Lauren Parlett, Yiling Yang, Cecilia Huang, Amanda Neikirk, Amy Davidow, Kevin Haynes, Tobias Gerhard, Carlos Rose, Brian Strom, and Daniel Horton
Background: The COVID-19 pandemic profoundly affected health care delivery and receipt. Changes in the management of juvenile idiopathic arthritis (JIA) have garnered less attention, in part due to the lower levels of COVID-19-associated morbidity and mortality in children. As a chronic disease in children, JIA is a valuable use-case. Objectives: Our hypothesis was that the pandemic led to fewer outpatient visits, fewer initiations of disease-modifying anti-rheumatic drugs (DMARDs), fewer intraarticular glucocorticoid (iaGC) injections, and more use of higher-risk, short-term therapies, including oral glucocorticoids (oGC) and opioids. Methods: Using commercial health plan members, we found incident JIA occurring Mar 2016 to Feb 2021 based on previously used algorithms. We required ≥12 months of enrollment prior to diagnosis and age < 18 on diagnosis. Anchored on 01Mar2020, we studied 90-day quarters (Q1 was Mar-May) in 2018-21 to determine rates of outpatient visits, DMARD initiation, iaGC injections, and dispensings of oGC and opioids. Rates were modeled with quarter, sex, region, age, JIA type, uveitis (DMARD and visits only), pediatric disease burden (visits only), SES quartile (visits only), and stratified by diagnosis recency (incident was within 1 year after diagnosis). We present adjusted percentages (95% CI) comparing 2020 to prior years. Results: In our cohort of 1294 patients with JIA, 71% were female, mean diagnosis age was 11 (SD 4.5), 15% had systemic JIA, and 131 were diagnosed in 2020. For all patients with JIA, outpatient visits declined by 11% (8%,14%) in 2020-21, most markedly in Q1 with later rebound. DMARD initiation showed no difference in rates [incident: -46% (-75%, 65%); prevalent: 5% (-56%, 151%)], though prevalent patients crude rates showed a sharp Q1 drop. iaGC injections showed no difference in either group [incident: -5% (-66%, 166%); prevalent: -27% (-68%, 24%)]; however, oGC was cut by nearly half [incident: -44%(-71%, 8%); prevalent: -43% (-62%, -15%)]. Opioid dispensing showed a 54% decrease among prevalent patients (-75%, -18%), but no difference among incident patients [-6% (-70%, 193%)]. Conclusions: Modest decreases in healthcare and drug utilization were observed early in the pandemic. Patients with prevalent JIA experienced overall decreases in JIA treatment utilization, delayed initiation of DMARDs, and reduced rates of oGC and opioids than in prior years. Unexpectedly, these results imply reduced use of potentially toxic and inappropriate short-term treatments for patients with prevalent JIA. Small outcome counts for incidence patients resulted in wide variance and no clear interpretation. More work is needed to understand how these changes in management influenced outcomes in this population.