Computerised clinical decision support system for the diagnosis of pulmonary thromboembolism: a preclinical pilot study.

BACKGROUND: Recommendations for the diagnosis of pulmonary embolism are available for healthcare providers. Yet, real practice data show existing gaps in the translation of evidence-based recommendations. This is a study to assess the effect of a computerised decision support system (CDSS) with an enhanced design based on best practices in content and reasoning representation for the diagnosis of pulmonary embolism. DESIGN: Randomised preclinical pilot study of paper-based clinical scenarios in the diagnosis of pulmonary embolism. Participants were clinicians (n=30) from three levels of experience: medical students, residents and physicians. Participants were randomised to two interventions for the diagnosis of pulmonary embolism: a didactic lecture versus a decision tree via a CDSS. The primary outcome of diagnostic pathway concordance (derived as a ratio of the number of correct diagnostic decision steps divided by the ideal number of diagnostic decision steps in diagnostic algorithms) was measured at baseline (five clinical scenarios) and after either intervention for a total of 10 clinical scenarios. RESULTS: The mean of diagnostic pathway concordance improved in both study groups: baseline mean=0.73, post mean for the CDSS group=0.90 (p<0.001, 95% CI 0.10-0.24); baseline mean=0.71, post mean for didactic lecture group=0.85 (p<0.001, 95% CI 0.07-0.2). There was no statistically significant difference between the two study groups or between the three levels of participants. INTERPRETATION: A computerised decision support system designed for both content and reasoning visualisation can improve clinicians' diagnostic decision-making.

Human-centred design processes for clinical decision support: A pulmonary embolism case study.

BACKGROUND: Clinical Decision Support Systems (CDSS) can make patient care more efficient, cost-effective, and guideline-concordant. Many are created by clinicians who understand the challenges, but may publish concepts before considering subtle but important design details. Human-Centred Design (HCD) approaches provide necessary methods ensuring solid CDSS design. This article highlights HCD approaches in a pulmonary embolism CDSS case study context. METHODS: This pulmonary embolism CDSS results from collaborative work between computer science, psychology, and medicine. HCD methods used include: evaluations of pre-clinical prototype recordings, iterative usability expert reviews with software refinement, formative usability testing, and (separately-published) clinical pilot study. RESULTS: HCD methods were instrumental in iteratively creating an easy to use and functionally-sound CDSS. Retrospective evaluations revealed that participants spent considerable time on items that were out of order from natural cognitive diagnostic workflows. Features missing between original and study version were noted, confusing interface elements reworked, and currently-active decision tree branches were visually emphasized. From iterative usability reviews, positioning of information within the decision tree was radically reworked, information separated into levels of support for different user groups, and supportive versus directive language issues addressed. Formative studies identified issues such as interface adjustments and hospital workflow integration. CONCLUSIONS: Human-centred design approaches provide methods for integrating the skills and knowledge of many disciplines, illustrated by example in this pulmonary embolism CDSS creation. Advantages of leveraging many design guidelines as well as revealing new design considerations that would otherwise have remained hidden are described. The findings reported here support future CDSS design through HCD inclusion.