Background/Purpose: Social isolation is an independent risk factor for poor health outcomes. It is unknown if the degree of social isolation differs between RA and OA patients after total hip replacement (THR), or if social isolation is associated with poorer outcomes. This study evaluates the association between social isolation, measured by the Berkman-Syme Social Network Index, and post-operative pain and function in RA and OA patients who have undergone THR.

Methods: Primary and revision RA and OA THR patients enrolled in a large volume, single center arthroplasty registry from May 2007- February 2011 and were alive at follow-up were eligible to participate. RA cases were validated via chart review and were matched 3:1 to OA cases on age (+/- 5 years), sex, procedure type (primary or revision) and year of surgery, to control for both expected differences between patient populations and time since surgery. The Western Ontario and McMaster Universities Arthritis Index (WOMAC) and demographic information were collected pre-operatively and between 2 and 5 years post-operatively. In addition, the Berkman-Syme Social Network Index (BSSNI) was administered between 2 and 5 years post-operatively. The BSSNI was divided into four categories, with the most isolated category being the referent group.

Results: 132/223 (59.2%) RA and 392/561 OA (69.8%) responded.  Demographics were similar in both groups (Table 1).  RA and OA had similar proportions of very socially isolated patients. Overall, the most socially isolated patients had worse postoperative WOMAC pain (p-value=0.013) and WOMAC function (p-value= 0.0025). This relationship was statistically significant for OA patients, (WOMAC pain p-value=0.0029 and WOMAC function p-value= 0.048), and showed similar trends but was not statistically significant in RA patients. In a multivariate regression controlling for disease (OA vs. RA), education, race , and SF-36 Physical Component Score (PCS), being socially isolated was independently associated with poor (WOMAC < 60) post-op pain, (OR 2.9; 95% CI 1.07-7.25, p-value=0.04) but not poor post-op function, (OR 2.1; 95% CI 0.77-5.52; p-value=0.15). This model also showed that RA was statistically significantly associated with poor post-op function (OR 3.2, 95% CI 1.57-6.52, p-value=0.001), but not with poor post-op pain (OR 1.92; 95% CI 0.86-4.28, p-value=0.11).

Conclusion:   Being socially isolated is associated with an almost 3x increased odds of poor pain after THR, controlling for multiple potential confounders.  Furthermore, social isolation appears to be more significant in OA, which comprises the vast majority of THR cases. Further prospective studies should be done to evaluate whether interventions to improve social networks can improve pain after THR.