Shifts in Multiple Myeloma Care Utilization of Veterans in the COVID-19 Era: Interrupted Time Series Analysis of Veterans Health Administration Data

Introduction The COVID-19 pandemic profoundly impacted medical care delivery and ushered in an era of telemedicine and virtual care. However, shifts in multiple myeloma (MM) care utilization patterns as a consequence of the COVID-19 pandemic are poorly characterized. We utilized electronic medical records from the Veterans Health Administration (VA), America's largest integrated health system, to study MM care utilization during the first year of the COVID-19 pandemic.

Methods We retrospectively analyzed medical records of veterans with MM who received care in the VA system and were prescribed MM oral medications from Jun 2019 to Dec 2020. We examined the following outcomes: proportion of in-person visits over all outpatient visits, number of oral MM prescriptions, number of laboratory tests (hemoglobin, creatinine, serum protein electrophoresis [SPEP], free light chain [FLC] ratio), and number of inpatient admissions. As state lockdowns commenced in Mar 2020, we defined the pre-lockdown period as Jun 2019 to Feb 2020 and the post-lockdown period as Apr 2020 to Dec 2020.

We performed interrupted time series (ITS) analysis on each outcome to assess for the step change between the pre- and post-lockdown periods via autoregressive moving average modeling. There were 9 time points in each period that corresponded to month. The counterfactual outcome was no observed change in defined outcomes in the post-lockdown compared to the pre-lockdown period.

Results We identified 3679 veterans with MM who met inclusion criteria; they received care in all 50 states. The median age was 72 years; 67% were white and 31% were black.

We assessed 93,455 outpatient visits attended by all 3679 veterans (Figure 1). Email, phone, and telehealth visits were considered virtual care. In the pre-lockdown period, in-person appointments stably accounted for approximately 93% of all visits, sharply dropping to 68% in Apr 2020. ITS showed a significant decrease of 26% in the percentage of in-person outpatient appointments (-0.26, [95% confidence interval (CI): -0.23 to -0.28], p<0.0001) comparing the post-lockdown to the pre-lockdown period.

We examined 37,388 oral MM medication prescriptions of all 3679 veterans (Figure 2). ITS showed no significant changes in the monthly number of ixazomib (+27, [95% CI: +60 to -6], p=0.12), lenalidomide (+22, [95% CI: +112 to -68], p=0.65), and pomalidomide (+58, [95% CI: +5 to +111], p=0.05) prescriptions comparing the post-lockdown to the pre-lockdown period.

We analyzed 326,415 laboratory tests collected from 3612 veterans (98% of study group). ITS showed significant decreases in the monthly number of performed creatinine (-1106, [95% CI: -395 to -1816], p<0.01), hemoglobin (-1096, [95% CI: -341 to -1851], p<0.01), and SPEP (-98, [95% CI: -39 to -157], p<0.01) tests comparing the post-lockdown to the pre-lockdown period. FLC testing was non-significant (-37, [95% CI: +147 to -221], p=0.70).

We evaluated 2857 admissions in 1286 veterans (35% of study group). ITS showed no significant change in the monthly number of admissions (-24, [95% CI: +9 to -57], p=0.18) comparing the post-lockdown to the pre-lockdown period.

Discussion We retroactively evaluated shifts in MM care utilization of veterans in the first year of the COVID-19 pandemic. After statewide lockdowns began in Mar 2020, in-person visits and laboratory testing significantly decreased as medical care quickly shifted to virtual care. However, prescriptions for MM oral medications were not affected. Similarly, inpatient admissions of veterans with MM did not increase. The impact on clinical outcomes as a result of increasing virtual care in MM remain to be fully assessed, although the lack of major observed shifts in outpatient scripts and inpatient admissions is encouraging.

No relevant conflicts of interest to declare.