A uniform framework for assessing the benefits of health care services: development of a checklist to identify the attributes of relevance for use in discrete choice experiments

Presenter: Linda Spijkers, Maastricht University Medical Center

Abstract

Introduction: Discrete choice experiments (DCE) are increasingly used to investigate the benefits of health care services. The theoretically well-founded technique is based on the premise that any good can be described by its attributes, and that an individual’s valuation of a good depends on the nature and levels of these attributes. In order to achieve external validity, the first step in a dce is to identify the attributes of importance to the choice between the services. The attributes relevant in health care services are not limited to health outcomes but may also include characteristics extending health. However, this crucial first step in the design of a DCE tends to receive only limited attention. In each single DCE the attributes of relevance are chosen, while a general accepted definition of what constitutes the benefit of health care services is lacking. As a result, the generalizability of findings from DCEs may be limited.Objectives: This paper describes a uniform framework for assessing the benefits of health care services, and the development of a checklist to systematically identify the attributes of relevance for use in DCEs.Methods: The classical framework for the quality of care (consisting of the components 'structure', 'process' and 'outcome') was used as starting point to develop a checklist to systematically select the attributes of relevance to assess the benefit of health care services. The components of the framework were divided into domains: 'structure' was divided into 'physical characteristics' and 'staff characteristics', 'process' into 'clinical care' and 'interpersonal care' and 'outcome' into 'health status'. This framework describes all possible (categories of) attributes.To create an overview of the attributes used in DCEs, we systematically searched the literature of the past five years and included papers that used a DCE to determine the value of a health care service. From each paper we extracted the attributes (and levels) and categorized these into the domains of the framework. Within these domains attributes were further categorized into particular topics and for each topic a specific question was formulated. Based on these questions a checklist was developed.Results:The search resulted in 323 papers of which 76 were included. In these studies, a total of 432 attributes were identified. In the component 'structure' 114 attributes (26.4%) were identified, consisting of 102 attributes in the domain 'physical characteristics' and 12 in the domain 'staff characteristics'. In the component 'process', 149 attributes (34.5%) were identified: 111 in the domain 'clinical care' and 38 in the domain 'interpersonal care'. In the component 'outcome' 169 'health status' attributes (39.1%) were identified. A checklist was developed consisting of questions relevant to the identification of attributes of importance.Conclusion: Using the classical framework for quality of care as starting point, all attributes from the 76 studies could be categorized. The resulting checklist is useful to systematically identify attributes that are relevant for use in a DCE in health care, thereby increasing external validity and generalizability of results from DCEs.