Making use of patient-reported outcome measures in health care : the case of EQ-5D in the Swedish national quality registries

Abstract

Introduction: Patient-reported outcome measures (PROMs) have been widely used in clinical trials and research, for example to monitor the health of specific populations or to evaluate treatment effects. In the Swedish national quality registries (NQRs), structured individual-level data for specific patient populations are collected in routine health care settings. In addition to disease-related information, most NQRs include PROMs data. The most common PROM in the NQRs is EQ-5D, which is a generic instrument that can be used for measuring and valuing health across different patient populations and disease areas. The EQ-5D questionnaire includes questions addressing five dimensions of health (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression) and a visual analogue scale (EQ VAS). In addition, EQ-5D provides an indirect method for obtaining health state values and is therefore widely used for the purpose of economic evaluation in health care. There has been increasing interest in the use of PROMs in routine health care. Ideally, routinely collected PROMs data could be used to inform decisions to improve the quality of care and, ultimately, the health of the population. In addition to the potential benefits in patient-clinician encounters, aggregate-level PROMs data could for example be used for assessing the effectiveness of treatments, for detecting variations between health care providers or regions, or as input for decision-making. Real-world cases in which PROMs have been widely implemented in routine health care are needed to better understand the actual use of routinely collected PROMs data.

Aim: The overall aim of the thesis was to increase knowledge on the use of PROMs collected in routine health care. Using the case of EQ-5D in the Swedish NQRs, this thesis addressed the overall aim by investigating current practices of routine collection and use of EQ-5D data, and by exploring measurement properties, thoughts behind patients’ responses, and applications of EQ-5D as an outcome measure.

Methods: All four studies included in this thesis examined aspects of making use of EQ-5D data, but differed in their designs, samples, data, and analyses. Study I provided an overview of how EQ-5D data were collected in the Swedish NQRs, and how the collected EQ-5D data were made available and used. Information for each registry was obtained from webpages and through personal communication with representatives from the registries. In Study II, twenty patients with type 1 diabetes participated in qualitative individual interviews. Participants were asked to describe their thoughts out loud while reporting and valuing their own health, using EQ-5D-5L, EQ VAS, and time trade-off (TTO). The interviews were analyzed using qualitative thematic analysis. Study III and Study IV were based on data obtained from two NQRs: the Swedish Amputation and Prosthetics Registry and the Swedish National Quality Registry for ECT. In Study III, patients responded to either EQ-5D-3L or EQ-5D-5L six months after a major lower limb amputation (LLA). The assessment and comparison of the measurement properties for the two EQ-5D versions included analyses of feasibility, response patterns, informativity, and convergent and known-group validity. In Study IV, multiple linear regression analyses were used to examine the association between pulse width and health-related quality of life (HRQoL) after electroconvulsive therapy (ECT) in patients with unipolar or bipolar major depression.

Findings: In Study I, the overview of current practices of the collection, presentation, and use of EQ-5D data in the NQRs showed that EQ-5D was administered across various populations and interventions, and often in combination with disease-specific measures. EQ-5D data were most frequently collected in registries targeting patients with conditions related to the musculoskeletal system. In Study III, the assessment of measurement properties indicated that there were advantages to using EQ-5D-5L over EQ-5D-3L in patients with a major LLA, mainly due to improved informativity and validity of the descriptive system. In Study II, thoughts expressed during the qualitative interviews revealed some variation in the aspects considered and the time perspectives and reference points used when assessing one’s own current health using the EQ-5D-5L, EQ VAS, and TTO. Some participants expressed a desire to discuss their responses with the health care provider. The overview showed several examples of ways in which routinely collected EQ-5D data were used, including assessment of interventions, health economic studies, benchmarking, in quality indicators, quality improvement, and in patient-clinician encounters. Still, 19 of the 41 NQRs reported that they were unaware whether the collected EQ-5D data were used for follow-up, quality improvement, or decision-making. In Study IV, which was an example of the use of EQ-5D data for assessing real-world outcomes associated with an intervention, the results showed no robust associations between pulse width and HRQoL after ECT. Nevertheless, relatively large mean improvements in HRQoL were observed for patients treated with ECT for unipolar or bipolar depression, regardless of the pulse width received.

Conclusions: The findings from this thesis contribute to the understanding of the use of PROMs data routinely collected in the Swedish NQRs, of considerations in the choice of EQ-5D version in specific populations, of ways in which EQ-5D data can be used to assess specific interventions, and of the interpretations and thoughts behind patient’s responses when reporting and valuing their own health. The case of EQ-5D in the Swedish NQRs revealed several opportunities of making use of routinely collected PROMs data with the ultimate intention to improve health care and patient health. Still, there is potential to increase the use of EQ-5D data for follow-up, quality improvement, or decision-making. For future implementations of PROMs in routine health care, it may be useful to further explore how routinely collected PROMs data could be of most use to patients and other stakeholders, and to further explore prerequisites for making use of routinely collected PROMs data at different levels of the health care system.