342 72-Hour Returns: A Trigger Tool for Diagnostic Error

Abstract

Patients who return to the emergency department (ED) within 72 hours of discharge are thought to represent failures in diagnosis, treatment and discharge planning. As a result, this metric is often seen as a surrogate for quality and used for ED quality assurance efforts. Little is known about the yield of this kind of review and the types of errors it identifies. The objective of this study was to identify the prevalence of error deficient care in patients who return within 72 hours and require admission, as well as the types and severity of harm in these patients. Retrospective review of all patients who presented to an urban, university-affiliated ED between October 1st, 2012 and September 30th, 2015 who were evaluated, discharged, and returned within 72 hours requiring hospital admission. As a part of departmental quality assurance activities, the records were reviewed by attending physician members of the Quality and Safety Committee. Any visit felt to represent substandard care was flagged for further investigation. These cases then received a secondary review by a different ED attending who classified the deviation in care for this study, if present, using a standard classification system. There were 413,167 ED visits during the study period with 2001 (0.48%) patients who returned within 72 hours and were admitted to the hospital. An event requiring further investigation was identified in 59 of these patients (2.95% of all 72-hour returns), and 50 of them (2.49% of all 72-hour returns and 0.0119% of all ED visit) were deemed to represent a deviation from optimal care (see Figure 1). Of these cases, 48 (96%) represented diagnostic error. When a standard diagnostic process of care framework was applied to these 48 encounters, the majority of cases represented failures in the initial diagnostic pathway (29 cases, 60.4%), including a failure to obtain a complete history and conduct a complete physical exam (3 cases, 6.2%), failure of ongoing monitoring of clinical status (4 cases, 8.3%), failure to establish a differential diagnosis (20 cases, 41.6%) and failure to order the appropriate diagnostic test (2 cases, 4.1%). The remainder of the cases represented failures in completing testing and results processing (13 cases, 27%), and failures in follow-up and coordination (6 cases, 12.5%). When the National Patient Safety Foundation’s Error Severity Codes were applied to the 50 cases that represented deviations from optimal care, 12 (24%) resulted in minor temporary harm, 30 (60%) in major temporary harm, 4 (8%) in minor permanent harm, 1 (2%) in major permanent harm and 3 (6%) resulted in death. Standard screening methodology of 72-hour ED returns resulting in admission has relatively low yield in identifying suboptimal care, with less than 3% of cases reviewed representing deviations from standard care. Of these, the majority represent cognitive errors in the diagnostic pathway. As such, these reviews may be useful as a tool for ongoing professional practice evaluation of individual clinicians, however likely serve less value in identifying departmental systems issues contributing to unsafe care. Future study should focus on understanding the yield of other departmental QA triggers such as safety reports and morbidity and mortality conferences.