BackgroundOsteoarthritis (OA) has a significant impact on individuals' ability to work. Our goal was to investigate the effects of the site of OA (knee, hip, hand, foot, lower back or neck) on employment reduction due to OA (EROA).Methods and FindingsThis study involved a random sample of 6,000 patients with OA selected from the Medical Service Plan database in British Columbia, Canada. A total of 5,491 were alive and had valid addresses, and of these, 2,259 responded (response rate = 41%), from which 2,134 provided usable data. Eligible participants were 19 or older with physician diagnosed OA based on administrative data between 1992 and 2006. Data of 688 residents were used (mean age 62.1 years (27 to 86); 60% women). EROA had three levels: no reduction; reduced hours; and total cessation due to OA. The (log) odds of EROA was regressed on OA sites, adjusting for age, sex, education and comorbidity. Odds ratios (ORs) represented the effect predicting total cessation and reduced hours/total cessation. The strongest effect was found in lower back OA, with OR = 2.08 (95% CI: 1.47, 2.94), followed by neck (OR = 1.59; 95% CI: 1.11, 2.27) and knee (OR = 1.43; 95% CI: 1.02, 2.01). We found an interaction between sex and foot OA (men: OR = 1.94; 95% CI: 1.05, 3.59; women: OR = 0.89; 95% CI = 0.57, 1.39). No significant effect was found for hip OA (OR = 1.33) or hand OA (OR = 1.11). Limitations of this study included a modest response rate, the lack of an OA negative group, the use of administrative databases to identify eligible participants, and the use of patient self-reported data.ConclusionsAfter adjusting for socio-demographic variables, comorbidity, and other OA disease sites, we find that OA of the lower back, neck and knee are significant predictors for EROA. Foot OA is only significantly associated with EROA in males. For multi-site combinations, ORs are multiplicative. These findings may be used to guide resource allocation for future development/improvement of vocational rehabilitation programs for site-specific OA.