This descriptive study assessed how recorded incidence of asthma, atopic dermatitis (AD), and multiple sclerosis (MS) changed before, during, and after the COVID-19 pandemic in a large US administrative claims database. The objective was to describe observed seasonal incidence rates (IRs) from 2018 through 2022 and compare each seasonal cohort in 2019–2022 to the corresponding season in 2018, with attention to age and sex subgroups. The authors framed the work as addressing methodological implications for incidence estimation using claims data across the pandemic period rather than testing causal effects of COVID-19 or countermeasures.
The study used Optum’s de-identified Clinformatics Data Mart Database (CDM), containing administrative claims for commercial and Medicare Advantage enrollees across all 50 US states. To account for seasonality, investigators created 20 season-based cohorts spanning winter 2018 (Dec 1, 2017–Feb 28, 2018) through fall 2022 (Sep 1–Nov 30, 2022). For each seasonal cohort, eligible individuals met five criteria: enrollment in the season for at least one day; 12 months of continuous enrollment prior to the index date; age ≥6 years; known sex; and absence of the outcome of interest in the prior 12 months. The index date was the earliest date meeting eligibility during the season.
Incident asthma and AD required at least two inpatient or outpatient ICD-10-CM diagnostic codes for the condition 30–365 days apart, with the date of the first qualifying code defined as the outcome date. AD codes included L20.0, L20.81, L20.82, L20.84, L20.89, and L20.9; asthma used J45. Incident MS required ≥2 ICD-10-CM codes for MS (G35) 30–365 days apart plus at least one dispensing for a disease-modifying therapy (DMT) within one year after the first qualifying code. For MS the authors noted previously reported high positive predictive value and sensitivity for this combined definition. The first qualifying code had to fall within the seasonal cohort timeframe; the second code and DMT dispensing could occur up to one year later.
Within each seasonal cohort, individuals were followed from the index date until outcome occurrence, death, disenrollment, or the end of the season. Incidence rates were calculated as the number of incident cases divided by person-time at risk (person-years) within each seasonal cohort. Exact methods were used to estimate 95% confidence intervals. Incidence rate ratios (IRRs) compared each seasonal cohort in 2019–2022 to the matching season in 2018 using Poisson regression with exact methods. Age- and sex-standardized IRs and IRRs used the 2022 US population composition by age and sex as the reference. Subgroup analyses estimated IRs and IRRs by age groups (6–11, 12–17, 18–64, ≥65 years) and by sex. Analyses were performed in SAS v9.4.
Across roughly 11 million individuals per seasonal cohort (average age 49–51 years, approximately 51% female), the authors observed a clear decline in recorded incidence for all three outcomes in spring 2020 compared with spring 2018. Reported reductions for spring 2020 were: asthma IRR 0.86 (14% decrease), AD IRR 0.72 (28% decrease), and MS IRR 0.77 (23% decrease). IRs for AD and MS returned to or exceeded pre-pandemic levels by summer 2020; asthma IRs took longer to recover, returning to or exceeding 2018 levels by spring 2021. Post-pandemic increases relative to 2018 were reported of up to +12% for asthma (since spring 2021), up to +35% for AD (since summer 2020), and up to +24% for MS (since spring 2021).
The early-pandemic declines were most marked in pediatric and older adult groups. For asthma in spring 2020, children aged 6–11 showed the largest reductions (female IRR 0.38, male IRR 0.40), and adolescents 12–17 had approximately 50% lower IRs for both sexes. Senior adults (≥65 years) also showed notable decreases (IRRs ~0.71–0.73). AD declines during spring 2020 affected all age and sex subgroups, with 6–11-year-olds most affected (IRR ~0.52). MS reductions were more pronounced in older adults (≥65 years) than in younger adults; IRs in those under 18 were not reported due to small counts. The analyses did not demonstrate consistent sex differences in the relative changes of IRs across periods.
The authors interpret the spring 2020 declines primarily as an apparent drop in recorded incidence driven by reduced healthcare utilization: suspension of non-urgent services, lower referrals, redeployment of clinical resources, and patient avoidance or fear of in-person visits. For pediatric asthma, the authors acknowledge that reduced exposure to respiratory viruses and other pandemic-related behavior changes (school closures, masking) may have contributed to a genuine reduction in triggers leading to diagnoses or exacerbations. The differing timing of the “catch-up” across conditions—earlier for AD and MS, later for asthma—suggests outcome-specific patterns in access and healthcare-seeking recovery.
The study highlights that data spanning the pandemic period may show artificially lowered incidence estimates for certain conditions, particularly in spring 2020 and disproportionately affecting children and older adults. Investigators designing incidence or prevalence studies that include 2020 data should exercise caution; the authors suggest considering exclusion of the early-pandemic period from trend analyses or ensuring sufficient post-pandemic follow-up (2021 and onward) to mitigate bias from delayed or missed diagnoses.
Limitations noted by the authors include reliance on administrative claims with potential for misclassification, the descriptive (non-causal) design, and that the Optum CDM data are licensed and not publicly available. The study was funded by Sanofi; several authors were employees of Sanofi and the funder purchased the data. The authors stated that the funder did not have additional roles in study design, analysis, or publication decisions. These disclosures and data-access details are reported in the source article.