Abstract
This study utilized experience sampling methodology (ESM) to examine the relationship of social interactions with daily pain and mood symptoms in people with osteoarthritis (OA) of the knee. Two hundred sixty-eight adults with physician-diagnosed OA of the knee underwent a baseline in-person interview and subsequent week-long ESM protocol to assess their daily activity patterns, pain, and mood via phone interview four times a day. A coding system was developed to assess presence and type of social interactions based on subject self-report of activity patterns. Multilevel modeling was used to examine between- and within-subject variation in outcomes based on both global and momentary measures of social activities, pain, and mood, while controlling for key demographic and potentially confounding variables. Positive associations were demonstrated between the ratio of positive to negative affect and both global (β = 0.49, p < .001) and momentary, especially positive (β = 0.24, p < .05), social activity patterns. Additionally, the association between negative affect and pain (β = −0.07, p < .01) was attenuated in those with more baseline social interactions. Social interaction has the potential to influence mood in adults with OA of the knee, both on a global scale, and through daily variations in interactions. These interactions seem to be directly related to mood, as well as the apparently attenuating the relationship between pain and depression. Daily social interactions showed a robust positive association with contemporaneous positive affect.
Acknowledgments
The authors gratefully acknowledge the assistance of Catherine Polster, Brian Cox, Francis J. Keefe, Jason A. DeCaro, Jessica Greenlee, Shelley Condon, Katy-Lauren Ford, Amber Ingram, Christina Pierpaoli, Jordan Williams, Karolina Zakoscielna, and others who assisted with data collection and management.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1 As noted in the Methods section, the coding of social interactions included both an assessment of social activity immediately prior to the phone call ESM survey (reported here), and an assessment of social activity ‘since the last phone call’. We also ran models using this second variable, and found substantively very similar results; therefore, in the interest of brevity we focus only on the ‘immediately prior’ social interactions variable from this point forward.
2 For negative affect, in addition to the Poisson model, we also tested the same model using negative binomial model and normal distribution assumptions. The coefficients changed only slightly across these approaches, and in each case the interaction between pain and social interactions was significant, suggesting that the result was not an artifact of the particular distribution chosen.