Description |
Electronic health records (EHRs) have been widely implemented in hospitals and outpatient settings and hold many promises for improved health outcomes. However, physicians have complained that the EHR has negative and unintended consequences, including interference with their relationship with the patient, high cognitive burden, low usability, and poor support for the medical decision-making process. In this dissertation, 4 studies were conducted to explore system features contributing to physician dissatisfaction and the EHR's impact on cognitive tasks. A problem-focused data display, designed to support clinical decision-making, was developed and evaluated. The first study addresses physicians' ability to manage the complex cognitive task of communicating with their patients during clinical exams, while accessing an EHR compared with use of a similar paper record. Communication skills were observed and evaluated in specific areas of clinical importance. Analysis shows that the simple transition from paper to computer does not appear to reduce physician's ability to communicate with their patients. Rather, use of the computer trended toward slight improvement of communication skills. Second, an ineffective active clinical decision support (CDS) tool-set was evaluated. A heuristic evaluation was conducted followed by a user study to identify problems that might be corrected. These problems were mapped to dual cognitive processes and a set of CDS-specific design guidelines was developed. In the third study, a complex medical condition, delirium, was identified as an important target for CDS development and critical incident interviews were conducted to discover information needs of physicians caring for patients with delirium. Uncertainty and high effort in obtaining needed information were among the important themes identified. Fourth, information from decision theory and prior studies was leveraged in developing a CDS tool to support rapid awareness and analytical investigation for physicians managing the care of patients at high risk of developing delirium. The tool received positive usefulness ratings from physicians and, in a randomized controlled trial, resulted in ordering of more delirium prevention strategies and an increase in confidence in delirium risk management. This work expands methods of CDS design and evaluation to support physician cognition by grouping and summarizing problem-focused patient data. |