Back

Examining inequality in healthcare utilization during pandemic disruptions

Lian, J.; Pei, S.; Gao, J.; Zhong, L.

2025-04-25 health informatics
10.1101/2025.04.24.25326361 medRxiv
Show abstract

ImportanceHealth crisis like recent COVID-19 pandemic has continuously disrupted healthcare utilization, particularly among non-COVID-19 patients. It remains unclear whether these disruptions were experienced equally across populations or if they disproportionately impacted specific demographic groups. ObjectiveTo assess inequality in healthcare utilization during the COVID-19 pandemic, stratified by demographic characteristics (age, race and ethnicity, income, and education) and medical specialties (emergency medicine, anesthesiology, and cardiology) across U.S. states. Design, Setting, and ParticipantsThis retrospective observational study analyzed electronic health records (EHRs) from millions of anonymized patients across U.S. states between 2019 and 2022. The Gini coefficient was used to quantify disparities in healthcare utilization. A regression discontinuity design was employed to assess the effect of declining telehealth usage on inequality. Main Outcomes and MeasuresTemporal inequality in healthcare utilization by demographic group and specialty, measured monthly using the Gini coefficient. ResultsCompared with the pre-pandemic period (2019), there was an approximate 45% decline in patient visits during the pandemic (2020-2022). Inequality in healthcare utilization increased consistently across demographic groups, with Gini coefficients rising by 14.5% (age), 2.3% (race and ethnicity), 7.0% (income), and 4.8% (education). States that maintained high levels of telehealth service usage experienced smaller increases in inequality. Conclusions and RelevanceThe pandemics significantly exacerbated disparities in healthcare utilization among non-COVID-19 patients. These findings underscore the importance of sustaining adaptive healthcare delivery strategies, such as telehealth, to mitigate systemic inequities during long-term public health emergencies.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.