Validation of PROMIS Anxiety Item Bank Computer Adaptive Test (CAT) Among Patients with Heart Failure

Nolan Marblestone*^1*Steven Chu*²Nicole Tomei¹Denzel Lodge¹Aarushi Bansal³Nathaniel Edwards¹Heather J Ross⁴Josef Stehlik⁵Desana Thayaparan²Jad Fadlallah¹Joshua G Lee⁶Istvan Mucsi^7*

• ¹Ajmera Transplant Centre & Division of Nephrology, University Health Network (UHN), Toronto, Ontario, Canada
• ²Ted Rogers Centre for Heart Research, University Health Network (UHN), Toronto, Ontario, Canada
• ³Department of Medicine, Division of Cardiology, McMaster University, Hamilton, Ontario, Canada
• ⁴Peter Munk Cardiac Centre, University Health Network (UHN), Toronto, Ontario, Canada
• ⁵Division of Cardiovascular Medicine, The University of Utah, Salt Lake City, Utah, United States
• ⁶Department of Medical Sciences, Western University, London, Ontario, Canada
• ⁷Ajmera Transplant Centre & Divison of Nephrology, University Health Network (UHN), Toronto, Ontario, Canada

Introduction: Anxiety is highly prevalent among patients with heart failure (HF), negatively affecting health related quality of life (HRQOL). The Patient-Reported Outcomes Measurement Information System (PROMIS) anxiety item bank computer adaptive testing (CAT) precisely assesses anxiety symptom severity. This study aims to assess construct validity and reliability of PROMIS-Anxiety CAT among patients hospitalized for HF. Methods: A cross-sectional convenience sample of adult patients hospitalized for HF, who completed PROMIS-A CAT, generalized anxiety disorder 7 (GAD-7), and other questionnaires electronically. Convergent validity was assessed by Spearman's rank correlation between PROMIS-A CAT, GAD-7, and other legacy measures. Known group analysis compared PROMIS-A CAT and GAD-7 scores between groups expected to have different levels of anxiety. Reliability of PROMIS-A CAT was calculated on the individual and group level from standard error of measurement, according to item response theory. Area under receiver-operating characteristics (ROC) curve and Youden's J statistic were used to identify a T-score cut-off for moderate/severe anxiety. Results: Of 333 participants, 87 (26%) had moderate/severe anxiety based on GAD-7 score (≥ 10). Participants completed on average (median [IQR]) 4(1) versus 7(0) items, with PROMIS-A CAT and GAD-7, respectively. PROMIS-A CAT T-scores were strongly correlated with GAD-7 scores (rho = 0.78) and moderately correlated with other legacy measures. Known-group analysis provided further support for construct validity of PROMIS-A CAT. Individual reliability for PROMIS-A CAT T-scores was >0.9 for 87% of the sample; mean reliability was 0.91. Based on ROC and Youden's J analyses, a T-score of 60 can be used to identify individuals with moderate/severe anxiety. Conclusion: These results support the validity and reliability of PROMIS-A CAT among patients hospitalized for HF.