Psychometric evaluation of the Theoretical Framework of Acceptability (TFA) among adolescent girls and young women in Uganda

The Theoretical Framework of Acceptability (TFA) provides a comprehensive lens for assessing acceptability of healthcare interventions; however, limited literature exists on its psychometric properties in low-resource settings like Uganda. We assessed construct validity and internal consistency reliability of HIV self-testing (HIVST) acceptability scores measured using the TFA among adolescent girls and young women (AGYW) in Uganda. We enrolled 377 AGYW aged 15–24 years in a cross-sectional study and evaluated the psychometric properties of the TFA-informed questionnaire adapted from the generic questionnaire developed by Sekhon et al. (2022). Acceptability was measured using the seven domains of the TFA – each assessed with a single 5-level Likert item question. The adapted questionnaire underwent expert review, translation into Luganda and pilot testing before psychometric evaluation. Construct validation of TFA was examined using the exploratory factor analysis (EFA), confirmatory factor analysis (CFA) and convergent and divergent validity, while reliability was assessed with the Cronbach’s alpha and McDonald’s Omega. Factor extraction was guided by the scree plot and Eigen values (> 1), and factor rotation was performed using oblique Oblimin method. Convergent and divergent validity were assessed by correlating the TFA acceptability scores with willingness to use HIVST and perceived HIV risk respectively. The median age of participants was 20 years (IQR: 18, 22). The questionnaire retained all the seven TFA domains with minor modifications to improve clarity. Although three factors were suggested by scree plot and Eigen values, the rotated solution of the EFA yielded a single factor. The polychoic correction-based CFA had strong factor loadings across all seven TFA domains ranging from 0.68 to 0.87, however, the global fit indices were poor (RMSEA = 0.189, CFI = 0.897, TLI = 0.845). The correlation between acceptability scores and willingness to use HIVST, and perceived HIV risk was 0.7 and 0.1 respectively. The overall Cronbach’s alpha and McDonald’s omega for the questionnaire items were 0.889 and 0.892 respectively. The TFA-informed acceptability scale demonstrated strong reliability and supportive construct validity among AGYW in Uganda. All seven TFA domains demonstrated strong loadings on a single latent factor, indicating that the TFA-informed questionnaire items function coherently as indicators of acceptability, although a single factor structure should be interpreted cautiously. Not applicable.