Most Patient-Reported Outcome Measures in CVD Fall Short

Only a minority of the instruments used in clinical and research settings to evaluate patient-reported outcome measures (PROMs) in CVD care adhere to quality standards set by the US Food and Drug Administration or to consensus-based metrics, according to a review of 50 such tools.

Health status PROMs that assess symptom burden, functional status, and health-related quality of life give the patient a voice in their care, provide tools for measuring treatment effects in research, and in some cases have been shown to be more-sensitive measures of health changes over time than clinician-elicited assessment.

Harriette Van Spall, MD, MPH (McMaster University, Hamilton, Canada), senior author of the review published in the Annals of Internal Medicine, said while traditional tools such as NYHA class and the Canadian Cardiovascular Society (CCS) scale for grading angina are the ones many physicians are most comfortable with, many disease-specific and generic PROMs are available, yet remain underutilized in CVD care.

Her group evaluated all of the existing health status PROMs that have been developed and validated in CVD populations.

“We found that few PROMs reported on the validation of all of the FDA-recommended psychometric properties,” Van Spall told TCTMD. “Few had properties that were rated as sufficient by COSMIN standards, which are standards that evaluate the validation process and reporting of PROMs.” The review also concluded that few PROMs had minimally important differences (MIDs) established, meaning there is no known threshold of improvement that is clinically relevant. Thus, results derived from those PROMs may be difficult to interpret.

“Not all PROMs are the same. When I select PROMs for my clinical trials, I often select them on the basis of what the patients prefer,” she added. “I will now have a narrower group of PROMs to present to them based on how they've been validated.”

A Map for Clinicians and Researchers

Van Spall and colleagues led by Derek S. Chew, MD, MSc (Libin Cardiovascular Institute, University of Calgary, Canada, and Duke Clinical Research Institute, Durham, NC), evaluated 50 health status PROMs described in 83 studies. Of these, 45 were disease specific and five were generic. The disease-specific tools were developed for a range of CV conditions, including arrhythmia, CAD, heart failure, and general CVD. The median number of items on each was 25 and the median time required for patients to complete them on paper or by tablet was 10 minutes.

Only 16% of PROMs reported on the validation of all psychometric properties recommended by the FDA, including invariance across sex, gender, and culture. Additionally, only 11 of 50 had available MID data or included a description of how to interpret changes in the scores over time. In terms of COSMIN standards, only two of 50 PROMs had all psychometric properties rated as sufficient in quality. In 64% of all PROMs, fewer than half of the psychometric properties were rated as sufficient.

None of the five generic PROMs, and only one in six of the disease-specific PROMs, fulfilled all the FDA requirements for product approval. Similarly, none of the generic PROMs and only two of the disease-specific PROMs had all nine COSMIN-defined properties classified as sufficient.

The PROMs that adhered most fully to both the FDA and COSMIN criteria were: Atrial Fibrillation Effect on QualiTy-of-life (AFEQT), Kansas City Cardiomyopathy Questionnaire (KCCQ), MacNew Heart Disease Health-related Quality of Life Questionnaire, Minnesota Living with Heart Failure Questionnaire (MLHFQ), Patient-Reported Outcomes Measurement Information System-Plus-Heart Failure (PROMIS-Plus-HF), Orthostatic Hypotension Questionnaire (OHQ), and HeartQOL.

The researchers note that while generic PROMs such as the EuroQol-5D (EQ-5D) are widely used in CVD clinical trials, “there seems to be a disconnect between the frequency of their use and the extent to which they fulfill FDA recommendations or meet COSMIN criteria for good measurement properties in CVD populations.” Among the issues found with the generic PROMs were insufficient content validity and the fact that none of them were developed with patient input.

“Although generic PROMs quantify a person's overall well-being, they do not include symptoms or functional limitations specific to CVD and are less sensitive to clinical changes compared with disease-specific instruments,” the researchers write. “Notwithstanding these limitations, generic PROMs

remain useful because they can be widely applied across different disease conditions to facilitate comparison.” They also can be useful to assess the health economics of pharmaceuticals or devices, they add.

“The take-home message is that we have now a comprehensive map of the characteristics and performance of each of these health status PROMs so that researchers and clinicians have a guide that can assist implementation of the right PROM for the right condition,” Van Spall said. “Our work highlights the need for careful adherence to accepted and standardized methodological criteria for PROM development and validation, and also clear reporting of the psychometric properties of PROMs.”