Identifying key domains of health-related quality of life for patients with Chronic Obstructive Pulmonary Disease: the patient perspective

Chronic Obstructive Pulmonary Disease (COPD) is one of the leading causes for mortality worldwide [1]. COPD is characterized by dyspnea, chronic cough, sputum production, a decreased exercise performance and reduced physical activity level [2],[3]. COPD cannot be cured; the main goal of COPD treatments is managing symptoms and their effect on the patient’s health-related quality of life (HRQoL) [3]. Consequently, HRQoL has become an important outcome measure in COPD research and clinical care.

HRQoL encompasses the physical, functional, emotional, and social well-being of the patient [4]. HRQoL is a highly subjective experience by definition and therefore measured by self-report questionnaires (patient reported outcome measures; PROMs). Many questionnaires are available to measure HRQoL in patients with COPD, covering a wide array of domains ranging from symptoms such as dyspnea, cough and wheezing, to social and emotional functioning (see [5] for an overview), and it can be a challenge to choose the appropriate questionnaire. Aspects to consider include domain coverage, test length (e.g. [6],[7]), recall period (for example, last year or last few weeks), the interpretation of (change in) scores, psychometric properties (e.g. [8]), and whether it is generic or disease-specific, among other things [9]. Since both generic and disease-specific instruments have desirable properties, several researchers advocate using both types of instruments to assess HRQoL in patients with COPD [10],[11]. Generic instruments can be used with any (patient) population, facilitating direct comparison among populations, including the general population. Disease-specific instruments are developed for a specific patient population, which may increase the sensitivity to measure intra-individual change after an intervention. However, clinicians tend to prefer short instruments for use in clinical practice, so the test battery should not be too lengthy. Unfortunately, short instruments often fail to provide a sufficiently detailed picture of the most relevant aspects of a patient’s HRQoL.

In recent years, Computerized Adaptive Testing (CAT) [12] based on Item Response Theory (IRT) [13] has been put forward as a possible solution to the dilemma. A CAT is a digital questionnaire tailored to the individual patient, resulting in each item (question) contributing valuable information, while maintaining comparability across patients. Item selection in a CAT is dependent on a patient’s response to previous items. In this way the latent trait estimate is continuously adjusted, until a specific level of measurement precision (reliability) is reached. In recent years, a large number of IRT-calibrated item banks have been developed to measure domains of HRQoL. This initiative is referred to as “PROMIS” (Patient Reported Outcomes Measurement Information System) [14], and resulted in item banks covering physical, mental, and social health. Both CATs and short forms have been developed using the item banks [15]. The PROMIS banks are generic, and have been calibrated using a large sample from the general population as well as subsamples of numerous patient groups, including COPD. However, it is still up to the user to decide which domains to select.

Since HRQoL is a highly subjective experience, it is not only paramount to use PROMs to measure it, but also to involve patients in the development and selection process of PROMs. It has been repeatedly shown that taking patient perspective into account during PROM development allows for the identification of problems with item content and coverage as well as problems with domain coverage and/or poor operationalization (e.g., [16]-[27]). These findings have inspired several guidelines concerning the development of PROMs. The European Organisation for Research and Treatment of Cancer (EORTQ) even proposes that patient perspective should be leading when “…compiling an exhaustive list of relevant QoL issues that cover the domain(s) of interest” [28]. Within the PROMIS framework, patient perspective has played an important role in item bank development [29]-[32]. Focus-groups were used to evaluate item content, coverage and wording, as well as the content validity of the item banks by evaluating the fit between the domain map that was drafted by a panel of experts, and concepts that focus-group participants identified as important aspects of HRQoL [33]-[36]. It should be noted that in this set-up, feedback from experts was used to conceptualize a domain-map, and subsequently input from patients (responding to open questions) was used to verify the domain map. Patients were not explicitly asked to comment on the domain-map, or to choose domains themselves.

Currently, there is no consensus as to which instruments are most appropriate to assess HRQoL in COPD patients. However, it has been argued that combining generic and disease-specific measures may provide the most useful estimations. PROMIS is probably the best generic instrument that is currently available to measure HRQoL; it consists of a wide range of generic item banks that have been developed with a rigorous and sound methodology. However, no core sets or recommendations are available to guide the selection process of PROMIS item banks for use with COPD patients (or any disease group for that matter); nor is it clear whether all domains relevant for COPD are covered by current PROMIS item banks. In line with aforementioned guidelines and published studies, we agree the patient perspective should play an important role in domain selection and assessment of domain coverage. The aim of this paper is to identify which domains of HRQoL are most important from the COPD patient’s perspective, and why. We will do this in two steps: first patients are asked to describe important domains freely; second, they are presented with PROMIS domains and are asked to select those most relevant to them. The patient is invited to explain the motives behind their choice, providing insight in the way they conceptualize and experience HRQoL. The results of this study can be used to inform instrument selection and development. This study is the first of an international collaboration (the Netherlands and Canada) that involve two research groups following the same steps (Additional file 1) to create a CAT to measure HRQoL in COPD patients.

Twenty-one COPD patients (13 men; mean age 66.1 (SD = 9.3) years) were included; 24% of them were recruited in Rotterdam; 19% were inpatients; 9.5% had GOLD stage I, 38% GOLD stage II, 43% GOLD stage III, and 9.5% GOLD stage IV. The face-to-face interviews took 15–30 minutes. See Table 3 for descriptive characteristics for each patient.

Physical health and social health emerged as important overarching themes, while domains associated with mental health were hardly chosen or mentioned. Physical health emerged as the most important theme from the spontaneous statements that patients made when asked about HRQoL in relation to COPD, while two of the six most frequently chosen PROMIS domains (fatigue and physical functioning) pertain to the overarching domain of physical health [40]. Coping with COPD emerged as the second most important theme from the spontaneous statements, while the remaining four PROMIS domains pertained to social health (instrumental support, ability to participate in social roles and activities, companionship, and emotional support) [40]. In addition to the obvious overlap between the outcomes of the open question and the card sorting task (PROMIS domains) with respect to physical health, there is a more subtle overlap with respect to coping. Analysis of the statements patients made to explain their choice of PROMIS domains, showed that coping was a frequently mentioned theme for three of the six most popular PROMIS domains. It should be noted that the statements elicited by the cues (PROMIS domains) were far richer in content than the spontaneous statements, providing us with a lot of information on the patients’ perspectives on these different domains of HRQoL. However, the open question elicited highly valuable information as well, since patients were free to describe HRQoL in their own words; this resulted in an important main theme (coping with COPD) that is not explicitly covered by any of the existing PROMIS domains. These findings underline the importance of using a funnel-shaped questioning strategy, as recommended by Brod and colleagues [27].

Surprisingly, very few patients mentioned or chose domains pertaining to mental health; only two patients mentioned anxiety when answering the open question, while one patient mentioned anger. During the card sorting task, four patients chose the PROMIS domain anxiety, and two patients chose anger. The PROMIS domain depression was not chosen by a single patient. These findings are striking, since the literature indicates that the prevalence of depression or the level of depressive symptoms is much higher in patients with COPD than in the general population; and also anxiety has been shown to be a serious problem in COPD [41]-[46]. Depression is still a taboo-subject for many; especially elderly people have been found to underreport depressive symptoms [47]. It should be noted, though, that although depression was not explicitly chosen, depressive feelings were mentioned by some patients in relation to fatigue, autonomy and acute exacerbations. The interviewed patients may have felt a bit daunted when confronted with the word directly. Additionally, many COPD patients are plagued by feelings of guilt [48], knowing that they are at least partly responsible for the onset of their disease (tobacco smoking). As a result they may feel the need to keep a stiff upper lip, believing they have no ‘right’ to complain, resulting in an underreporting of psychological distress. We discussed this point with a number of COPD experts (researchers and pulmonologists), and several of them pointed out that it requires quite some probing in an interview/consultation to get the patients to open up about their depressive feelings. We therefore presume that the patients not choosing depression may have been a method-effect, and it cannot be ruled out that this domain is in fact relevant to patients with COPD. When it comes to anxiety, Strang and colleagues [46] found that patients experienced high levels of anxiety related to COPD symptoms, mostly in connection with the themes death anxiety (fear of suffocation, awareness of death, fear of dying, separation anxiety) and life anxiety (fear of living, fear of the future). Furthermore, patients also talked about coping strategies they used to deal with the feeling of suffocation, or fear of dying. In their study, Strang and colleagues started with a very general question (“Can you describe what living with COPD means to you?”), but continued with probing questions, including “What are your experiences of anxiety?” Their asking about anxiety directly may partly explain the differences between the results of our open question and their findings. When comparing these anxiety-related themes to the PROMIS anxiety item bank, it becomes apparent that the PROMIS anxiety items are formulated in a very general way. This may be an explanation as to why only few patients in this study chose this domain. It may make more sense to include disease-specific anxiety items that are directly related to COPD symptoms (such as shortness of breath) in an instrument that aims to measure HRQoL in COPD patients, rather than using the PROMIS anxiety domain.

When comparing the most important themes and domains that emerged in this study, to domains covered by the most frequently used instruments to assess HRQoL in COPD [5], a few things stand out. Among the domains most frequently covered by existing instruments are energy/fatigue and social functioning; our findings clearly show support for these being important domains from the patient perspective. However, although many existing instruments focus at least in part on COPD symptoms (dyspnea, cough, phlegm, chest tightness), only few focus on the impact of COPD on daily life [5]. Our results show that this is an important shortcoming of the existing instruments, as many patients indicated they struggle with the impact of COPD on daily activities, as well as finding a way to cope with this negative impact. Most patients indicated that physical activity triggers or aggravates their symptoms, and several patients indicated that they have to avoid or adapt their daily activities. For some patients, these limitations were so severe that they considered themselves as invalids. Many patients struggle daily with their dependency on others in performing daily life tasks; they feel ashamed and frustrated that they are no longer self-reliant.

It should be noted that some patients found it difficult or refused to follow the instructions. Four patients did not pick any PROMIS domain. They argued that they were in good physical shape or were at least able to do everything they wanted; or they fully accepted the consequences of COPD and were able to adjust to it. It may be that the PROMIS domains were too general, which could be solved by adding a disease-specific module to the PROMIS domain framework. Some patients had comorbid disorders and indicated that it was difficult for them to identify the exact impact of COPD. Other patients attributed part of the symptoms that impacted their HRQoL to the ‘normal’ ageing process. Some patients had trouble to pick exactly five important domains. Patients that picked fewer than five domains indicated that they had trouble seeing the difference among some of the domains.

The most relevant (sub)domains of HRQoL for patients with COPD were: physical health (fatigue, physical functioning), social health (instrumental support, ability to participate in social roles and activities, companionship, and emotional support), and coping with COPD. The latter is not explicitly covered by any of the existing PROMIS domains, or by the most popular traditional questionnaires that are used to measure HRQoL in COPD. Our results reaffirm what has been pointed out in previous studies [27]: engaging patients in PROM development and evaluation is crucial to ensure content validity. The identified PROMIS domains could form the basis for a CAT measuring HRQoL in COPD patients. We recommend the development of a disease-specific module to be included in the CAT, to capture important HRQoL aspects not covered by the current PROMIS domains. When selecting or developing instruments, we argue that it is important to take into account expert opinion as well as patient experience (see Additional file 1). Therefore, we will conduct interviews with healthcare professionals in a future study. To ensure the cross-cultural relevance and external validity of our findings, we will repeat these interviews (with patients and healthcare professionals) in other countries as well, starting with Canada.