Patient-reported outcome measures and patient engagement in heart failure clinical trials: multi-stakeholder perspectives

Introduction

In addition to increased risks of mortality and hospitalization, heart failure (HF) is associated with symptoms that can result in physical and social limitations, impaired quality of life, and increased rates of depression, anxiety, and stress.^{1, 2} While these latter outcomes have an important impact on patients' lives, they are not always captured as primary, or even secondary, endpoints in clinical trials.^{3, 4}

One method for quantifying patients' lived experiences is through the formal collection of patient-reported outcomes (PROs) that directly assess the patient's experience, health behaviours, and the impact of the disease and its treatment on patient health status.⁵ There are three general components of patient health status: symptom burden, functional status (physical, mental, emotional, or social functional limitations), and health-related quality of life (HRQoL) (Table 1).³ Since each patient views their own health in different ways, HRQoL is measured distinctly from symptoms or functional limitations. Thus, it is essential to assess this aspect with specifically designed tools.³

Patient-reported outcomes provide a unique perspective of the patient's experience of their illness and the effects of its treatment. These outcomes are relevant, not just to the patients, their families, and treating clinicians, but also to trialists, regulatory agencies, payers, and industry partners who seek to better understand these patient-centred outcomes. For example, PRO data may be used by trial sponsors to support a new product indication or labelling claim based on patient-reported treatment benefits, while payers may use these data to help determine the comparative value of different therapies.⁶

The Global CardioVascular Clinical Trialists (CVCT) Workshop brings together cardiovascular clinical trialists, patients, clinicians, and regulatory, government and industry representatives to discuss challenges facing cardiovascular clinical research. This paper is based on discussions at the 18th Global CVCT meeting (www.globalcvctforum.com) and aims to look at the benefits and challenges around the use of PROs to improve patient care, to inform drug development, and to aid regulatory and reimbursement decisions. The full session presentations and video recording can be accessed on the CVCT website.

Examples of patient-reported outcome measures used in heart failure

General measures of health status (e.g. 36-item Short Form Health Survey [SF-36], EuroQol-5 Dimension [EQ-5D]) can provide the clinician and patient with useful information, but are not disease-specific and are designed to quantify total impact of all current health conditions on overall physical and emotional functioning. PROMs are commonly used in HF, and a number of HF-specific PRO measures (PROMs) including the Minnesota Living with Heart Failure Questionnaire (MLHFQ) and the Kansas City Cardiomyopathy Questionnaire (KCCQ) (Table 2), are now becoming more commonly used.^{4, 7} These tools not only assess HF-specific health status, but have prognostic utility, with the KCCQ shown to be more sensitive to change^{8, 9} and superior to the MLHFQ for the prediction of clinical outcomes (death/transplant/left ventricular assist device implantation and hospitalization).¹⁰

It is important that PROMs be developed with patients, family members, and caregivers, and many may need ongoing updating as environmental and other issues faced by patients may change over time. It is important to consult patients from diverse demographic groups and communities.

Minimal clinically important difference

While the primary analyses of most clinical trials compare the mean differences across populations of studied patients, these can be difficult to interpret from a clinical perspective. In contrast, categorizing score changes into gradations of clinically important changes can provide deeper insight into how much better (or worse) patients are likely to feel after a particular treatment versus another. It is also necessary to understand the degree of change that is associated with increased risks of clinical events. The minimum difference in score of any outcome that patients can perceive as either beneficial or harmful is referred to as the minimal clinically important difference (MCID).^{11, 12} The MCID provides a method for interpreting changes in scores to help inform treatment decisions, and is also useful when planning clinical trials and calculating sample size.^{11, 12} However, a majority of PROs do not have established MCIDs, and studies typically report whether there is a statistically significant difference in health status between treatment groups, but do not put this in the context of MCID or larger clinically important changes.⁷ Population-average changes often do not differ by thresholds that are greater than the MCID for individual patients, which can make inferences about the magnitude of health status improvement with the interventions difficult to appreciate. Responder analysis, which describes the proportions of patients in each treatment arm who experience changes in their health status of varying clinical magnitudes, is one way of assessing clinical relevance in this setting, although there are arguments against this analytic approach too.^{13, 14} In particular, assessment at a single time point is poor evidence to classify a patient as a responder.

Estimating MCID can be done using either an expert consensus, a distribution-based method, or an anchor-based method. The distribution-based method compares changes in scores and small, moderate, and large changes in patient health status. This method relies on the statistical properties of the distribution of scores, and is used to estimate whether the magnitude of change in different groups in response to an intervention is more than would be expected from chance alone.^{11, 12, 15} An anchor-based determination compares the patient's numerical change in score for an outcome measure (e.g. changes in KCCQ) with an external subjective, independent evaluation of improvement (e.g. Patient Global Assessment [PGA]).^{11, 16}

The KCCQ is the HF-specific PRO that has undergone the most extensive research to define clinically important changes.^{15, 17-20} These efforts have led to a general consensus that a change in KCCQ total score of >5 points has been considered to be the MCID associated with a small improvement in patient health status.¹⁷ Changes of 10 and 20 points are considered to be moderate to large, and large to very large, clinical changes in patient health status. Such thresholds can transform the mean differences in scores along a 100-point range into the proportions of patients (e.g. number needed to treat) experiencing different magnitudes of improvement with treatment. It is also important to understand the stability and time course of benefit of a given intervention; some interventions may provide benefit early while others may take longer, and similarly, some benefits may be short-term and others more sustained.²¹

Using an anchor-based methodology (based on PGA), and data from the FAIR-HF (Ferinject Assessment in Patients with Iron Deficiency and Chronic Heart Failure) trial, the MCID threshold for a small positive change was associated with a change in KCCQ overall summary score of 3.6.¹⁶ This was somewhat lower than the typically accepted 5-point change described above.¹⁷ This demonstrates the utility of anchor-based methodology and supports the concept that even small improvements in HRQoL of a treatment over placebo can be clinically meaningful (Table 3).^22-26

In addition, HRQoL may decline over the course of the disease due to deterioration of HF, but also due to other factors that may not be amenable to HF treatment, such as physical and social deterioration related to aging, non-cardiovascular comorbid conditions, and psychological factors.^{22, 27} Therefore, if a treatment can improve or at least stabilize a patient's HRQoL that could be considered beneficial.

Using patient-reported outcomes to improve patient care

Patient-reported outcome measures are not yet widely incorporated into routine clinical practice for patients with HF, with clinicians generally placing greater importance on echocardiographic and biomarker responses, or using non-standardized ways to obtain subjective patient feedback and gauge symptomatic treatment response.²⁸ Yet PROs reflect what patients actually perceive and have been shown to improve the accuracy of clinicians' assessments of patients' New York Heart Association (NYHA) classification when available to the assessing physician.²⁹ Moreover, they can be used to assess their experience and behaviours, monitor therapeutic effectiveness, and prompt discussions relating to prognosis or end-of-life care. PROs may also be useful in guiding referrals for invasive cardiac care (e.g. percutaneous coronary intervention or coronary artery bypass grafting), non-cardiac care (e.g. psychiatric care, physiotherapy, or occupational therapy, and for assessing the quality of delivered healthcare). This suggests a potential role for interdisciplinary patient advocacy efforts in support of the widespread collection, standardization and implementation of PROs in clinical practice. Quality of life should be consistently and correctly assessed and managed throughout the patient journey. Incorporating routine measurements of HRQoL in clinical practice using PROMs can help both clinicians and patients to discuss the individual's experiences, and to consider whether the treatment programme is meeting their expectations. The integration of the routine use of PROMs in clinical practice requires investing in the resources to collect, score and display the results for use by patients and their providers. Digital technologies, integrated in the electronic medical record are a promising strategy to help make this feasible. The instruments may be offered through patient portals, a smart phone app, or through tablets in clinics, thus enabling some patients to complete the questionnaires while awaiting their appointment. This is an active area of investigation and the benefits of routinely using PROMs in clinical care are just beginning in cardiology,^29-31 although their benefits have been proven in cancer.^{32, 33} Patient organizations can provide coaching and support for patients to self-manage their condition and improve their quality of life. Formal assessments can also highlight unmet needs and areas for potential improvement in care delivery.³⁴

Patient perspective

The priorities and expectations of patients and clinicians often differ. Clinicians frequently are more concerned with cardiac anatomy and physiology, reducing mortality, and specific management issues, while patients just wish to feel better.^{28, 35} However, both clinicians and patients generally assign high priority to improving HRQoL. Multiple studies have reported that patients often prefer a healthier life, with the majority willing to accept some decrease in longevity in return for improved health status.^{1, 36, 37} For example, in a survey of over 500 HF patients by the Pumping Marvellous Foundation, only 3% said that how long they live was the most important aspect of HF treatment, while 97% prioritized quality of life or both longevity and quality of life (Figure 1).²⁸ A survey within a family medicine clinic assessing disability and quality of life among HF patients found that almost 20% had severe disability, and another 28% had moderate disability.¹ The main areas affected were life activities, getting around, and participation in society.

Using patient-reported outcomes to better inform drug development and clinical trial design

Despite there being a significant increase in the inclusion of PROs in HF clinical trials over the past 20 years,⁴ there remains uncertainty around their interpretation and their role in the product development lifecycle. To further optimize study design requires a clearer understanding of the downstream utility of PRO data.

Using patient-reported outcomes to better inform regulatory and reimbursement decisions

Incorporating patient-reported outcomes into clinical practice guidelines

Despite recognition of the importance of functional capacity and HRQoL for patients with HF, recommendations in HF guidelines continue to be based predominantly on morbidity and mortality outcomes.^{56, 57} Recommendations based on PRO data are generally not class I, even when the data fulfil the criterion of ‘data derived from multiple randomized clinical trials or meta-analyses’. For example, in the 2021 ESC HF guidelines the recommendations relating to iron supplementation continue to be class IIa (‘should be considered’) rather than class I (‘is recommended’ or ‘is indicated’).⁵⁶

This is likely related to ongoing issues around standards for reporting and interpreting PRO data. In HF clinical trials, the NYHA class is not reported as a mean for each treatment arm, but rather as the proportion of patients within each NYHA class. This categorization facilitates understanding of the applicability of the trial to a certain patient population. Yet this approach is not used for HRQoL variables, with each trial reporting their own mean scores for the PROM used. An established and agreed-upon range for PROMs such as KCCQ or MLHFQ scores would enable more standardized reporting and interpretation of PRO data.¹⁵

Relevant questions to be considered include how to assess PRO fluctuations to identify responders and how to convert patient- and population-improvement concepts into interpretable results for action by the clinician, including how to define the MCID for within-patient changes in different patient populations. Until these questions are answered, guideline recommendations will continue to receive a lower-class rating despite a high level of evidence. One approach would be to issue a class I recommendation where warranted especially in rare diseases, but one that is accompanied with a specific statement acknowledging the lack of long-term mortality/morbidity data. Such a shift in guidance would further acknowledge the importance of improvements in PROs.

Beyond patient-reported outcomes: involving patients in clinical trial design and execution

The focus of PROMs has been on developing and validating instruments that have optimal test characteristics. However, this often fails to consider the user-friendliness of these tools. Although using PROMs is a step in the right direction, it is important to ensure that they are being designed, and this type of data is being collected, from a diverse population that is representative of individuals with the disease. The burden of disease extends to family members and caregivers, and their perspective should also be solicited when designing PROMs. Having comprehensive representation of patients in clinical trials will ensure that HRQoL, efficacy and safety data are more relevant, and generalizable to broader populations.

‘Patient-centric’, the concept of having the patient at the centre of clinical research and care, underscores that the goal of generating new knowledge is to improve patient care and outcomes. However, for some people this implies patients in the middle with things going on around them. Patient-centred research is meant to anchor research on the values and priorities of patients and to include them as partners, but true patient partnership continues to be underused. Even in Patient-Centered Outcomes Research Institute (PCORI) funded studies, engagement often included only passive consultants, while true collaboration and shared leadership occurred in only 37% of trials. A report from the PCORI funded ADAPTABLE (Aspirin Dosing: A Patient-centric Trial Assessing Benefits and Long-term Effectiveness) study concluded that patients must be an active part of the steering committee, and that true patient engagement requires not just one or two ‘token’ patients but a group of at least five or six diverse patients to produce effective, shared ideas and confidence.⁵⁸ It is critical that patients are seen as partners in their healthcare management and in clinical trials from development through trial design, execution, and dissemination.

The CVCT Forum has a long history of integrating patients as speakers and panellists during its annual meetings (www.globalcvctforum.com/patient-trialists). Each patient brings their own perspectives and priorities, their own experiences and knowledge about their disease to the table, and thus they are one of the most valuable resources in the development of improved treatments for HF.

Summary

Steps are needed to address the challenges around the standardization and use of PROs in cardiovascular trials, regulatory and payers' decisions and patient care. One of the most important aspects is to ensure that patients, family members, and caregivers from diverse demographic groups and communities are involved in the development of PROMs, as well as the design, analysis and interpretation of PROs in clinical trials. There is a need for standardization not only of the PROM instruments, but also in procedures for analysis, interpretation and reporting of PRO data. A summary of recommendation is shown in the Graphical Abstract and Table 4.