The association of physical activity and depression in patients with, or at risk of, osteoarthritis is captured equally well by patient reported outcomes (PROs) and accelerometer measurements - Analyses of data from the Osteoarthritis Initiative

摘要

Objectives Osteoarthritis (OA) patients are at increased risk of depression, and low levels of physical activity (PA) are a potential warning sign of depression. PA can be estimated by patient reported outcomes (PROs) or measured with accelerometers (ACCs). We explored which of these two best captures depression in patients with, or at risk of, OA. Methods 48-months data from the Osteoarthritis Initiative were cross-sectionally analysed. The dichotomized Centre for Epidemiological Studies Depression Scale score was used as outcome (depression y/n). The Physical Activity Scale for the Elderly (PASE) was selected as PRO. ACC-data comprised average minutes of daily moderate to vigorous activity. Two multivariable models (PRO-model/ACC-model) were compared directly and indirectly using areas under the curve (AUC) for the predicted probability of depression, penalized model selection criteria (PMSC) and log-likelihood ratio tests. Results AUCs from the ACC (0.71 [95% CI 0.67; 0.75]) and PRO model (0.72 [95% CI 0.68; 0.76]) were not significantly different (p?=?0.28). Differences in PMSC were small (<10). The log-likelihood ratio test for the comparison of the ACC (ll ?505.22) with the base model (ll ?505.98) was not significant (LR chi2 = 1.52; p?=?0.22), but the PRO model (ll ?501.55) had a better fit than the base model (LR chi2 = 8.87; p < 0.01). Conclusions PRO and ACC data perform similarly in capturing depression. Indirect comparison even pleads for PROs. Costs of accelerometers and the additional burden for patients are in support of the PASE as an appropriate alternative to screen for depression in OA patients