Who Gets Referred? Sociodemographic Gaps in Access to Subspecialty Care for Advanced Parkinson’s Disease

The objective of this study was to characterize sociodemographic differences in subspecialty referrals for advanced Parkinson’s disease (aPD). Only 9% of patients with aPD receive subspecialty care despite evidence that such care improves outcomes, with care further limited for minoritized groups. Retrospective cohort study of patients with aPD, (ICD-10 diagnoses of PD with fluctuations and/or dyskinesia (G20.A2, B1 and B2)), in a large integrated health network, between 2020 and 2024. Referrals to the affiliated academic subspecialty clinic were defined as system-entered referrals to a movement specialist. Referral completion (referred patients with a completed encounter with a movement specialist) was tracked. Demographic and clinical characteristics were compared using t-tests and chi-square tests. Among 3,037 patients with aPD (mean age 74.1±9.1 years; 62.7% male), 690 (22.7%) were referred to subspecialists, of whom 533 (77.2%) completed at least one visit. Patients referred were younger (72.0 vs. 74.7 years, p<0.001), partnered (69.7% vs. 64.8%, p=0.044), and more likely to reside in urban counties (56.7% vs. 33.4%, p<0.001). Referred patients were predominantly from outside the health system (81.9%, p<0.001). Referral completion was more common among older (72.2 vs. 71.3 years, p=0.025), male (64.4% vs. 51.0%, p=0.002), partnered (71.5% vs. 63.1%, p=0.032), and externally referred patients (84.0% vs. 74.5%, p=0.006) compared to those who were referred but did not complete a visit. Race/ethnicity distributions did not differ significantly. In one of the largest integrated health systems in the region, only one in five patients with aPD were referred to an academic movement disorders center. Limits in access to subspecialty care showed differences at social and geographic levels. Findings suggest that referrals to specialists may be an important intervention node and highlight the need to measure social vulnerabilities as multidimensional risks rather than focusing on single dimensions such as race/ethnicity.