OBJECTIVES

: To evaluate and compare inappropriate medication use in Parkinson’s Disease (PD) using the TriNetX Electronic Medical Record (EMR) database and the National Ambulatory Medical Care Surveys (NAMCS).

METHODS

: Data from the NAMCS and the TriNetX EMR from 2011-2015 were analyzed. The study sample included elderly patients aged ≥ 65 years with a diagnosis of PD based on the (ICD-9-CM and ICD-10-CM). Inappropriate medications were defined based on the 2015 American Geriatric Society Beers’ criteria and included anti-psychotics and anti-emetics. Descriptive analyses of the NAMCS data involving appropriate weights were conducted using SAS 9.4 whereas analyses of the EMR data were performed using the in-built analytics platform.

RESULTS

: Analyses of the NAMCS data revealed about 6.63 million outpatient visits (95% CI, 5.78 -7.54) for PD during 2011-2015. Inappropriate anti-psychotic and anti-emetic drugs accounted for 0.30% (and 0.25% of the office visits, respectively. Promethazine was the most commonly prescribed inappropriate anti-emetic (0.30%); olanzapine (0.21%), haloperidol (0.20%), risperidone (0.20%) and lurasidone (0.30%) were the most commonly prescribed inappropriate anti-psychotics. The EMR data analyses found 58,833 patients with PD during 2011-2015. Inappropriate antipsychotic and antiemetic drugs were prescribed in 13% and 19% of PD patients, respectively. Among inappropriate anti-emetics, promethazine (8%), metoclopramide (6%) and prochlorperazine (5%) were the most commonly prescribed. The most commonly prescribed inappropriate anti-psychotics were olanzapine (3%), haloperidol (5%) and risperidone (3%). Hypertension as a co-morbidity was recorded in 54% PD patients in TriNetX compared to 13% visits in NAMCS. Similarly, depression (34% vs 8.47%), Type 2 diabetes mellitus (21% vs 5%) and sleep disorders (28% vs 8%) were captured in higher proportions in TriNetX.

CONCLUSIONS

: While the patient-level findings from TriNetX EMR data are not directly comparable to the visit-level estimates from NAMCS, the study findings suggest that EMR data sources can provide strong study samples for outcomes research.