Evaluating the Usability of a Dosimetry Quality Assurance Checklist and Associated Workload, Performance and Patient Safety in Clinical Settings

Abstract

<h3>Purpose/Objective(s)</h3> The majority of errors during radiation therapy (RT) treatment planning and associated quality assurance (QA) processes are ascribed to ‘human error'. Most centers typically use in-house QA checklists but without formal human factors engineering evaluations. We herein comprehensively evaluate the perceived usability of a dosimetry-QA checklist and associated cognitive workload, performance, and patient safety in real clinical settings at an academic medical center. <h3>Materials/Methods</h3> Four dosimetrists (female = 2, male = 2) performed pre-treatment QA tasks on 76 different RT plans with varying degrees of difficulty (low (35), medium (35), and hard (6) based on subjective rating by physicists) in the clinic. For each plan, the perceived usability and workload were assessed using a variety of validated tools (see results). We compare these scores with the recommended standards in human-computer interaction and human factors and ergonomics. Performance was measured using the time taken to complete the QA tasks on 46 different RT plans. Patient safety was measured as the number of errors leading to patient harm. <h3>Results</h3> See Table 1. The perceived usability of the dosimetry-QA checklist meets only 1 out of 6 standards assessed. The workload assessments are within the recommended standards. The average time taken to complete pre-treatment QA tasks is 37.25 minutes. No errors leading to patient harm were reported. <h3>Conclusion</h3> Our institutional dosimetry-QA checklist has suboptimal usability, but the associated workload is acceptable. QA checklists should be evaluated for usability and redesigned to reduce the risk of human errors, which are likely, at least in part, to result from the use of sub optimally designed tools. Future studies evaluating the association between usability and workload of QA checklists with patient safety indicators may be helpful.