The goal of the project is to create a framework and guidelines for designing advanced interactive information visualizations that provide patient-centered cognitive support by enhancing the accessibility and understanding of patient data, thus addressing key clinical tasks that are currently the source of substantial errors and inefficiency.
Please see the The University of Maryland's SHARPC page on Reconciliation for full details.
Two prototypes were completed in the 1st first year that presented two completely different user interfaces and ways for clinicians to address medication reconciliation (i.e. reconciling two lists of medications into a single reconciled list) in two different use-case scenarios. These interfaces (Twinlist and MEDREC) were built on the substratum of a novel medication reconciliation algorithm that removes the tediousness of a fully manual reconciliation without diminishing the decision making power of the clinician.
Our medication reconciliation algorithm and the user interfaces are part of the Pan-Sharp project now run on the SMART platform currently under active development by Harvard’s SHARP project.
Please see the The University of Maryland's SHARPC page on Results Management for full details.
A prototype had been developed in the early part of the project to illustrate designs that facilitate the management of orders to reduce the problem of “missed labs”.
Our design makes use of a hierarchy of process definitions, which when combined with a database of actors and organizations, provides input for an Interface Generator. The software architecture produces a domain independent system that can be widely used and easily modified. A set of principles ensures that lab results are returned and acted on: (1) definition of agent temporal responsibilities, (2) generation of actor action sheets that offer appropriate choices at each step of the process, integrated in the result viewing screens, and (3) use of predictions of estimated time of completion for each step to increase awareness of delays and better prioritization of actions.
Our visual approach to retrospective analysis uses visualization to review the performance of different actors involved in multistep processes.
Please see the The University of Maryland's SHARPC page on Wrong Patient Selection for full details.
Wrong patient selection is a severe problem in patient safety; from ordering medication to performing surgery. CPOE systems can sometimes even increase the rate of wrong patient selection. We reviewed the origin of wrong patient errors, and suggested user interface remedies; we built a prototype system to demonstrate them.
Please see the The University of Maryland's SHARPC page on Treatment Explorer for full details.
We designed a new visualization to present summaries of outcome data to patients (or to the patient-provider team). We reviewed the literature and propose a framework for design (see paper below) and are now developing an interactive web-based prototype showing 4 alternative designs. The most promising prototype uses step by step animation to explain the new visualization. We are also preparing a user study to compare a text only design with the new designs.
In this very exploratory project we applied network analysis and visualization tools to study a corpus of patient discharge summaries, exploring the relationships between patients and their associated symptoms, diseases, drugs, and procedures. Cody Dunne used the technology he had been developing in his PhD thesis (including motif simplification and group-in-a-box layouts) to reduce clutter as well as interactive text displays to show the origin of each relationship.
Our research on the Systematic Yet Flexible Systems Analysis (SYFSA) framework guides interface design even as it is continuously refined by our practical experience in applying it. We expect that SYFSA will enable the creation of Healthcare IT systems that encourage best practices while simultaneously accommodating the tremendous variety of real-world healthcare workflow. Systematic Yet Flexible Systems can provide visual and other cues to encourage evidence-based best practices along with visual feedback of progress while simultaneously supporting the need to deviate from these standards in some cases.
A Mathematica notebook containing code to automate SYFSA analysis is available on request from Todd R Johnson.
The Pan-SHARP project is a collaboration between all SHARP projects, demonstrating interoperability in a SmartApp.
The medication reconciliation project was selected as the focus and showcase of a Pan-SHARP collaboration. The Pan-SHARP collaboration, under Project 4’s lead, incorporates innovations from all four SHARP projects and MD PnP to deliver an innovative medication reconciliation solution with state-of-the-art technology. Project 4 members are leading this project.