Better criteria, validation needed for patient-reported outcome measures in CV research
An analysis of 50 patient-reported outcome measures used for CV outcomes trials shows most do not have validated psychometric properties recommended by the FDA and lack established minimally important differences, researchers reported.
“Despite recognition of the stated importance of patient-reported outcome measures in clinical and research settings, they remain underused in CVD trials,” Harriette Van Spall, MD, MPH, associate professor in the division of cardiology at McMaster University and a scientist with the Population Health Research Institute in Ontario, Canada, and colleagues wrote in Annals of Internal Medicine. “Both clinicians and researchers may lack familiarity with most generic and disease-specific patient-reported outcome measures available for use in CVD. To date, there has not been a comprehensive review of instruments, their psychometric properties and the process of validation. Although the U.S. FDA has endorsed a pathway for drug and device approval based solely on improved health states, concordance of these patient-reported outcome measures to FDA recommendations for development and validation is unclear.”
Generic vs. disease-specific measures
Van Spall and colleagues analyzed data from 83 studies that recruited participants with CVD and described the development and validation process of instruments assessing health status and/or health-related quality of life. Researchers stratified 50 health status patient-reported outcome measures (PROMs) by CVD type. Of the 50 PROMs included, five were generic and 45 were disease-specific.
Of the generic PROMs, three were validated for participants with CAD and three were validated for participants with HF. Only the Short Form-36 Health Survey reported interpretability data and the method of establishing minimally important difference.
The remaining 45 disease-specific PROMs were developed for patients with a range of CV conditions, including arrhythmia (22%); CAD (18%); HF (36%); general CVD (13%); and other conditions (11%), including implanted ventricular assist devices or congenital heart disease. Only 20% of disease-specific PROMs had available minimally important difference data; most (70%) of the disease-specific PROMs with minimally important difference data used the distribution-based approach.
Researchers found that no generic PROMs and just eight disease-specific PROMs adhered to all FDA framework recommendations for the validation of psychometric properties.
“Generic PROMs did not stringently assess content validity using the FDA-recommended process, which prefers solicited input from patients with CVD (via focus groups or 1-on-1 interviews) to validate and confirm the relevance of PROM items to the patient population under study,” the researchers wrote.
Of the 45 disease-specific PROMs, only six reported cross-cultural validation or measurement variance, such as differences across subgroups by age, sex or language. The researchers also noted that 13% of disease-specific PROMs were supported by moderate- or high-quality evidence consistently across the remaining properties of content validity, structural validity, internal consistency, reliability, measurement error, criterion validity, construct validity and responsiveness.
Assessing psychometric properties
Researchers also applied the Consensus-based Standards for the Selection of Health Measurement Instruments (COSMIN) criteria to PROMs to assess whether psychometric properties were optimally measured. Of the 50 PROMs, only two disease-specific PROMs had all nine properties classified as “sufficient” or optimally measured. The average proportion of psychometric properties considered sufficient was 38% for the five generic PROMs and 47% for disease-specific PROMs.
“Of 50 PROMs validated in CVD populations, only a handful adhered to all FDA recommendations, had psychometric properties rated as sufficient by COSMIN, or had minimally important differences established,” the researchers wrote. “Given the use of PROMs to guide FDA approvals of drugs and devices in CVDs, there is a need for better adherence to quality standards in PROM validation studies.”
Collapse