Unraveling Diagnostic Error: Delving Deeply to Identify Hidden Human Factors

In my view, the IOM’s report did not address the impact of social, economic, and workplace stressors on human performance. These stressors may impact cognition, compounding the pitfalls associated with common diagnostic biases and pressure-driven overutilization of clinical heuristics that may direct clinicians away from looking for exceptions to rules and for a broader range of experiences. As a result, then, these stressors affect diagnostic accuracy.

Enhancing quality of investigations

The National Patient Safety Foundation (U.S.) and the Health Quality Council of Alberta (Canada) have developed investigation methodologies (National Patient Safety Foundation, 2015; Duchscherer & Davies, 2012) designed to improve investigations’ quality, consistency, and reliability, thus resulting in actionable recommendations for improvement. These methodologies employ a Donabedian (2005) focus on structures, processes, and outcomes for identifying the factors contributing to patient safety incidents. Both approaches encourage deep diving (Weick & Sutcliffe, 2007), which is vitally important to high-reliability organizations that value the contributions of frontline staff. Both approaches, too, specifically avoid attempts to assign personal blame.

In too many instances, investigations look to assign blame to one or more individuals without fully identifying and understanding the factors that interfere with human performance, both in the diagnostic and procedural realms. In my view, if an investigation concludes that an individual was uniquely responsible for a patient safety incident, then that investigation has not delved deeply enough to identify the human factors contributing to that individual’s error. Only by identifying these human factors will real opportunities for improvements be identified and actionable steps be taken to improve outcomes. Reasons for mistakes must be identified and modulated for us to achieve success, and blame has no role in this.

Investigating the factors affecting human performance may be part of a “root cause” analysis, but more appropriately, it may require a separate investigation that includes experts in human factors analysis, especially if professional performance issues may need to be addressed. The following case studies point out some of the pitfalls of assessing individual human performance within the context of a patient safety incident investigation.

Case study 1: Emergency department chaos and diagnostic delay

A 3-year-old child presented to a busy urban emergency department (ED) with low-grade fever (38oC), lassitude, anorexia, urinary frequency of several days’ duration, and a history of minor urinary tract infections. Her single-parent mother was a recent Haitian immigrant with limited English fluency. The child was initially assessed by a triage nurse, baseline labs were ordered per ED protocol, and the child was assigned a non-urgent priority for physician evaluation.

The child was first seen by a physician two hours later. The physician spoke some French, and she obtained an additional history of polyuria and maternal sickle cell trait. The father’s sickle cell status was unknown. The child looked reasonably well, with no CVA tenderness, and a presumptive diagnosis of urinary tract infection was made, pending lab results. The physician left instructions with a nursing assistant to be contacted when the results, expected shortly, were available. Unfortunately, the laboratory studies, including a blood count, blood culture, electrolyte panel, and urinalysis, were never performed, though the specimens had been obtained at triage. When an ED nurse became aware of this, she prompted the lab, and the studies were then performed after the initial long delay.