PRODIAB: Perspectives on the use of patient-reported outcome measures among diabetic patients

Patient Reported Outcome Measures (PROMs) are recognized by health authorities as fundamental and can be evaluated by several questionnaires. The main complexity of evaluating PROMs in People with Diabetes (PwD), be it type 1 (PwT1D) or type 2 (PwT2D), is choosing the right tool (generic or specific) with the right constructs. This study explores the use of PROMs in PwD. The main objective was to compare generic and specific QoL PROMs in a diabetic population. The secondary objective was to assess potential overlaps of assessed constructs in the different PROMs frequently used in diabetes. PRODIAB was an online quantitative survey conducted between January and February 2022. The scientific committee selected the following oft-used questionnaires: EuroQol 5-Dimensions 5-Level (EQ-5D-5L), Audit of Diabetes Dependent Quality of Life (ADDQOL), Treatment Burden Questionnaire (TBQ), Hospital Anxiety and-Depression Scale (HADS-A; HADS-D) and Patient-Health-Questionnaire-9 (PHQ-9). All PROMs used were validated in French. Responses from 2,796 French PwD were analyzed. Compared to PwT2D, PwT1D reported a better general Quality of Life (QoL) on the EQ-5D-5L index (0.75 vs 0.66; p < 0.001) and a lower specific QoL on the ADDQOL (-2.8 vs -2.5; p < 0.001). Adjusted analysis showed that age was inversely associated with EQ-5D-5L index score (-0.11; p < 0.001) and a positively associated with ADDQOL score (0.14; p < 0.001). All PROMs were significantly correlated with each other and the HADS-A, HADS-D and PHQ9 (r 0.60 to 0.72) even more so. While principal component analyses suggested that all PROMs measured only one dimension (mental health) (eigenvalue=3.39; first dimension percentage of variance 56.5), adjusted analyses suggested that the EQ-5D-5L is not adequate for assessing the specific impact of diabetes on QoL. Our study emphasizes the importance of identifying the constructs assessed by each PROM as well as the target population in order to choose the best-suited questionnaire type with the most appropriate constructs.