MedicalResearch.com Interview with:
Daniel R. Murphy, M.D., M.B.A.
Assistant Professor – Interim Director of GIM at Baylor Clinic
Department of Medicine
Health Svc Research & General Internal Medicine
Baylor College of Medicine
MedicalResearch.com: What is the background for this study? What are the main findings?
Dr. Murphy: Electronic health records (EHRs) have improved communication in health care, but they have not eliminated the problem of patients failing to receive appropriate and timely follow up after abnormal test results. For example, after a chest x-ray result where a radiologist identifies a potentially cancerous mass and suggests additional evaluation, about 8% of patients do not receive follow-up imaging or have a visit with an appropriate specialist within 30 days. Identifying patients experiencing a delay with traditional methods, like randomly reviewing charts, is not practical. Fortunately, EHRs collect large amounts of data each day that can be useful in automating the process of identifying such patients.
We evaluated whether an electronic “trigger” algorithm designed to detect delays in follow up of abnormal lung imaging tests could help medical facilities identify patients likely to have experienced a delay. Of 40,218 imaging tests performed, the trigger found 655 with a possible delay. Reviewing a subset of these records showed that 61% were truly delays in care that required action. We also found that the trigger had a sensitivity of 99%, indicating that it missed very few actual delays.
MedicalResearch.com: What should readers take away from your report?
Dr. Murphy: The electronic trigger achieved an accuracy sufficient for future practical use, even in large clinical data repositories. Large-scale application of triggers could leverage economies of scale by allowing multiple sites to use a centralized team to monitor and act on delays.
MedicalResearch.com: What recommendations do you have for future research as a result of this study?
Dr. Murphy: While our triggers were successful in identifying delays, more work is needed to determine the best way to use this information to impact patient care. For example, it is still unclear who is the best recipient for this information: physicians who are already overloaded with EHR-delivered information, the clinical team who can manage the administrative tasks of ensuring follow up progresses as expected and involved physicians when necessary, or facility-level patient safety personnel.
MedicalResearch.com: Is there anything else you would like to add?
Dr. Murphy: A large amount of clinical information and justification for physicians to take no or alternative action is contained within free text notes, which were not accessible to our trigger. Future triggers can potentially take advantage of new text mining techniques, like natural language processing, which that can interpret physician’s notes and make use of this information when deciding to flag a record as a delay or not. This would further improve the efficiencies gained with use of the triggers.
Chest. 2016 May 10. pii: S0012-3692(16)48968-3. doi: 10.1016/j.chest.2016.05.001. [Epub ahead of print]
Computerized Triggers of Big Data to Detect Delays in Follow-up of Chest Imaging Results.
Murphy DR1, Meyer AN2, Bhise V2, Russo E2, Sittig DF3, Wei L2, Wu L4, Singh H2.
Note: Content is Not intended as medical advice. Please consult your health care provider regarding your specific medical condition and questions.
More Medical Research Interviews on MedicalResearch.com.