Tracking and predicting the treatment adherence of patients under rehabilitation: a three-wave longitudinal validation study for the Rehabilitation Adherence Inventory

This study aimed to develop and validate a new measurement tool, the Rehabilitation Adherence Inventory (RAI), to measure patients’ rehabilitation adherence. We recruited 236 patients with anterior cruciate ligament (ACL) ruptures from the United Kingdom (Mage = 33.58 ± 10.03, range = 18 to 59; female = 46.2%). Participants completed a survey, that measured their rehabilitation adherence, rehabilitation volume, psychological needs support, autonomous motivation, and intention at baseline, and at the 2nd and 4th month. Factorial, convergent, discriminant, concurrent, predictive, ecological validity and test–retest reliability of the RAI were tested via exploratory factor analysis (EFA), confirmatory factor analysis (CFA), and structural equation modelling (SEM). All the EFAs, CFAs, and SEMs yielded acceptable to excellent goodness-of-fit, χ2 = 10.51 to 224.12, df = 9 to 161, CFI > 0.95, TLI > 0.95, RMSEA <0.09 [90%C I < 0.06 to 0.12], SRMR <0.04. Results fully supported the RAI’s factorial, convergent, discriminant, and ecological validity, and test–retest reliability. The concurrent and predictive validity of the RAI was only partially supported because the RAI scores at baseline was positively associated with rehabilitation frequency at all time points (r = 0.34 to 0.38, p < 0.001), but its corresponding associations with rehabilitation duration were not statistically significant (p = 0.07 to 0.93). Overall, our findings suggest that this six-item RAI is a reliable and valid tool for evaluating patients’ rehabilitation adherence.