Jaya Tripathi
The MITRE Corporation, USA
Title: Identifying risky drug-seeking behavior at the point-of-care
Biography
Biography: Jaya Tripathi
Abstract
Statement of the Problem: There have been many efforts in the last few years to address the opioid epidemic. Despite the efforts, deaths involving opioid overdoses have continued to rise. One effort has been mandates in 41 states requiring a prescriber to check the state’s PDMP (Prescription Drug Monitoring Program) before prescribing a controlled substance. Prescribers must carefully assess each patient’s risk of medication abuse. This task is challenging, as it can be difficult to distinguish between a person who may be visiting multiple clinicians because of substance abuse disorder and legitimate patients with acute exacerbations of chronic pain. Other challenges faced by physicians include lack of PDMP-EHR integration, and a lack of decision-making tools for pain management. Thus these PDMP datasets remain under-utilized in terms of applying advanced analytic and visualization techniques that would help with decision-making at the point of care.
Methodology & Theoretical Orientation: At the outset of our research, we did a landscape study and assessed that there is considerable subjectivity in pain management approaches; decision-making is limited to physicians’ gestalt, to high-level guidance practices like ‘traffic rules’, to CDC guidelines, etc. We applied advanced analytic and visualization techniques to a large PDMP dataset of over 12m. records. Using the outcomes from our models, we developed the MeDSS tool. We first conducted a usability study with the Board of Pharmacy and improved our tool based on the feedback. We then did a cross-over study whose purpose was to determine how the MeDSS tool improves efficiency and recognition of high-risk factors and reduces subjectivity in opioid-prescribing decision-making. 50 physicians from Brigham and Women’s Hospital in Boston participated in the study, 25 were shown MeDSS and 25 the standard PDMP view.
Conclusion & Significance: The key findings are that the new tool reduced the subjectivity in decision-making. The test group was able to quickly glean Important information such as payment type, harmful polypharmacy, cross-state activity, etc. We also found a difference in the decision to prescribe an opioid for some of the patient profiles.