Prevalence of Parkinson's disease across North America

Abstract

Estimates of the prevalence of Parkinson's disease in North America have varied widely and many estimates are based on small numbers of cases and from small regional subpopulations. We sought to estimate the prevalence of Parkinson's disease in North America by combining data from a multi-study sampling strategy in diverse geographic regions and/or data sources. Five separate cohort studies in California (2), Minnesota (1), Hawaii USA (1), and Ontario, Canada (1) estimated the prevalence of PD from health-care records (3), active ascertainment through facilities, large group, and neurology practices (1), and longitudinal follow-up of a population cohort (1). US Medicare program data provided complementary estimates for the corresponding regions. Using our age- and sex-specific meta-estimates from California, Minnesota, and Ontario and the US population structure from 2010, we estimate the overall prevalence of PD among those aged ≥45 years to be 572 per 100,000 (95% confidence interval 537-614) that there were 680,000 individuals in the US aged ≥45 years with PD in 2010 and that that number will rise to approximately 930,000 in 2020 and 1,238,000 in 2030 based on the US Census Bureau population projections. Regional variations in prevalence were also observed in both the project results and the Medicare-based calculations with which they were compared. The estimates generated by the Hawaiian study were lower across age categories. These estimates can guide health-care planning but should be considered minimum estimates. Some heterogeneity exists that remains to be understood.

Fig. 1

Prevalence of PD aged ≥45 years by age group and sex: a Females. b Males. HAAS Honolulu-Asia Aging Study, REP Rochester Epidemiology Project, CPDR-PP California Parkinson’s disease Registry-Pilot Project, KPNC Kaiser Permanente Northern California Integrated Health Care System. In each row, the squares or diamonds are centered on the point estimate of the prevalence and whiskers represent 95% confidence intervals. Point estimates for HAAS are indicated by circles instead of squares because the meta-estimates exclude the HAAS study due to methodologic differences between this and the other studies

Fig. 2

Prevalence estimates by study compared to Medicare data from the corresponding counties. The corresponding counties were as follows: For California PD Registry Pilot Project: Fresno, Kern, Santa Clara, and Tulare, for Rochester Epidemiology Project: Olmsted, for KPNC: Alameda, Amador, Contra Costa, El Dorado, Fresno, Kings, Madera, Marin, Mariposa, Merced, Napa, Placer, Sacramento, San Francisco, San Joaquin, San Mateo, Santa Clara, Santa Cruz, Solano, Sonoma, Stanislaus, Sutter, Tulare, Yolo, and Yuba counties, for HAAS: Honolulu