Predictors of physical activity recording in routine mental healthcare

Physical activity (PA) is beneficial for those with mental illness. However, it is unclear what influences clinicians' decisions to record PA. We conducted the first study to examine PA recording and its predictors in clinical settings, using data from routine mental health records in a secondary mental health service. A retrospective study using anonymised electronic health record data was conducted. Patients with any psychiatric diagnosis who had PA recorded were matched to controls (1:5) who were active in hospital at that time and did not have PA recorded. Multivariable logistic regressions were conducted to examine how sociodemographic factors and health status predict PA recording. 5034 patients (839 with PA recorded and 4195 matched controls) with mental disorders were included in the analysis [mean (SD) age = 44 (17) years; 50.0% males]. Being hospitalized in the past year (OR = 8.9; 95%CI = 6.8-11.54) and having hallucinations (OR = 1.75; 95%CI = 1.37-2.24) were significantly associated with higher odds for PA recording. Mental disorder diagnosis of organic mental disorders, neurotic/stress-related disorders and disorders of psychological development were significantly associated with lower odds of PA recording. These results highlight which patients were more likely to have PA recorded. These findings can be used to advocate for more training and resources in mental health care around PA counselling and recording.