Introduction

In the last decade, treatment strategies for rheumatoid arthritis (RA) have evolved from the old “pyramid” approach to the early, proactive, and consistent use of disease-modifying antirheumatic drugs (DMARDs) [1–6]. To minimize the delays in diagnosis, treatment initiation, and monitoring, a number of health service models, including the use of both pharmacological and nonpharmacological treatment modalities, have been subsequently developed and evaluated [7].

A health service model may be labeled according to the setting in which the treatment modalities are implemented (e.g., inpatient team [8]), the leadership (e.g., the clinical nurse specialist model [9] and the physical therapist practitioner model [10]), or the mode of delivery (e.g., telehealth [11]). However, many labels have multiple meanings, which create communication challenges among health care professionals, researchers, and other decision makers. For example, models involving the use of communication equipment are commonly called “telehealth” and yet this term may refer to models with different modes of delivery including rheumatologist consultation provided through videoconferencing with the patient at the family physician’s office [11–13], through Internet and e-mails [14], or through telephone follow-up [15]. In addition, the lack of relevant detail regarding the health services model such as the objective(s) of the model, provider(s) of the service, and process of service delivery creates additional barriers when making decisions about patient care or implementing specific health service models into clinical practice.

Current standards for reporting clinical trials [16], including the most recent recommendations on the reporting of nonpharmacological interventions [17], focus on the inclusion of items that convey information about the study design, analysis, results, and interpretations of the findings. However, they provide little detail on how to systematically describe the features of complex health service interventions. This information would provide the reader with the ability to make informed judgments regarding the suitability of the model for their own patient population and to implement it in their practice setting.

The purpose of this study was to develop a reporting framework for health service models in RA care. To begin, we summarized how health service models in RA management were described through a systematic review of primary studies. We focused on RA instead of other types of arthritis because most health service models were developed for this population [7]. The essential components for reporting health service models were then identified and used to develop a framework for research reporting. We envision that this framework will facilitate knowledge exchange between researchers and research users.

Materials and methods

The research team consisted of individuals with expertise in clinical epidemiology (SO, LL, and TVV), research on nonpharmacological interventions in rheumatology (LL and TVV), and qualitative research (JK). In addition, one member was a health care administrator (CL) and one was a full-time physical therapist (HF). We conducted a search of the health sciences literature from 1990 to December 2005 in CINAHL, the Cochrane Library, EMBASE, Healthstar, Medline, PEDro, PsycINFO, and Social Work Abstracts. The search strategy was developed in Medline by an experienced librarian in collaboration with the research team and modified as required for the other databases (Appendix A). To ensure that the literature search and review strategy were comprehensive, we sought input via online discussions with health service researchers, clinicians, health care administrators, and patient representatives who participated in the 2005 CARE III Conference [18].

Eligibility criteria and article selection

Primary studies that evaluated a health service delivery model for managing RA, published in either English or French, were included in this review. We defined a health service delivery model for the management of RA as an approach to delivering health services/interventions which may include services provided by nonhealth care professionals and/or health care professionals to adults with RA across the disease continuum. Examples include multidisciplinary team care, use of allied health professionals in advanced practice roles, telemedicine, patient-initiated care, and targeted care (e.g., foot or hand outpatient program). Studies involving a single treatment (e.g., a surgical procedure, a medication, or a physical therapy intervention) or a study population without RA, review articles, and editorials were excluded. Two team members (HF and SO) applied the eligibility criteria on the first 100 titles from the literature search and achieved good agreement (kappa = 0.71).

The selection of articles involved a two-step process. First, two of three team members (CL, HF, and SO) independently evaluated all retrieved studies using the bibliographic record (i.e., title, authors, keywords, abstract). Potentially relevant records or those that did not contain sufficient information to determine eligibility received a review of the full article from two of the same three members. Disagreements were resolved by consensus and reasons for exclusion were noted.

Data extraction

A data extraction form was developed by the research team to extract the components of health services from the literature. These components were based on our collective knowledge and experience on health service models in RA management. The data extraction form was pilot tested by the team using a randomly selected set of five eligible papers until consensus was reached and no further modifications were required. All team members participated in the review, with the information extracted by a primary reviewer and verified by a second reviewer. Disagreements were resolved by consensus. Data extracted included the name of the model, setting, health professionals involved, modes and frequency of communication among health professionals, coordinator of the health services, initiator of the referral, the process of health service delivery, intervention(s) provided/received, frequency of health professional visits, length of the intervention, level of care provided (e.g., community, primary, secondary, or tertiary care) [19] and stage of disease according to the American College of Rheumatology (ACR) criteria for progression [20].

Development of reporting framework

Findings from the systematic review were used to inform the development of the reporting framework. The components that the team deemed important when describing the features of health service interventions for RA management (e.g., setting, health care professional(s) involved, etc.) were retained and expanded upon for the framework.

The framework was comprised of six fundamental dimensions relating to health service delivery models: Why?, Who?, What?, When?, Where?, and How? Within the “How?” dimension, multiple questions were addressed. Information under each question were classified as a component (i.e., mutually exclusive categories consisting of information that were vital to describing a health service model) or subcomponent (i.e., optional information). The research team, using a nominal group approach, carried out this process through independent reflection as well as five face-to-face meetings, with one team member that had experience with this approach (JK) attending as a facilitator [21]. Additional literature nominated by the team were used to facilitate the selection of components and subcomponents [7, 22, 23].

The nominal group process began with a discussion on the components of health service delivery models to include from the literature review in the reporting framework. Next, the team developed an overall structure for the reporting framework. This was followed by independent reflection on how best to classify the information into components and subcomponents as previously defined. Responses were collated and consensus was sought during the five facilitated face-to-face meetings. The document was then revised by JK and reviewed by the team for final editing.

Results

Of the 2,033 citations retrieved and screened, 78 papers (68 primary studies [91% in English] and ten companion papers) met the eligibility criteria (Fig. 1) and 93 descriptions of health service models were identified (Appendix B). While the vast majority of models (95.7%) reported the health professionals involved, the mode and frequency of communications between health professionals and patients were only described in 43.0% and 36.6% of the models, respectively (Table 1). The majority of models also reported the setting where services were provided (89.2%), the level of care (89.2%), and the coordinator of the overall health service delivery (61.3%). Around half of the models described the length of the intervention (59.1%), the initiator of the referral (54.8%), the services provided to and received by patients (53.8%), and the disease duration (51.6%). However, less than one third reported the frequency of health professional visits, only 8.6% stated the ACR classification of global functioning of the study participants, and none of the models reported the stage of disease according to the ACR criteria for progression.

Fig. 1
figure 1

Literature search strategy

Full size image
Table 1 Attributes of health service delivery models for the management of RA reported in the literature
Full size table

Reporting framework for health service models

To synthesize the information from the literature review on the components deemed important when describing health services interventions for RA management, the six questions about health service models were further modified as follows:

  1. 1.

    Why was the health service delivery model founded?

  2. 2.

    Who was involved?

  3. 3.

    What were the roles of the individuals participating?

  4. 4.

    When were the health services/interventions provided and/or received?

  5. 5.

    Where were the health services/interventions provided and/or received?

  6. 6.

    How were the services/interventions accessed? How are the services/interventions implemented? How long was the intervention? How did the individuals involved communicate? How was the health service delivery model supported and sustained?

The nominal group process indentified and defined were components and subcomponents for each set of questions (Table 2). To illustrate the application of the reporting framework, examples are provided using five papers on health service models published between 2006 and 2007 (Table 3) [24–28].

Table 2 Framework for reporting health service delivery models for managing rheumatoid arthritis
Full size table
Table 3 Application of the reporting framework for health service delivery models in the management of RA using primary studies from the literature
Full size table

The “Why” question addresses the model’s goals, which may include improving patient outcomes (e.g., reducing pain, reducing work disability, improving quality of life, etc.) or improving the process of service delivery (e.g., reducing delay for diagnosis, improving the coordination of the health service delivery, minimizing duplication of services, or reducing healthcare costs).

The “Who” question focuses on the reporting of the providers and users of the health service. Providers may include health care professionals (e.g., registered nurses, primary care physicians, specialists, and allied health professionals) and/or nonhealth care professionals (e.g., informal caregivers, healthcare administrators, and job counselors). User of health services may include the patient and patient’s family and/or friends.

The “What” question addresses the roles of the provider(s) and user(s) involved in the health service delivery model. Providers’ roles may include being the champion for the model, providing a specific service (e.g., screening, triaging cases, providing education), or coordinating the overall delivery of the health services. For users of health services, roles may include initiating a clinical visit, participating in treatment goal setting, or practicing self-management techniques.

The “When” question addresses the reporting of the disease duration, i.e., time from onset of symptoms and/or diagnosis.

The “Where” question focuses on to the location where the health services were provided or received. The components include the country of origin, the setting (e.g., inpatient hospital program, outpatient hospital program, community clinic, patient’s home, etc.) and the level of care. The potential categories for levels of care that authors can use include the following: community care, primary care, secondary care, follow-up, orthopedic surgical consultation, or preorthopedic/postorthopedic surgery. Community care involves services and/or information received between the onset of the patient’s symptoms and their first visit to the family physician. Primary care includes services provided by the family physician and/or other primary care team members. Secondary care services are provided by rheumatologists or other specially trained health care professionals, such as rheumatology nurse practitioners. Follow-up services are typically provided to monitor disease activity and adverse events due to medications. An orthopedic consultation is provided by an orthopedic surgeon prior to the patient’s surgical intervention. Lastly, preorthopedic and postorthopedic surgery services are provided prior to and following the surgical intervention in order to prepare the patient for surgery and monitor their recovery and progress [7].

Lastly, the “How” question addresses five topics, including (1) the referral process, (2) the process through which the interventions are implemented, (3) the duration of the interventions, (4) the mode of communication of the individuals involved, and (5) resources for supporting and sustaining the health service model.

The referral process may include self-referral, referral from a health care professional/case manager, or recruitment via a research study. The methods in which the intervention can be implemented include the use of a care pathway or treatment guidelines, the provision of individualized care, and/or by simply following the study protocol. Information regarding how individuals communicate in the health service delivery model can be captured by a description of the modes of communication, for instance, team conferences, individual face-to-face meetings, telephone, e-mail, etc. Lastly, the resources used to support/sustain the health service delivery model are captured from a financial perspective (e.g., public and/or private funding) and a human perspective (e.g., champion or founder of the model of care).

Discussion

This review has found several trends in the reporting of health service delivery models in the management of RA within the peer-reviewed literature. Although the vast majority of models in RA management reported on the individuals involved in providing the service, the setting, and the level of care, only half provided information about the length of the intervention or the process of health service delivery. Results also showed shortcomings in the description about the patient population, with less than 10% having reported on the functional level and the stage of disease progression and only half reported on the disease duration. Overall, the quality of reporting the features of health service interventions in the management of RA is poor.

We argue that the discrepancies in reporting may have contributed to the slow progress in the development, implementation, and evaluation of effective models in RA management. The first study about the team care model was published more than 40 years ago [29]; since then, few articles and systematic reviews on health service delivery models for RA have been published [7, 30]. Some models, including team care and nurse-led clinics, have demonstrated effectiveness in improving patient outcomes. However, it remains a challenge to identify what makes a model work and how to implement an effective model in specific settings.

More than 15 years ago, Yelin pointed out that the active ingredients of team care could be the result of the formal elements such as the health professionals involved and their interaction at structured meetings, the informal elements such as how people communicate, or both [31]. This had led to the question, “What’s inside the team care box?” [31]. Researchers have been encourage to clearly define and test different components of team care and other models in order to identify what contributes to an effective model. But little has been changed in the reporting of health service models. Since Yelin published his editorial, health care professionals and other decision makers still experience difficulties in sorting out the elements that make a model work, the requirement for implementing a model locally, and the resources needed for sustaining a model. This issue, noted at the 2005 Summit on Standards for Arthritis Prevention and Care meeting, has contributed to the inability to propose a definite standard on health service delivery models for Canada [32].

From the experience of other initiatives that aim to improve research reporting, we anticipate that our reporting framework will help facilitate the development, implementation, and evaluation of health service models in the clinical setting. Prior to the mid-1990s, the quality of clinical trial reporting was poor [33] and this led to the development of the Consolidated Standards of Reporting Trials (CONSORT) statement to provide standards for research reporting [16, 17, 34, 35]. A recent extension of the CONSORT statement to randomized trials on nonpharmacological treatment provide guidelines on the reporting of participants (i.e., eligibility criteria for the centers and those performing the interventions), interventions (i.e., precise details of both the experimental treatment and comparator), and corresponding examples of good reporting. There is now increasing evidence that the CONSORT statement may have contributed to improve reporting in some fields [36–38]; however, it is not sufficient to serve as a standard for reporting all the attributes of a health service delivery intervention for the management of RA. Furthermore, although theoretical evaluation frameworks for programs or interventions are available, a majority of them are general frameworks and do not address specific aspects related to complex interventions in managing chronic diseases [39]. For example, while the UK Medical Research Council published a framework that outlines the appropriate methods to adopt when developing or evaluating complex interventions, the description of the intervention is not highlighted [40]. This stresses the need for a reporting framework that addresses the description of the health service delivery intervention itself.

The role of the patient is vital in the way a health care organization that provides services (versus tangible products) is managed and evaluated. For this reason, the inclusion of the patient’s role is a definite strength in the proposed reporting framework [41]. However, despite the rigorous process used to develop this framework, there are three limitations with regards to this study. First, there was no theoretical framework or reporting format for health service interventions available at the time we developed our data extraction form. Since, general health service research reporting paper has been published by Glasziou et al. which provides guidelines on how to describe nonpharmacological treatments in research studies and hence complements our reporting framework [42]. Second, the framework was developed using results from a systematic review of the literature and a nominal group technique involving six research team members; therefore, the components and subcomponents might not reflect all the relevant information needed by a decision maker when considering a health service model. Third, the included studies were published between 1990 and 2005. Although several articles on RA models were published after this time period, we believe that they have little effect on the general trend of our findings because, to our knowledge, there were no major advances in the reporting methodology in this field between 2005 and 2008.

In conclusion, there is room for improvement within the arthritis field in describing health service delivery models in the peer-reviewed literature. The proposed reporting framework offers a practical solution by providing guidance to researchers to prepare research reports. Further evaluation is required to ensure that the framework covers all relevant information from the perspectives of health care professionals and other decision makers, can be easily applied, and improves research reporting. The proposed framework has the potential to facilitate the improved reporting of health service interventions in other chronic conditions and we encourage its adoption and modification for the reporting of health service interventions in other fields.