Integrating patient-reported outcomes in healthcare policy, research and practice

Keywords::

• clinical practice
• clinical research
• decision-making
• health policy
• health-related quality of life
• patient-reported outcomes

As use and understanding of health-related quality of life (HRQoL) and other patient-reported outcomes (PROs) have increased, researchers, clinicians and policy-makers have begun exploring different applications for these patient-centered outcome measures. Lipscomb et al. describe three arenas of application [1]. The macro-level involves use of PROs for population surveillance and policy. The meso-level includes applications of PROs in descriptive and analytical studies, such as clinical trials comparing interventions. The micro-level addresses PRO use for individual patient–clinician interactions. While PROs have, to date, been used most commonly at the meso-level, focus on the micro- and macro-levels has been developing in recent years.

The International Society for Quality of Life Research sought to explore different applications of HRQoL and other PROs and designated as the theme for its 16th annual conference: “Integrating HRQoL in Health Care Policy, Research, and Practice”. Three plenary sessions addressed use of PROs in each of these three areas. The policy plenary focused on using PROs for cross-national comparisons of population health [2–4]; the research plenary focused on using PROs for adverse symptom reporting in clinical trials [5–7]; and the practice plenary focused on using PROs to evaluate the quality of healthcare delivery [8–10]. In this article, we discuss the cross-cutting issues from the three sessions regarding how PROs are (or are not) used and the barriers to integrating PROs more completely. We also discuss proposed strategies for addressing these issues, thereby improving the application of PROs across healthcare policy, research and practice.

Several recurrent themes emerged across the three plenary sessions. First, great strides have been made over the past several decades in documenting the value of the patient’s perspective. Second, despite this value of the patient’s input, there is some hesitance to include PROs as one of the key sources of information in healthcare decision-making owing to issues related to measurement and interpretation. Third, methods for addressing the measurement and interpretation challenges are currently available and new methods are being developed. We discuss each of these themes in turn below.

In all three sessions, it was widely recognized that the patient’s voice provides a unique perspective and is valued. Salomon noted that this has not always been the case, reporting on the evolution of self-rated health measures from ‘irrelevant’ to ‘important’ for population health assessment [4]. PROs’ ascendance in importance is also evident in the evaluation of healthcare delivery, where patient ratings have been referred to as the ‘bottom line’ [10] and where they are increasingly being applied in various contexts [8–10]. Feeny noted that a multidisciplinary team of clinicians, having been introduced to the use of PROs to assist in managing individual patients in a formal study, came to rely on the information from PROs and subsequently made the routine use of PROs part of standard care [8,11]. Burke noted the importance of the patient’s perspective when balancing the benefits and risks of medical products [7]. The release of the US FDA guidance for industry Patient-Reported Outcome Measures: Use in Medical Product Development to Support Labeling Claims underscores the increasing importance of PROs for evaluating the impact of new medical interventions [12]. Thus, across all three sessions, the value of the patient’s voice was clear.

Despite this value placed on the patient’s perspective, there is some hesitance to include PROs as one of the key sources of information in healthcare decision-making. That is, while the patient’s voice provides a unique perspective, it is not this perspective that is generally given primacy. There are a variety of reasons why this may be the case. Two prominent issues that were raised in the plenary sessions were challenges in measuring PROs and in interpreting the findings.

With regard to measuring PROs, several speakers focused on the challenges associated with developing and applying items that are cross-culturally appropriate for population health assessments [2–4]. Salomon also demonstrated that responses to the same question can vary depending on question order or sampling strategy [4]. For evaluating the quality of healthcare delivery, Browne noted the importance of PROs that are sensitive and specific measures of quality so that causal links can be established [9]. Santanello stressed the importance of maintaining the traditional method of collecting adverse event information from patients by using a healthcare provider who can help probe symptoms and link symptoms to other data (e.g., laboratory data and signs) and the potential need for further evaluation [6]. She cited an example in which patients were unable to report edema reliably, and a second example where PROs could not distinguish adverse effects of non-steroidal anti-inflammatory drug from nonspecific gastrointestinal symptoms in patients receiving placebo. On the other hand, Santanello also pointed out that PRO diaries could be useful memory aids. As discussed later, addressing these measurement challenges will promote further integration of PROs in decision making.

Interpretation challenges associated with PROs also need to be addressed. Bullinger and Ravens-Sieberer suggest that health is a universal concept, but that there is cross-cultural variation in how the concept is endorsed [2], making it difficult to determine the meaning of findings across countries. Similarly, in evaluating the quality of healthcare delivery, Browne emphasized that it is not enough to know how one provider compares to others, rather, that providers need to be able to use the outcomes data to identify what areas require improvement [9]. The complexity of interpreting adverse event reports is illustrated by Basch’s presentation of data demonstrating that patient reports of adverse symptoms are more highly correlated with measures of functioning and well-being such as HRQoL, while clinician reports of adverse symptoms are more highly correlated with clinical end points, such as death or hospitalization [5]. Similarly, Santanello suggested that PRO reports of adverse symptoms are more appropriate for tolerability assessments than adverse event reporting [6]. As PROs are used more commonly and researchers and clinicians become more familiar with the meaning of the findings, interpretability will improve.

It should also be noted that many of the challenges mentioned earlier relating to measurement and interpretation are not unique to PROs. Hahn et al. compared the reliability of PROs to other commonly used clinical measures and found that PROs compared favorably [13]. Still, there is clearly room for improvement. Several suggestions were proposed during the plenary sessions. The Patient-Reported Outcomes Measurement Information System (PROMIS) was highlighted as a possible means of improving measurement rigor [10]. The PROMIS item banks and computer-adaptive tests are designed to improve measurement precision while limiting respondent burden [14]. Salomon reviewed three techniques that could be useful to improve the comparability of data across cultures: identifying item cut points using gold standard validation measures, using anchoring vignettes to identify and adjust for cut point differences, and using comparable and homogenous groups to establish cross-population comparability [4]. These techniques aid interpretation by estimating what variation results from measurement challenges and what variation is probably owing to real differences between groups. For evaluating the quality of healthcare delivery, Browne provided an example in breast reconstructive surgery demonstrating how applying clinically meaningful labels to outcome scores can help surgeons compare their outcomes to those of other surgeons and thereby identify areas for improvement [9]. In this strategy, descriptive labels are assigned to numeric scores (e.g., breasts equal in shape and size unclothed is better than acceptable shape when clothed). To help improve the measurement of PROs at the individual patient level such as that described by Feeny [8,11], Donaldson provides several useful strategies, including more frequent low-burden assessment, more realistic confidence levels and the use of population data to provide context to scores [15].

In conclusion, it is clear that the patient’s voice is valued across the different applications of PROs. Still, there is hesitance to include PROs as one of the key sources of information in healthcare decision-making. This hesitance appears to result from challenges in measuring PROs and interpreting the meaning of PRO findings, but methodologic advances are helping address these challenges in measurement and interpretation. Thus, there is reason for optimism that PROs will increasingly be integrated in decision-making as familiarity with the measures continues to grow and methods to address measurement and meaning are developed and adopted.

Financial & competing interests disclosure

Claire Snyder is supported by a Mentored Research Scholar Grant from the American Cancer Society (MRSG-08-011-01-CPPB). Claire Snyder and Michael Brundage served as co-chairs of the 16th Annual Conference of the International Society for Quality of Life Research (ISOQOL). They currently serve as members of the ISOQOL Board. The authors appreciate comments received on an earlier draft, which was circulated to the ISOQOL Executive Committee and the nine plenary speakers. The views expressed in this article are solely those of the authors, and do not necessarily represent the opinions or policy of ISOQOL. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.