Open Access

Rica Hartmann, Florian Avermann, Christoff Zalpour, Annika Griefahn

• Background Low Back Pain (LBP) is a common health problem worldwide. In recent years, the use of mobile applications for the treatment of various diseases has increased, due to the Corona pandemic. Objective The aim of this study is to investigate the extent to which artificial intelligence (AI)-assisted exercise recommendations can reduce pain and pain-related impairments in daily life for patients with LBP, compared to standard care. Methods To answer the research question, an 8-week app-based exercise program was conducted in the intervention group. To measure the influence of the exercise program, pain development and pain-related impairment in daily life have been evaluated. A so-called rehabilitation sports group served as the control group. The main factors for statistical analysis were factor time and group comparison. For statistical calculations, a mixed analysis of variance for pain development was conducted. A separate check for confounders was made. For pain impairment in daily life nonparametric tests with the mean of change between the time points are conducted. Results The intervention group showed a reduction in pain development of 1.4 points compared to an increase of 0.1 points in the control group on the numeric rating scale. There is a significant interaction of time and group for pain development. Regarding pain-related impairments in daily life, the intervention group has a reduction of the oswestry disability index scores by 3.8 points compared to an increase of 2.3 in the control group. The biggest differences become apparent 8 weeks after the start of treatment. The significant results have a medium to strong effect. Conclusion The results shown here suggest that the use of digital AI-based exercise recommendations in patients with LBP leads to pain reduction and a reduction in pain-related impairments in daily living compared to traditional group exercise therapy.