Modeling the Effects of Different Interventions on Kinesiophobia, Pain, and Quality of Life in Women with Lumbar Disc Herniation Using Artificial Neural Networks

<title>Abstract</title> <p>This study aimed to examine the effects of swimming exercise, physical therapy, and a inactive lifestyle on pain, kinesiophobia, and quality of life in women diagnosed with lumbar disc herniation and to model the SF-12 physical health score using kinesiophobia and pain levels through Artificial Neural Networks (ANNs). A total of 54 female volunteers participated in the study. The participants had a mean age of 52.83 ± 13.30 years, a mean height of 162.18 ± 4.45 cm, a mean body weight of 75.37 ± 13.53 kg, a mean body mass index of 28.67 ± 5.11 kg/m², and a mean disease duration of 8.98 ± 7.85 years. Data were collected using a personal information form, the Visual Analogue Scale, the Kinesiophobia Scale, and the SF-12 Quality of Life Questionnaire. Normality was assessed using the Shapiro–Wilk test, group differences were evaluated by MANOVA, and the statistical significance level was set at p &lt; 0.05. In the ANN models, the total kinesiophobia score and pain score were used as input variables, while the SF-12 physical health score was used as the output variable. A separate model was developed for each pairwise group comparison. A significant difference was found among the groups in terms of pain level (p &lt; 0.01), with the swimming group exhibiting lower pain scores than the physical therapy and inactive groups. No significant differences were observed in kinesiophobia or physical quality-of-life scores. In the Swimming–Physical Therapy model, MSE was 0.675, MAPE was 3.242%, and R was 0.991. In the Physical Therapy–Inactive model, MSE was 0.705, MAPE was 3.410%, and R was 0.991. In the Swimming–Inactive model, MSE was 1.429, MAPE was 5.363%, and R was 0.973. The results indicate that swimming exercise may support pain management and that ANN models can predict physical quality-of-life scores with a high level of agreement.</p>