Background: Depression and anxiety are frequent among kidney transplant recipients (KTRs). Patient reported outcome measures are used to assess these constructs, but concerns remain about measurement precision and questionnaire burden. The NIH PROsetta Stone project developed a common metric system which associates scores from scales measuring similar concepts across known legacy tools with the Patient Reported Outcomes Measurement System (PROMIS-57). Here, we evaluate the accuracy of the depression and anxiety domains of the PROMIS-57 profile questionnaire among KTRs.
Methods: Participants of this cross-sectional, convenience sample of stable KTRs completed the PROMIS-57 (includes PROMIS-29), Generalized Anxiety Disorder Scale (GAD-7) and Patient Health Questionnaire (PHQ-9) questionnaires. Raw scores of legacy tools (GAD-7 and PHQ-9) were converted to PROMIS T-scores using PROsetta Stone© crosswalk files. Spearman correlations were conducted between calculated PROMIS (from legacy scores), and reported PROMIS-57 scores. The cut off score of 10 on GAD-7 and PHQ-9 legacy scales were used to indicate clinically significant (moderate to severe) depression or anxiety, respectively. The corresponding PROsetta stone cut off scores on the reported PROMIS-57 and -29 scales were used to categorize severe to moderate depression and anxiety. We computed the sensitivity, specificity, positive predictive and negative predictive values for the categorization. Lastly, we calculated Cohens Kappa values to assess the degree of agreement between legacy instruments and respective PROMIS-57 and-29 domains to assign patients to “depression” and “anxiety” categories.
Results: Our sample included 150 KTRs (mean (±SD) age was 50 (±17) years, 57% male, 57% white. Based on legacy instruments, 7% had moderate to severe anxiety and 8% had depression while reported PROMIS-57 scores yielded 9% with anxiety and 11% with depression. Calculated anxiety scores showed strong correlations with reported PROMIS-57 (r=0.677, p<0.001), and PROMIS-29 (r=0.760, p<0.001) anxiety scores. Calculated depression scores showed strong correlations with reported PROMIS-57 (r=0.760, p<0.001), and PROMIS-29 (r=0.68, p<0.001) depression scores. The legacy cut offs used for severe to moderate anxiety and depression had high specificity (anxiety=0.95, depression=0.93) and moderate sensitivity (anxiety=0.70, depression=0.58). The Kappa values indicated moderate agreement between GAD-7 categorization of anxiety versus PROMIS-57 (K=0.55) and PROMIS-29 (K=0.56). Similarly, there was moderate agreement between PHQ-9 classification of depression versus PROMIS-57 (K=0.45) and PROMIS-29 (K=0.52).
Conclusions:  The PROMIS-57 and -29 depression and anxiety domains are valid self-report tools that can be used to assess depressive and anxiety symptoms. Furthermore, the shorter questionnaire seems to be a good alternative to reduce questionnaire burden.