2023 HSR&D/QUERI National Conference

4016 — The Impact of Including Medicare Data to Research on Veterans with COVID-19

Lead/Presenter: Kristin de Groot,  Resource Center - VIReC
All Authors: de Groot K (VA Information Resource Center), Kan D (VA Information Resource Center) Zhang Q (VA Information Resource Center) Weismantle K (VA Information Resource Center) Stevens J (VA Information Resource Center) Souden M (VA Information Resource Center)

Many Veterans who test positive for COVID-19 in the VA also receive healthcare outside the VA that is paid for by Medicare. The purpose of this study is to assess Medicare utilization for these Veterans and to identify which groups are most likely to utilize Medicare in the 2-month period following COVID-19 diagnosis (post-COVID period). We will also determine the benefit of adding Medicare data to ascertain additional comorbidities for Veterans with COVID-19.

Veterans who tested positive for COVID-19 in a VA facility were identified from the VA’s COVID-19 Shared Data Resource. Due to the lag in Medicare data availability, the cohort was limited to Veterans who tested positive prior to February 2022 to include 60 days of follow-up. Medicare fee-for-service claims data were examined during the post-COVID period and logistic regression was used to examine predictors of Medicare utilization. Two variations of the Charlson Comorbidity Index (CCI) were calculated with 12 months of pre-COVID-19 data, one using only VA data and one using combined VA and Medicare data.

We identified 582,959 Veterans who tested positive for COVID-19 in VA facilities between March 2020 and February 2022. Not surprisingly, almost all (98%) Veterans 65+ were enrolled in Medicare and 43% had Medicare utilization. More than half (57%) of Veterans with Medicare utilization also had a COVID-19 diagnosis in Medicare. We identified almost 14,000 Medicare inpatient stays during the post-COVID period, 77% of which included a COVID-19 diagnosis and 37% included ICU days. Predictors of Medicare utilization were examined for Veterans 65+. The odds of Medicare utilization in the post-COVID period were highest among Veterans in VA Priority Group 4, Veterans of white race, and increased significantly with age and number of comorbidities found in the VA data. When using only VA data, 10% of Veterans had a CCI score of 3 or higher, whereas using combined VA and Medicare data, 20% had a score of 3 or higher. Veterans age 85+ or in VA Priority Group 4 had the greatest disparity between the VA CCI index and the VA and Medicare CCI index. For all 17 conditions contained within the CCI, the number of patients identified with each condition doubled when adding Medicare data.

Many Veterans who tested positive for COVID-19 in VA received post-COVID care outside of VA paid for by Medicare, including almost half of those age 65+. Veterans most likely to use Medicare after testing positive in VA were older and had more comorbidities. Including Medicare data allowed for the identification of significantly more comorbidities in older Veterans compared to using VA data only.

Researchers should add Medicare data to their analyses to obtain a more complete picture of healthcare utilization post COVID-19 infection. Comorbidities identified utilizing only VA data should be considered incomplete for Veterans who receive healthcare outside VA. In the future, as the data become available, other sources of non-VA healthcare should also be considered for inclusion, such as Medicaid and Medicare Advantage.