COMORBIDITIES IN MULTIPLE SCLEROSIS- A DESCRIPTIVE LONGITUDINAL CLUSTERING ANALYSIS OF HEALTHCARE RESOURCE UTILIZATION AMONG MS PATIENTS IN IBM EXPLORYS DATABASE
Author(s)
Schneider D1, Alevras D2, Saxena S3, Zhang H4, Wu Y5
1Teva Pharmaceuticals, Frazer, PA, USA, 2IBM Watson Health, Wayne, PA, USA, 3IBM Watson Health, New York, NY, USA, 4IBM Watson Health, Rochester, MN, USA, 5Teva Branded Pharmaceuticals Products R&D Inc, Frazer, PA, USA
OBJECTIVES: The comorbidities of multiple sclerosis (MS) patients are well described in the literature, with depression and hypertension being most prevalent. This research aims to use clustering analysis to characterize the timing of healthcare resource utilization for patients based on comorbidity burden (CB) and timing of disability progression (DP). METHODS: Retrospective analysis of EHRs of MS patients in the IBM Explorys database (2000-2017), a longitudinal, de-identified clinical database for >50 million unique patients in the US, was conducted using iterative, knowledge-driven clustering of patient-level data. Clustering features with predefined scores were based on comorbidities, MS-related events, follow up, healthcare interactions (physician encounters, medications, and hospitalizations), and physician reported observations. Clusters of highest and lowest tertiles were selected to map clinical journeys of patients. Twenty-six events were aggregated by order of occurrence along a given patient journey. Median time-to-event and median age were reported for each event per occurrence. RESULTS: A total of 27,744 MS patients were analyzed. Four clusters were defined as A: Early Disability/Low Comorbidity (n=876), B: Early Disability/High Comorbidity (n=5237), C: Late Disability/Low Comorbidity (n=5308), D: Late Disability/High Comorbidity (n=751). Mean # of relapses overall respectively (0.5, 3.0, 0.2, 1.3); Std. Dev (1.5, 7.0, 0.8, 3.0); median follow-up years (4.49, 5.84, 5.06, 6.98); % female (73, 81, 76, 79); % African-American (16, 20, 14, 13); number of comorbidities with >10% prevalence (17, 133, 6, 75); median encounters 2-5 years since diagnosis: MS-related (3, 9, 2,4), all encounters (5, 28, 4,19); median comorbidities per year between ages 46-55 (2, 5, 1, 3). CONCLUSIONS: In this clustering analysis we found that MS patients with high CB and more rapid MS DP utilize healthcare resources to a much greater extent than other MS patients. Increased multidisciplinary management of medically complex MS patients may be warranted to reduce healthcare utilization.
Conference/Value in Health Info
2017-11, ISPOR Europe 2017, Glasgow, Scotland
Value in Health, Vol. 20, No. 9 (October 2017)
Code
PND1
Topic
Epidemiology & Public Health
Disease
Cardiovascular Disorders, Mental Health, Multiple Diseases, Neurological Disorders