Publications

IMPORTANCE

Although the general US population had fewer emergency department (ED) visits during the COVID-19 pandemic, patterns of use among high users are unknown.

OBJECTIVES

To examine natural trends in ED visits among high users of health and social services during an extended period and assess whether these trends differed during COVID-19.

DESIGN, SETTING, AND PARTICIPANTS

This retrospective cohort study combined data from 9 unique cohorts, 1 for each fiscal year (July 1 to June 30) from 2012 to 2021, and used mixed-effects, negative binomial regression to model ED visits over time and assess ED use among the top 5% of high users of multiple systems during COVID-19. Data were obtained from the Coordinated Care Management System, a San Francisco Department of Public Health platform that integrates medical and social information with service use.

EXPOSURES

Fiscal year 2020 was defined as the COVID-19 year.

MAIN OUTCOMES AND MEASURES

Measured variables were age, gender, language, race and ethnicity, homelessness, insurance status, jail health encounters, mental health and substance use diagnoses, and mortality. The main outcome was annual mean ED visit counts. Incidence rate ratios (IRRs) were used to describe changes in ED visit rates both over time and in COVID-19 vs non-COVID-19 years.

RESULTS

Of the 8967 participants, 3289 (36.7%) identified as White, 3005 (33.5%) as Black, and 1513 (16.9%) as Latinx; and 7932 (88.5%) preferred English. The mean (SD) age was 46.7 (14.2) years, 6071 (67.7%) identified as men, and 7042 (78.5%) had experienced homelessness. A statistically significant decrease was found in annual mean ED visits among high users for every year of follow-up until year 8, with the largest decrease occurring in the first year of follow-up (IRR, 0.41; 95% CI, 0.40-0.43). However, during the pandemic, ED visits decreased 25% beyond the mean reduction seen in prepandemic years (IRR, 0.75; 95% CI, 0.72-0.79).

CONCLUSIONS AND RELEVANCE

In this study, multiple cohorts of the top 5% of high users of multiple health care systems in San Francisco had sustained annual decreases in ED visits from 2012 to 2021, with significantly greater decreases during COVID-19. Further research is needed to elucidate pandemic-specific factors associated with these findings and understand how this change in use was associated with health outcomes.

The COVID-19 pandemic has underscored how underlying disparities in environmental and health conditions exacerbate vulnerability during public health emergencies in low-income and communities of color. Neglected epidemics-high rates of pollution, chronic disease, and racial and socioeconomic health disparities-have continued amid persistent systemic racism and declining investment in public health. Recognized too late due to shortcomings in public health data tracking, COVID-19 has surged through vulnerable communities. Improved public health tracking is critical for informing the country's recovery from COVID-19, and it can be leveraged to measure and reduce health disparities and strengthen community resilience to respond more effectively to the next public health crisis. We emphasize how public health tracking agencies can engage communities in data collection and reporting; we also discuss the complementary role that communities can take to mobilize data to change policies and institutions, strengthening resilience through increased information and capacity driven by community priorities. Success requires the continuous collection of timely data at a community scale, and public health agencies partnering with communities to use the information in decision making and evaluation to ensure progress over time. We highlight community-engaged data collection and reporting-community air monitoring in Imperial County, CA-as an example of working with communities to improve public health data collection and reporting, increase community dialogue and engagement in governmental decision making, and inform public health tracking to reduce health disparities and strengthen community resilience.