity between 2019 and 2022 were analyzed and fit models within disjoint population subgroups were combined via multilevel meta-analysis models to determine whether each comorbidity's association with mortality changed over time.
Results: This study included 59 514 042 beneficiaries in 2019 (32 351 732 females [54.4%]; 50 814 834 aged 65 years or older [85.4%]) and 63 202 599 beneficiaries in 2022 (34 377 560 females [54.4%]; 55 197 435 aged 65 years or older [87.3%]). The mean number of coded comorbidities per beneficiary decreased from 3.85 to 3.62 for FFS and 4.54 to 4.39 for MA (-0.15) between 2019 and 2022. In FFS, 19 comorbidities (63.3%) decreased, 2 (6.7%) were unchanged, and 9 (30%) increased. In MA, 14 comorbidities (46.7%) decreased, 7 (23.3%) were unchanged, and 9 (30%) increased. In multivariable analyses pooled across FFS and MA, 11 comorbidities (36.7%) were more positively associated with mortality, and 4 (13.3%) were more negatively associated with mortality in Q1 2022 than in Q1 2019.
Conclusions and relevance: Coded levels of many comorbidities in Medicare were lower postpandemic than prepandemic, and the association of these codes with mortality changed. This suggests that across the COVID-19 pandemic comorbidity capture was affected by utilization changes, and the association between comorbidities and mortality changed as a result. Analyses that include immediately pre- or post-COVID-19 data and condition inferences on membership in utilization-based disease groups, or use claims-based risk adjustment, may be subject to bias.