Author + information
- Eldrin F. Lewis, MD, MPH∗ ()
- Cardiovascular Division, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts
- ↵∗Reprint requests and correspondence:
Dr. Eldrin F. Lewis, Cardiovascular Division, Brigham and Women’s Hospital, 75 Francis Street, Boston, Massachusetts 02115.
- European Medicines Agency
- Food and Drug Administration
- heart failure
- patient-reported outcomes
- quality of life
- regulatory approval
Health-related quality of life (HRQL) is a heterogeneous concept that attempts to represent a person’s perception of his or her well-being reflective of the balance between expectations and the reality of limitations, disease burden, and clinical course. HRQL is an important target for outcomes in clinical care, research, and patient goals, and thus tools have been established to measure HRQL and other patient-reported outcomes (PRO), which all are measured directly from the patient without interpretation and captures symptom status, functional status, and general health perceptions (1). Given increased attention in clinical research, the quality of the HRQL publications has improved over the decades. Approximately 7,000 publications involving heart failure (HF) patients have listed HRQL as a key word, and clinicians are more adept at interpreting these findings. Patient-centered care is defined as “Health care that establishes a partnership among practitioners, patients, and their families (when appropriate) to ensure that decisions respect patients′ wants, needs, and preferences and that patients have the education and support they need to make decisions and participate in their own care” (2). For many patients with HF, HRQL is equal to (and sometimes more important) than survival; thus, comprehensive care of the patient is achieved when all outcomes, including HRQL, are characterized in clinical trials that evaluate novel therapies.
Among patients with HF, PRO instruments can be used to evaluate efficacy and effectiveness of novel treatments, assess side effects from medicines or procedures, determine compliance with trial procedures and therapy, assess impact of therapy and patient’s disease on family and caregivers, and determine the relevance of these outcomes in the context of patients characteristics, belief system, and cultural or social environment (3,4). These PRO instruments may also serve as a surrogate outcome measure in clinical trials prior to initiation of the Phase 3 study if there is close association between changes in these instruments and clinical outcomes such as HF hospitalizations and mortality. In clinical practice, PRO instruments can be useful in a disease management program to monitor disease progression, provide audit feedback to the patient in order to facilitate medical decision-making, identify patients at risk for adverse events (including hospitalizations and death), and monitor the effectiveness of clinical management strategies and disease management interventions (5,6). This latter role allows clinicians to monitor the speed of recovery as well as the sustainability of recovery of HRQL, both of which may have large influences on the patient’s decision to proceed with risky interventions. For regulatory approval, PRO instruments could be helpful in the evaluation of new therapies involving patients with impaired HRQL related to a specific condition treated with an intervention, such as pain related to cancer or rheumatologic disorders, dyspnea related to asthma, pulmonary hypertension, or HF, and angina related to coronary artery disease (7). Symptom improvements seem to be a target for regulatory approval (7), such as seen in ranolazine, which was approved for treatment of chronic angina in November 2008. HF is complicated given varying symptom burdens for different patients and coexistence of comorbid conditions, each with its own symptom profile.
In this issue of JACC: Heart Failure, Psotka et al. (8) review attempts to review the PRO instruments utilized in research to determine the potential fulfillment of the regulatory requirements of the Food and Drug Administration (FDA) Guidance of Use of PRO measures for regulatory approval. Nineteen instruments were identified and reviewed, with the most common being Kansas City Cardiomyopathy and Minnesota Living with Heart Failure instruments. After a comprehensive review, the authors concluded that none of the instruments met the strict regulatory requirements and thus it is unlikely that the FDA would provide approval for a PRO indication based upon the use of these instruments. As the authors note, several comprehensive reviews of HRQL in HF have been performed and their focus is on the contemporary use of these instruments in the context of the recent guidance. The strengths include a great summary of the domains covered in psychometric properties of all instruments. The main criticism is the lack of a solution for achieving claims for improving HRQL in HF, which admittedly is in part due to the failed track record of recent PRO claims. It would also be helpful to provide an estimated time for completion of each instrument. Less widely used instruments that may not have been used in clinical trials may be missed. An assessment of adherence to the FDA guidance and its role in submission is beyond the scope of the paper.
The U.S. FDA and European Medicines Agency are important regulatory agencies that evaluate the benefits and risks of novel therapies and also legally determine what a commercial organization could say about the benefits and risks of a particular therapy (4,7). Because of the increasing use of PRO data, drug manufacturers started seeking claims for novel therapies which resulted in the final guidance for industry regarding PRO and medical product development to support labeling claims (9). A recent review of PRO labeled claims by the FDA suggested that 86% were for symptoms, 25% for functioning, and only 1 was for HRQL (4). Interestingly, generic instruments were used to support claims for symptoms and function, including Short Form-36, physical component summary, visual analog scales for multiple symptoms, functional assessment of chronic illness therapy–fatigue, and the Patient Global Assessment of Change. In another analysis of reasons for rejection of PRO claims submitted to the FDA, 38% of the cases were due to “fit for purpose” problems related to lack of adherence to minimum qualifications of a tool, instrument that was not specific for the trial population, and content validity issues (10). Another one-quarter of PRO claim denials were due to trial design (e.g., open-label interventions that can magnify the effect size), with smaller contributions due to statistical analysis, including multiple testing due to several instruments used, multiple domains covered, and lack of declaration of the primary PRO outcome measure. Only 14% of the denials were due to lack of statistical differences that demonstrated efficacy. These articles dovetail with the Psotka et al. (8) efforts to provide a strategy for marked refinements in the PRO process and subsequent claims for new treatments.
As suggested by Psotka et al. (8), the optimal PRO instrument has yet to be developed. This instrument must meet all criteria described in the FDA guidance. In addition there are several other features that would be helpful. First, it should be able to measure meaningful domains for all patients with a particular disease that takes into account the heterogeneous nature of these patients because of the disease progression, their backgrounds and culture, and goals in life. Instruments utilizing adaptable designs that expands (or collapses) domains that are not important on the basis of a few screening questions may minimize item burden and redundancy. Second, there are variable contributions of functions and symptoms to a patient’s HRQL. The activities that are required on a daily basis can greatly influence the impact of symptoms and function on a patient’s HRQL and this should be captured in a facile manner. Third, comorbid conditions have a marked impact on a patient’s sense of well-being, and may influence the responsiveness of any intervention on HRQL change scores (11). An instrument that captures disease-specific HRQL, generic cardiovascular (CV) HRQL, and overall HRQL would allow for comparison across the spectrum of CV and non-CV disease while maintaining the responsiveness of a disease-specific instrument.
As researchers and clinicians, we should take the challenge of the 2009 guidance for PRO instruments to refine our process for conduction of research that involves symptomatic patients with HF. Six simple strategies can both increase confidence in the data that clinicians and researchers review for HF patients and may reverse the trend in success rates of claims utilizing PRO data.
1. Standardize reporting of PRO data in manuscripts: At minimum, the baseline scores, mean change scores between baseline and the pre-specified time point, and magnitude of between-group differences should be reported in clinical trial PRO manuscripts.
2. Develop robust data analysis plan to analyze PRO data: The primary efficacy measure should be pre-specified with particular attention to which domain(s) will be improved by treatment. If possible, subjects should be blinded to the treatment arm. If not, then a longer follow-up period may mitigate some of the improvements in HRQL that occur due to placebo effect and increased attention from the study visits.
3. Collect PRO data beyond the study visit used to assess primary efficacy: HRQL improves relatively quickly for many interventions and shorter follow-up minimizes the impact of issues such as lost to follow-up, death, and stopping study drug on efficacy. However, collecting PRO data for beyond 3 to 6 months allows one to understand the long-term effect of an intervention on HRQL.
4. Assess the anticipated effect size and domains affected by intervention before Phase 3 trial: Investigators would not typically perform a Phase 3 study without some estimate of the impact of a novel therapy on the endpoints being measured (e.g., death and HF hospitalization). However, HRQL is often not measured systematically before embarking on a large trial and multiple instruments are used without a clear understanding of the PRO domains that will be influenced. Use of responder analyses to determine the hazard ratio for achieving meaningful improvements in HRQL and proportion of responders to a therapy can also provide some additional information beyond the p value. Also, identification of predictors of “HRQL nonresponders” to therapies could be helpful to clinicians and patients as these therapies are used in clinical practice.
5. Minimize missing data: This will be achieved by emphasizing the importance of these endpoints by: 1) including the PRO case report forms and research design in the main protocol; 2) increasing the scrutiny for missing data, including a form that has to be completed by the investigators when the data is not collected; and 3) decreasing the burden by only including instruments that are required to characterize the PRO of the trial population.
6. Enroll patients with minimum HRQL impairment: All patients are often evaluated to determine change scores in HRQL, regardless of baseline impairments. Given floor and ceiling effects of these instruments, it may be difficult to show a meaningful difference in the change scores. Just as a hypertension clinical trial often requires a minimum systolic blood pressure as eligibility requirements to determine the impact of an antihypertensive therapy on outcomes, there should be a minimum degree of HRQL impairment for novel therapies.
In conclusion, the paper by Psotka et al. (8) greatly enhances our understanding of some of the requirements of PRO tools in HF to achieve regulatory approval. By raising the bar on the rigor of HRQL assessment, we will be able to focus on a meaningful outcome for the millions of patients suffering with HF.
↵∗ Editorials published in JACC: Heart Failure reflect the views of the authors and do not necessarily represent the views of JACC: Heart Failure or the American College of Cardiology.
Dr. Lewis has reported that he has no relationships relevant to the contents of this paper to disclose.
- 2016 American College of Cardiology Foundation
- Committee on Quality of Health Care in America and Institute of Medicine
- Joyce E.,
- Chung C.,
- Badloe S.,
- et al.
- Kelkar A.A.,
- Spertus J.,
- Pang P.,
- et al.
- Psotka M.A.,
- von Maltzahn R.,
- Anatchkova M.,
- et al.
- ↵U.S. Department of Health and Human Services. Guidance for Industry. Patient-Reported Outcome Measures: Use in Medical Product Development to Support Labeling Claims. December 2009. Available at: http://www.fda.gov/downloads/Drugs/.../Guidances/UCM193282.pdf. Accessed May 31, 2016.
- DeMuro C.,
- Clark M.,
- Mordin M.,
- Fehnel S.,
- Copley-Merriman C.,
- Gnanasakthy A.
- Lewis E.F.,
- Kim H.Y.,
- Claggett B.,
- et al.