User-Centered Collaborative Design and Development of an Inpatient Safety Dashboard.

Patient safety remains a key concern in hospital care. This article summarizes the iterative participatory development, features, functions, and preliminary evaluation of a patient safety dashboard for interdisciplinary rounding teams on inpatient medical services. This electronic health record (EHR)-embedded dashboard collects real-time data covering 13 safety domains through web services and applies logic to generate stratified alerts with an interactive check-box function. The technological infrastructure is adaptable to other EHR environments. Surveyed users perceived the tool as highly usable and useful. Integration of the dashboard into clinical care is intended to promote communication about patient safety and facilitate identification and management of safety concerns.

A system dynamics model of clinical decision thresholds for the detection of developmental-behavioral disorders.

BACKGROUND: Clinical decision-making has been conceptualized as a sequence of two separate processes: assessment of patients' functioning and application of a decision threshold to determine whether the evidence is sufficient to justify a given decision. A range of factors, including use of evidence-based screening instruments, has the potential to influence either or both processes. However, implementation studies seldom specify or assess the mechanism by which screening is hypothesized to influence clinical decision-making, thus limiting their ability to address unexpected findings regarding clinicians' behavior. Building on prior theory and empirical evidence, we created a system dynamics (SD) model of how physicians' clinical decisions are influenced by their assessments of patients and by factors that may influence decision thresholds, such as knowledge of past patient outcomes. Using developmental-behavioral disorders as a case example, we then explore how referral decisions may be influenced by changes in context. Specifically, we compare predictions from the SD model to published implementation trials of evidence-based screening to understand physicians' management of positive screening results and changes in referral rates. We also conduct virtual experiments regarding the influence of a variety of interventions that may influence physicians' thresholds, including improved access to co-located mental health care and improved feedback systems regarding patient outcomes. RESULTS: Results of the SD model were consistent with recent implementation trials. For example, the SD model suggests that if screening improves physicians' accuracy of assessment without also influencing decision thresholds, then a significant proportion of children with positive screens will not be referred and the effect of screening implementation on referral rates will be modest-results that are consistent with a large proportion of published screening trials. Consistent with prior theory, virtual experiments suggest that physicians' decision thresholds can be influenced and detection of disabilities improved by increasing access to referral sources and enhancing feedback regarding false negative cases. CONCLUSIONS: The SD model of clinical decision-making offers a theoretically based framework to improve understanding of physicians' behavior and the results of screening implementation trials. The SD model is also useful for initial testing of hypothesized strategies to increase detection of under-identified medical conditions.