U.S. flag

An official website of the Department of Health and Human Services

SHARE:

FacebookXEmail AHRQPrintShare

Connected Emergency Care Patient Safety Learning Lab (CEC PSLL)

Principal Investigator:Jeremiah Hinson, M.D., Ph.D., Johns Hopkins University, Baltimore, MD; formerly Scott Levin, M.S., Ph.D., Johns Hopkins University, Baltimore, MD
AHRQ Grant No.: HS26640
Project Period: 09/30/18-12/31/23
Description: The goal of CEC PSLL was to improve care in the emergency department (ED) by designing technological tools such as clinical decision support (CDS) that further connected physicians to patients’ pre-, post-, and intra-encounters.

The specific aims were to:

  1. Optimize diagnostic performance for patients with suspected respiratory infection.
  2. Increase the specificity of antibiotic treatment for patients with respiratory infection.
  3. Improve transition of care outcomes after the ED encounter was complete by reducing: (a) unnecessary hospitalizations and sudden care-level changes for those admitted; and (b) 30-day postencounter acute care use for those discharged.

CEC PSLL’s multidisciplinary team of physicians, nurses, pharmacists, administrators, and engineers followed a five-phase innovation cycle (i.e., problem analysis, design, development, implementation, and evaluation). They used agile methodology to create a data infrastructure and resources to problem solve critical gaps in care, particularly in the management of infectious diseases. Resources included an electronic health record (EHR) data repository, machine learning (ML) models, and a platform for individualized audit and feedback.1

When COVID-19 emerged in 2020, the lab shifted its focus to make important contributions with its data science assets, including an EHR-derived data repository, to study respiratory infections. This adjustment allowed researchers to meaningfully support institutional and regional decision making during the pandemic.1 For example, researchers could measure the relative frequency of SARS-CoV-2 infection across racial and ethnic groups. Of 37,727 patients, more than 16 percent (n=6,162) tested positive for SARS-CoV-2. Latino patients’ positivity rate was significantly higher (42.6%) compared with White patients (8.8%), Black patients (17.6%), and patients of other racial/ethnic backgrounds (17.2%).

The lab’s data architecture served as the backbone for this project and helped develop ML models that estimated short-term risk of clinical deterioration: 

  • CEC PSLL researchers collaborated with their leaders and a research team funded by the Centers for Disease Control and Prevention to rapidly develop and deploy a CDS system to improve decision making for patients diagnosed with or suspected of having COVID-19. 
  • Researchers worked with designers at the Maryland Institute College of Art to create Linking Outcomes Of Patients—a patient outcomes feedback platform—for emergency medicine clinicians. This platform used EHR data to give clinicians insights into short- and long-term patient outcomes. 
  • Researchers developed an antibiotic-prescribing audit and patient feedback system. The system used EHR data audit and peer comparison with behavioral feedback to reduce inappropriate antibiotic prescribing for acute respiratory infections.

These data-informed CDS tools paved the way for more informed, efficient, and equitable healthcare practices less susceptible to racial and socioeconomic biases.1-4 

Researchers also used the lab’s data science assets to understand the influence of race and socioeconomic status on antibiotic prescribing for acute respiratory infections in the ED. Multivariable regression was used to examine the association between patient demographics (race, ethnicity, area deprivation index) and likelihood of receiving an inappropriate antibiotic prescription, controlling for variables such as age, sex, and comorbidities.1,5

Of 147,401 ED patients, 7.6 percent of cases involved inappropriate antibiotic prescribing, with White patients being 1.32 times as likely to receive inappropriate prescriptions. In addition, disparities were evident along socioeconomic lines, with patients from wealthier areas more likely to receive inappropriate antibiotics than those from areas with greater socioeconomic deprivation.1,5

To date, this PSLL’s work has resulted in at least 18 peer-reviewed publications that have been cited more than 230 times in other publications, as well as posters and presentations at conferences across the United States.

Publications

2024

2023

2022

2021

2020

References

  1. Hinson J. Final Report: Connected Emergency Care (CEC) Patient Safety Learning Lab. Baltimore, Maryland: Johns Hopkins University School of Medicine, Department of Emergency Medicine; 2024, pp. 1-12.
  2. Hinson JS, et al. Multisite implementation of a workflow-integrated machine learning system to optimize COVID-19 hospital admission decisions. NPJ Digit Med 2022 Jul 16;5(1):94.
  3. Stonko DP, et al. A pilot machine learning study using trauma admission data to identify risk for high length of stay. Surg Innov 2023 Jun;30(3):356-365.
  4. Hinson JS, et al. Multisite development and validation of machine learning models to predict severe outcomes and guide decision‐making for emergency department patients with influenza. J Am Coll Emerg Physicians Open 2024 Apr;5(2):e13117.
  5. Klein E, et al. Racial and socioeconomic disparities evident in inappropriate antibiotic prescribing in the emergency department. Ann Emerg Med 2024 Aug;84(2):101-110.
Page last reviewed January 2025
Page originally created January 2025

Internet Citation: Connected Emergency Care Patient Safety Learning Lab (CEC PSLL). Content last reviewed January 2025. Agency for Healthcare Research and Quality, Rockville, MD.
https://www.ahrq.gov/patient-safety/resources/learning-lab/connected-emergency-care-long-desc.html

Click to copy citation