This past month a pair of articles were published online in the EMJ concerning diagnostic error in the Emergency Department. This pair of articles, with the fabulous Hardeep Singh as senior author on each, attempt to describe the underlying foundations of error.
These are particularly important in the context of the ongoing resource constraints facing medical care. A common strategy to reduce the potential for error, and, certainly, the “easy way out”, is simply to overtest. This strategy seems to be associated with a pattern of reduced medicolegal liability, although it is a bit of a stretch to apply this surrogate to diagnostic accuracy. However, besides the obvious up-front costs associated with overtesting, there is an under-appreciated incidence of false-positive results in the setting of low pretest probability.
The first of two articles, by Medford-Davis, focuses on the errors associated with abdominal pain presentations in the Emergency Department. These authors reviewed patients with an initial presentation of abdominal pain, who then subsequently returned to the same institution within 14 days and were admitted. These cases, totalling 100 in all, were retrospectively examined for potential diagnostic error.
Rather distressingly, a full one-third of the return visits may have resulted in part to some diagnostic error on the initial visit, although the reviewers were not in good agreement on the presence of error in all cases. A small number of adverse outcomes were related to non-remedial patient factors, such as poor compliance, poor follow-up, and other factors relating to the study occurring at a facility comprised of an inner-city, indigent, and undocumented immigrant population. The bulk of errors, however, are tucked into the patient encounter itself. The initial root of errors is in the initial evaluation – if elements of history or physical exam are inadequately described, there is a cascading effect of missed indicated testing. Then, a second roadblock seemed to occur on the processing of information resulting from initial orders. Many patients returning for subsequent admission had abnormal studies that were incompletely addressed at the initial visit.
The second article, by Okafor, analyzed the content of cases from an institutional error reporting system. Rather than focusing on the specific phase of care in which errors occurred, these authors focus on the primary domains of error. They found, unsurprisingly, the cognitive dimensions of errors are complex, and include issues such as faulty data gathering, premature closure, and misinterpretation of results. However, they importantly observe errors rarely occurred in isolation. Cognitive errors were frequently complicated by process or resource-related issues, as well as simple factors relating to patient complexity. In all, three-quarters of errors resulted from a combination of factors.
The basic takeaway here probably relates to some of the most difficult issues to address in the Emergency Department. Diagnostic accuracy will be maximized when properly resourced physicians have the time they need to evaluate their most complex patients. Resources, time, and complexity, however, are not frequently seen in their optimal concentrations in the ED – but it is important to recognize any quality improvement efforts would be remiss without accounting for these factors, and focusing solely on individual physician remediation.