Point-of-care Ultrasound in Morbidity and Mortality Cases in Emergency Medicine: Who Benefits the Most?

Who Benefits the Most?

Andrew J. Goldsmith, MD, MBA; Hamid Shokoohi, MD MPH; Michael Loesche, MD, PhD; Ravish C. Patel; Heidi Kimberly, MD; Andrew Liteplo, MD


Western J Emerg Med. 2020;21(6):172-178. 

In This Article


Study Setting and Population

This retrospective study was done at two large academic EDs with annual volumes of 120,000 and 70,000 patients. Both institutions have an emergency ultrasound (US) division, emergency US fellowship program, and share a four-year EM residency training program with 60 EM residents postgraduate years 1–4 (PGY). This study was reviewed by the institutional review board and determined to be exempt.

Selection of Participants

Cases were reviewed monthly in the departmental M&M conference as part of routine departmental quality assurance. PGY-4 EM residents prepared M&M cases for review with a faculty EM attending physician as part of this process, and not for research purposes. All PGY-4 EM residents were asked to participate in the study survey. Participation was voluntary. There were no exclusion criteria.

Study Design

All ED cases were subject to review from July 2018–June 2019. All cases that resulted in a death in the ED, all deaths within 24 hours of an ED encounter, and all upgrades to an ICU within 24 hours were automatically reviewed for possible clinical or system errors. In addition, any cases referred by nursing, ED providers, and providers from other departments were reviewed.

For each M&M session, a PGY-4 resident was provided a list of all cases for review over a designated time period. This resulted in approximately 60–75 cases over about a 40-day period. Each M&M conference was specific to one hospital. All information was obtained by retrospective chart review. After reviewing each case, the PGY-4 resident submitted a summary of each case to a faculty mentor, an attending physician responsible for departmental M&M review. Together, the PGY-4 resident and attending physician identified all cases that were considered to have potential patient care concerns while in the ED. This review was done as routine departmental quality assurance and not for purposes of the study.

After all cases were reviewed, a study investigator surveyed the PGY-4 residents about all of the cases in which there were possible concerns about patient care as determined by the PGY-4 EM resident and an EM attending. The survey addressed questions regarding the use of POCUS in M&M cases. Specifically, the resident was asked:

  • In cases when POCUS was performed, did POCUS contribute to the M&M?

  • In cases when POCUS was not performed, would it likely have prevented the M&M if it had been done?

  • If so, which application(s) would have helped, and how?

Residents were instructed that when assessing the potential of POCUS to prevent M&M, they should assume that POCUS would have been appropriately performed, interpreted, and integrated. In addition, we collected information regarding patients' initial chief complaints and initial triage vital signs. An US fellow and/or an US fellowship-trained EM attending administered the survey. Verbal consent was obtained for all participants. The PGY-4 resident was blinded to the purposes of the study. An EM attending with fellowship training in POCUS was then presented with the same cases. The attending, blinded to the resident's assessment, was then asked the same questions. Additionally, the attending was asked specifically if he or she would have performed a POCUS if presented with the same clinical case and timeline.

We collected and managed study data using REDCap (Vanderbilt University, Nashville, TN) electronic data capture tools hosted at Massachusetts General Hospital.[14,15] The same software stored and de-identified all demographic and clinical data obtained.

Data Analysis

All data obtained was de-identified, exported to, and analyzed in Microsoft Office 365 Excel (Microsoft Corporation, Redmond, WA). We used descriptive statistical analysis to compare the data. We calculated the overall percentage of cases where M&M may have been affected by POCUS, as assessed by a PGY-4 resident. We then performed subgroup analyses stratified by chief complaint, vital signs, type of POCUS, and how POCUS may have affected M&M. A kappa value was then calculated for interobserver agreement between the PGY-4 residents and the US EM attending. We calculated proportion confidence intervals (CI) of 95% using our sample sizes.