Enhancing understanding of clinician cognitive limitations will be useful in helping clinicians engage in more accurate diagnostic reasoning and in avoiding harmful diagnostic errors. CLT offers a useful framework for understanding diagnostic reasoning and inherent cognitive limitations. Although imperfect, methods for measuring cognitive load do exist.
More research into the causal pathways is needed, but we think we can reasonably state that high cognitive load likely does impact diagnostic accuracy. With this premise in mind, we propose the following hypotheses as the most important to further explore (Table 1). If they prove correct, we have suggested ways to better protect clinician cognitive load with the subsequent assumption that diagnostic accuracy should improve.
Table 1. Cognitive load hypotheses and opportunities to develop enhanced diagnostic accuracy
| Hypothesis | Opportunities for Implementation |
|---|---|
| Cognitive load is an independent variable that affects diagnostic accuracy and should be accounted for when designing clinician workforce structures. |
|
| EHR audit log data can serve as a proxy for cognitive load and should be used to help create EHR interventions that protect clinician cognitive load. |
|
| Human factors engineers and system usability testing can create technology that decreases clinician cognitive load. |
|
| Cognitively protected physical spaces decrease extrinsic cognitive load to allow more capacity for intrinsic and germane cognitive processing. |
|
| Cognitive load theory should be integrated into organizational change efforts to better protect healthcare worker cognition. |
|
Clinical Example
What might an environment that focuses on optimizing the cognitive environment look like? A clinical example is again helpful. Consider the hospitalist admitting the patient with congestive heart failure and, unbeknownst to them, multifocal pneumonia. Because this hospitalist works for a division that understands the limits of cognition, their leadership strategically incorporates additional measures of workload via the NASA-TLX. This approach allows customizable limits on the number of patients the hospitalist is expected to care for and acknowledges that provider workload is likely impacted by total census and patient complexity.
In addition, at this hospital, cognitive task analysis was coupled with task-evoked pupillary eye tracking and heart rate variability measurements during EHR interface development. This analysis resulted in an information visualization system that minimizes extrinsic cognitive load by avoiding complicated visuals and data overload. The hospital’s EHR has also been customized to allow various levels of electronic messaging urgency, which helps the hospitalist designate protected time during high-risk and error-prone periods and hold nonurgent interruptive alerts, thereby reducing multitasking.
Finally, the emergency room has been thoughtfully designed to reduce alarm fatigue and protect patient rooms from loud noises and unnecessary visual stimuli. Because the cognitive environment has been optimized, the hospitalist avoids cognitive overload and can appropriately engage in dual-process thinking; they realize that their patient has both heart failure and multifocal pneumonia. Appropriate antibiotics are started and sepsis, with its associated high morbidity and mortality, is avoided.
Conclusion
Diagnostic accuracy is garnering increasing interest from researchers, frontline clinicians, and healthcare systems focused on creating safer environments for patients. CLT provides a framework that can be used to better understand cognition, along with its limitations. Further investigation into the interplay between cognitive load and diagnostic accuracy will help create strategies to support clinicians as they optimize their diagnostic accuracy.
