Pain Research and Management (Jan 2022)
Validation of a Model Predicting That Physical Activities Improve Health-Related Quality of Life in Older Japanese Adults with Pain, Dysesthesia, and Kinesiophobia after Lumbar Surgery: Structural Equation Modeling
Abstract
Objectives. This study assessed the validity of a hypothesized model predicting that physical activity improves health-related quality of life (HRQOL) in older Japanese adults with pain, dysesthesia, and kinesiophobia following lumbar surgery. Methods. We included 431 elderly patients who underwent surgery for lumbar spinal stenosis at two hospitals. The frequency of physical activity, pain, dysesthesia, kinesiophobia (somatic focus and activity avoidance), and HRQOL were investigated using a questionnaire. Missing values were complemented by the stochastic regression imputation. We constructed the following model. (i) physical activity affects pain, dysesthesia, and kinesiophobia. (ii) pain, dysesthesia, and kinesiophobia separately affect HRQOL. This hypothetical model was tested by structural equation modeling. The model was improved based on a modified index. Results. Of the 431 respondents, 297 (median age 72 years, range 65–91 years; 158 men and 139 women) were analyzed (68.9%). The fit of the model improved based on the modification index and was acceptable comparative fit index, 0.948; Tucker–Lewis index, 0.919; root mean square error of approximation, 0.048 (90% confidence interval, 0.026–0.069), and standardized root mean square residual (0.046). The paths by which physical activities reduced pain or dysesthesia (standardized pass coefficients, −0.406) and somatic focus (−0.301) and consequently improved HRQOL were significant (pain/dysesthesia, −0.684; somatic focus, −0.218). Discussion. Our hypothesized model predicting that physical activity improves HRQOL in terms of pain, dysesthesia, and kinesiophobia in older Japanese adults after lumbar surgery was validated using cross-sectional data. Interventional studies on physical activity based on this model are required to establish the model.