Systematic review of health-related quality of life models

Introduction

Health-related quality of life (HRQOL) has been identified as a goal for all people across all life stages by leading health organizations [1-3]. HRQOL, that is, quality of life relative to one’s health or disease status, is a concern of policymakers, researchers, and health care practitioners [4]. Especially important is the need to align HRQOL research priorities with the needs and values of patients and their families. Because of the multidimensional aspects of HRQOL, and the varied use of this term across many different health and disease conditions, researchers have used a variety of HRQOL models to guide their research. A conceptual model is a schematic representation of a theory that acts as a heuristic device to provide a better understanding of a phenomenon (e.g., HRQOL) by depicting interrelationships among concepts [5]. The term conceptual model has been used interchangeably as “conceptual framework, theoretical model, or theoretically based conceptual model [6].”

There are many HRQOL models applied across different health and illness conditions, across the lifespan, and among individuals, their families, and communities. HRQOL is commonly conceived as dynamic, subjective, and multidimensional, and the dimensions often include physical, social, psychological, and spiritual factors [7]. However, the specific dimensions are labeled differently by different authors [7]. For example, these broad dimensions subsume more specific dimensions such as emotions, cognitive function, economic status, and intelligence [8], and they may incorporate friends and family [9]. While the theoretical underpinnings of HRQOL may be consistent across models, variations in terminology for analogous concepts make comparison across studies very difficult. Differing conceptualizations of HRQOL limit the ability to have a coherent body of evidence to guide further HRQOL research and practice. Common HRQOL models provide essential structure to the conceptualization of HRQOL using common language that can be shared across studies. Identification and evaluation of common HRQOL models can help guide research and practice toward promoting or attaining optimum HRQOL for populations of interest. Thus, the purposes of this paper were to (a) identify the most frequently used HRQOL models found in the literature over the past ten years and (b) review and critique the most commonly used models using established criteria by Bredow [10]. Although Bredow’s [10] criteria were developed to critique middle-range theories in nursing research, they represent a comprehensive approach to theory analysis for review and critique of HRQOL models [10]. The overall goal was to determine the relevance of HRQOL models to research and practice aimed at improving HRQOL.

Methods

Several search engines were used to locate relevant articles. Initially, PubMed, MEDLINE, CINAHL, and PsychINFO were searched using the keywords quality of life, health-related quality of life, conceptual framework, conceptual model, and theory. Both quality of life and health-related quality of life were searched because these terms have been used interchangeably in the literature.

We limited our search to English language articles published between January 1, 1999 and August 31, 2010. The inclusion criteria were published articles pertaining to HRQOL models that had been used to guide (a) literature reviews, (b) instrument development studies, (c) descriptive or correlational studies, (d) intervention studies, or (e) practice. Articles in which research findings were used to derive HRQOL models were also included. We did not limit our search to specific populations (e.g., children, adolescents, adults, older adults) because we wanted a broad representation of the use of HRQOL models. Exclusion criteria were articles that did not pertain to humans, were non-English, or involved studies and information published only as dissertations, abstracts, editorials, or clinical opinion. Relevant articles were identified from the literature search using a three-step process. In the first step, authors working in pairs reviewed the article titles based on inclusion/exclusion criteria. In cases in which there was a lack of consensus between the two reviewers, a third reviewer was sought, consistent with methods outlined by the Joanna Briggs Institute [11]. In the second step, titles that met the criteria were further evaluated. Authors, again working in pairs, reviewed abstracts and reached agreement about whether the abstracts met inclusion/exclusion criteria. In the third step, the identified articles were obtained and evaluated by the same pairs of authors. Full text articles were reviewed and again were included only if the pairs agreed the article met the criteria.

The paired authors then extracted and consolidated pertinent information from the articles into a review table. Column variables in the table were: author and date, country of origin, purpose, methods, design, and model. Each row represented a unique article. All authors engaged in group discussion to reach consensus on articles to be included in the review and to determine the format in which to present review findings. After reviewing articles, the most commonly used HRQOL models were identified, fulfilling the first purpose of this review. For purpose two, each of the most commonly used HRQOL models was critiqued by the author pairs using established criteria by Bredow [10]. After considering several alternatives, we chose Bredow because he incorporated the most comprehensive criteria for evaluating theories, frameworks, and models [10]. Although these criteria are used to evaluate middle range theories in nursing research, they are also useful in evaluating quality of life theories [10]. A description of Bredow’s [10] criteria appears in the first column of Table 1, and is summarized below.

Bredow’s [10] criteria for evaluating theories were organized around two major areas: internal and external criticism. Internal criticism involves a judgment about the internal components of the theory, whereas external criticism involves a judgment about the match between the theory and context of its use. When evaluating internal criticism, the evaluator assesses the adequacy (thoroughness in addressing topic), clarity (clearness of statements), consistency (congruency in semantics, etc.), logical development (support from evidence), and level of theory development. To make judgments about external criticism, the evaluator assesses the complexity (number of concepts/variables, from parsimonious to complex), discrimination (uniqueness), reality convergence (relevant assumptions), pragmatism (ability to use in the real world), scope (narrow to broad use for practice), significance (impact of theory), and utility (ability to produce hypotheses). Critique information for each of the commonly used models was summarized in a table after consensus had been reached by two (or sometimes three) authors.

Results

The disposition of the search results is shown in Figure 1 as a PRISMA flow diagram [14]. Searching the three databases with the selected keywords yielded a total of 1,602 records. Author review excluded 50 records because they were duplicates, books, dissertations, presentations, or could not be located. This left 1,552 titles to screen. Author pairs excluded 1,334 titles because they did not meet inclusion criteria. This left 218 abstracts to be screened, of which 70 did not meet inclusion criteria; 148 progressed to the full text assessment for eligibility. Of the 148 full text articles assessed, 48 were eliminated because a HRQOL model had not been derived from or used to guide the research, review, and/or findings. This process resulted in a total of 100 articles being included in this review (see Figure 1).

Of the 100 articles, 46 were quantitative. Of the remaining 54, 16 were qualitative research, 1 was mixed methods research, 15 involved instrument development, 20 were literature reviews, 1 described a model revision, and 1 was a consensus paper. The 46 quantitative studies were mainly descriptive studies (n = 31), with a few being correlational (n = 13), or randomized controlled trials (n = 2). Sample sizes ranged from 10 [15] to 69,031 participants [16]. The 100 articles came from 21 different countries including Australia (n = 4), Austria (n = 1), Brazil (n = 3), Canada (n = 12), China (n = 3), Finland (n = 1), Germany (n = 4), India (n = 1), Ireland (n = 1), Israel (n = 1), Italy (n = 2), Japan (n = 1), Netherlands (n = 7), Norway (n = 2), Spain (n = 3), Sweden (n = 2), Taiwan (n = 1), Thailand (n = 4), Ukraine (n = 1), United Kingdom (n = 10), and the United States (n = 49). Of these 100 articles, 9 involved more than one country.

Discussion and recommendations

There are two important findings from this review. First, there has been little consistency in HRQOL models within the literature of the past 10 years. Approximately three-fourths of the articles reviewed used an existing HRQOL model as a guide; however, most of these applied a variety of different models, rather than using a common model found in the literature such as the Wilson and Cleary model [12]. Thus, there were wide variations in terminology for analogous HRQOL concepts, making cross-study comparisons virtually impossible. This seriously limits the ability to have a coherent body of evidence to guide further HRQOL research and practice. Second, the most commonly used models were based on work by Wilson and Cleary [12], the revised model by Ferrans and colleagues [13,19], and the WHO [3,20,21]. A majority of the researchers using these models could be doing so because of an absence of better alternatives. However, based on our findings, we recommend that authors consider the advantages of using one of the three commonly used global models in research to more quickly advance the science in the area of HRQOL. Our findings show that Wilson and Cleary [12], as well as the revisions of Wilson and Cleary’s model proposed by Ferrans et al. [13], together are the most frequently referenced in the HRQOL literature, representing nearly a quarter of all of the articles reviewed. Ferrans and colleagues’ [13] model provides clear conceptual and operational definitions, and it also clarifies relationships among concepts to guide research and practice. The WHO ICF model [3] may be useful in specific HRQOL studies; however, it has more potential for application to studies of an epidemiological, sociological, or educational nature.

There are a great many models in the HRQOL literature that have not been adequately tested or refined. Cross-comparisons across diseases could be done if authors had at least used a global HRQOL model as a starting point. In fact, many single-use models included the same concepts as the three global HRQOL models but labeled them differently. In the future, when a common global HRQOL model is not used, authors should clearly delineate why a context- or disease-specific model is preferred. Increasing the consistency in models used across studies would help increase our understanding of this important concept.

Of the 23 articles citing the three most common HRQOL models, most articles were descriptive, correlational, or literature reviews. Importantly, future HRQOL research should involve comparisons of intervention outcomes. Only one randomized controlled trial was found that used the most commonly cited Wilson and Cleary [12] model [25]. Although disease-specific or situation-specific models may be better for testing interventions, the global models should still be useful as a template and a jumping-off point for adaptations to specific contexts. In addition, using an existing model can advance the state of the science of HRQOL by contributing new information about the applicability of the selected model to research and practice, thus leading to model refinement such as the revised model proposed by Ferrans and colleagues [13]. This underscores the need to start with the best available HRQOL models and build upon them, rather than creating new models.

Conclusion

In summary, based on this systematic review of the literature, Ferrans et al., [13] revision of Wilson and Cleary’s [12] model appears to have the greatest potential to guide HRQOL research and practice. We recommend Ferrans and colleagues’ [13] model because they added individual and environmental characteristics to the popular Wilson and Cleary [12] model to better explain HRQOL. Although the WHO ICF model has been considered a model of HRQOL, it is more of a mapping and classification framework than a guide for hypothesis generation in the area of HRQOL. Use of one model, such as Ferrans et al. [13] revised HRQOL model, will provide more opportunities for testing and refinement of the model and more evidence about which relationships among HRQOL concepts are common to different populations. Finally, and maybe most importantly, using one model will help in comparing HRQOL across studies and populations, contribute to the development of more intervention studies, and more quickly advance the science in the area of HRQOL.

Abbreviations

HRQOL: Health-related Quality of Life; QOL: Quality of Life; WHO: World Health Organization; WHO ICF: World Health Organization International Classification of Functioning, Disability, and Health.

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

TB provided overall leadership and contributed to the conception and design, participated in the review and critique process, drafted sections of the manuscript, and revised it critically for intellectual content. SMM, JSC, JMB, JLO, KMH, MLE, &; JLW contributed to the conception and design, participated in the review and critique process, drafted sections of the manuscript, and revised it critically for intellectual content. KAH acquired articles for review, abstracted findings to tables, contributed to analysis and interpretation, provided reference management, and drafted sections of the manuscript. All authors read and approved the final manuscript.

Supplementary Material

Acknowledgements

Funding for this study was provided by Indiana University School of Nursing Research Investment Funds and the Center for Enhancing Quality of Life in Chronic Illness. The authors thank Phyllis Dexter, RN, PhD, Assistant Scientist and Editor, Indiana University School of Nursing Center for Nursing Research, for her editorial assistance. The authors also wish to thank Randi L. Stocker, Research Librarian at Indiana University Purdue University at Indianapolis, for her assistance with literature searches.