Back

Patient and Hospital Characteristics Associated with Disparities in Acute Stroke Treatment in Community and Academic Hospitals

Shen, J.; Corlin, L.; Cote, K. E.; Pudlo, M. E.; Leung, L. Y.

2023-09-26 neurology
10.1101/2023.09.25.23296119 medRxiv
Show abstract

ImportanceSystemic biases and barriers may affect identification, emergency transportation, and care coordination for people with stroke. We assessed patient- and hospital-level factors for associations with pre-hospital and emergency department care processes. We compared trends for patients first presenting to an academic medical center versus community hospitals. ObjectiveAssess whether patient and hospital characteristics were associated with differences in emergency medical services utilization, stroke code activation, and time to treatment for patients first hospitalized with stroke at academic medical center versus community hospitals. We hypothesized that disparities exist by patient characteristics (e.g., race and ethnicity, primary language) within each clinical setting, and that differences exist across the clinical settings. DesignRetrospective cohort study using data from the electronic health record at an academic medical center (Tufts Medical Center). SettingTertiary care referral center. Participants542 patients aged [≥]18 years hospitalized with stroke (96% with acute ischemic stroke) between 1/1/2018-12/31/2020, including patients who presented directly to the academic medical center and patients transferred from community hospitals. Main OutcomesEmergency medical services use, stroke code activation, door-to-computed tomography time, and door-to-needle time. ResultsAcademic medical center non-Hispanic Asians (odds ratio (OR)=0.25; 95% confidence interval (CI)=0.13-0.47) and Hispanics (OR=0.19; 95% CI=0.05-0.72) and community hospital non-Hispanic Black/African-Americans (OR=0.17; 95% CI=0.05-0.62) were less likely to use emergency medical services compared to non-Hispanic whites. Patients with non-English primary language were less likely to use emergency medical services (OR=0.38; 95% CI=0.23-0.63) compared to English-speaking patients overall, in academic medical centers (OR=0.46; 95% CI=0.25-0.83) and community hospitals (OR=0.18; 95% CI=0.06-0.51). Community hospital Hispanics were less likely to have stroke code activation (OR=0.24; 95% CI=0.05-0.86) compared to non-Hispanic whites. Patients first presenting to a community hospital versus an academic medical center were less likely to have stroke code activation (OR=0.12; 95% CI=0.07-0.19), had shorter door-to-computed tomography time (31% shorter; 95% CI=15-43% shorter), and had longer door-to-needle time (29% longer; 95% CI=5-58% longer). ConclusionsPatient-level factors and hospital setting were associated with differences in acute care suggesting opportunities for community outreach on emergency medical service use, interventions to alleviate language barriers, and approaches to address systemic racism affecting stroke care. Key PointsO_ST_ABSQuestionC_ST_ABSAre there differences in acute stroke care between patients first hospitalized at academic medical centers versus community hospitals? FindingsIn this retrospective cohort study (n=542), academic medical center patients identifying as non-Hispanic Asian or Hispanic, community hospital patients identifying as non-Hispanic Black/African-American, and patients with a non-English primary language were less likely to use emergency medical services. Community hospital patients identifying as Hispanic were less likely to have stroke code activation. Community hospital patients were less likely to have stroke code activation, had shorter door-to-computed tomography time, and had longer door-to-needle time. MeaningTargeted intervention and outreach are needed.

Matching journals

The top 4 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.