Drug Seeking Behavior (DSB) Detection
Customer
State-wide Health Information Network (POC)
Challenges
-
Identifying patients that are seeking opiate drugs are
difficult to identify quantify as their visit are spread
across many providers.
-
Physicians are required to make diagnosis decisions without
support from longitudinal data.
-
Statewide Health Information Networks already collect the
data needed via Clinical Care Documents (CCDs) and
Admit/Discharge/Transfer (ADT) data
-
Patient identifiers are both confidential (PHI) and lack
standards for easy matching
Solutions
-
Patient matching systems were already in place
-
Create a Machine Learning model to identify the same patient
with varying names
-
Create a Machine Learning Model to identify abnormal
frequency of visits by those patient
-
Create a report for review for Emergency Department
Physician Leadership of potential DSB patients that visited
their facilities in the last 30 days.
-
Create an API for Emergency Department Information Systems
(EDIS) to query upon patient arrival potential
Business Impact
-
M/L model provided a 2nd opinion of patient matching
-
Better standardization of CCD documents across provider
systems would increase detection
-
Real-time Notification is needed to support the Emergency
Department Physicians