Comorbid Diseases and Medication Adherence in a MS Population Using EMR-Based Data

Background: MS is the most common cause of neurologic disability in young adults, with a peak age of onset between 20 and 40 years. Data on prevalence and treatment patterns in U.S. populations has typically been based on claims data. Clinic-based data from electronic health records (EHRs) may provide complementary estimates of treatment and disease patterns.

Objectives: To describe medications used and comorbidities for MS patients as documented in the EHR for a large multi-payor health system in order to learn more about current prescribing and patient characteristics.

Methods: Geisinger Health System (GHS) is a multi-payor delivery system offering healthcare services to residents of 31 of 67 counties in Pennsylvania with 1.5 million patient visits per year and an outpatient EHR since 1996. Using the EHR, we identified patients with an MS diagnosis and at least 1 prescription order for an MS drug, and queried all medications and ICD-9 diagnosis codes from inpatient/outpatient encounters from 2004-2010 to describe the distribution of drugs used and comorbidities diagnosed before and after MS treatment.

Results: We identified 1,603 patients (22% male, 98% Caucasian, mean age 44) with a mean observation time of 30 months with an MS diagnosis (range 6-51 months). The most common initially-ordered MS drug was Copaxone (46%), followed by Avonex (24%), Betaseron (18%), Rebif (11%) and Tysabri (2%), with 394 patients (25%) switching drugs during the period. Most commonly prescribed non-MS drugs were antidepressants (65% of patients), corticosteroids (59%), analgesics-narcotic (49%), anticonvulsants (47%) and skeletal muscle relaxants (47%). Most common comorbid diagnoses prior to the first MS diagnosis were: hyperlipidemia (10%), hypertension (9%), arthritis (9%), malignancy (7%), and burning/numbness (5%). Most frequent new-onset diagnoses following an MS diagnosis were hyperlipidemia (13%), arthritis (12%), malignancy (11%), hypertension (10%) and insomnia (4%).

Conclusions: : Clinic-based, multi-payor EHR data is a valuable source of evidence to complement claims databases when characterizing MS patients' treatment and disease characteristics.