Search

For patients with a chronic disease such as prostate cancer, their possible journeys through treatment can be mapped as a state diagram, which now can be implemented as an electronic health record (EHR) Care Path, generating novel data for analysis and visualization. A prostate cancer Problem List form captured treatment path assignment, treatment response, and recurrence. Patients reported their symptom burden via the Expanded Prostate Cancer Index Composite (EPIC) questionnaire, completed by patients at defined intervals either at home via mobile device or computer, or in clinic on a tablet. Patients move through the Care Path via state transitions triggered automatically via rule. New types of EHR data on each patient's journey–pathway sequence and time-in-state–automatically ensue, enabling novel analyses. In the first 3 months after go-live, 408 patients were being actively managed on the Care Path. Data visualizations display not only each individual patient's journey through the system but also (using R) an aggregated view of the patterns of all patients' journeys. Combining a Care Path modeled as a state diagram with a Problem List form and online questionnaire(s) for patient-reported outcomes proves powerful for collecting chronic disease registry data as a byproduct of patient care–including novel state sequence and state dwell time data. Ready access to such data can accelerate the "Practice-to-Knowledge, Knowledge-to-Practice" cycles crucial to a Learning Health System.

A tool that automatically identifies subject domains of examination questions is useful in at least three ways: (1) it can help learners hone their ability to perform this subject identification task, which is an important skill in several highstakes examinations, (2) in the context of educational content repositories, it can assist both maintainers in organizing and learners in querying content, and (3) it can be used to assess the appropriateness of questions on an examination. In this paper, we present such a tool for engineering examination questions using a novel technique that builds on existing efforts by India’s National Programme on Technology Enhanced Learning (NPTEL). In particular, we rely on textual transcripts as well as links between videos and questions from GATE (the annual Graduate Aptitude Test in Engineering taken by approximately one million candidates) that have been manually generated under this initiative. Our approach also addresses a key weakness in prior work based on queries by keywords, where terms used in multiple subject domains result in erroneous search results.