Measuring therapeutic relationship in the care of patients with haemophilia: A scoping review

Abstract Objective We conducted a scoping review of the tools used to measure therapeutic relationship in patients with haemophilia. Background Haemophilia is an inherited bleeding disorder caused by a deficiency of a clotting factor in the blood. Therapeutic relationship is foundational to the management of patients with chronic diseases like haemophilia. A reliable and valid measurement tool for assessing therapeutic relationship is needed to evaluate the quality of care received by these patients, and to rigorously study the association between therapeutic relationship and the outcomes of treatment. Methods We adopted the Arksey and O'Malley framework for scoping studies. The following electronic databases were searched for studies that measured a construct related to therapeutic relationships in haemophilia care: MEDLINE, EMBASE, CINAHL, PsycINFO and Scopus. We inventoried these studies, identified the measurement tools used, and described each tool by purpose, content, measurement properties and target population. We identified gaps in the current evidence and directions for future research. Results There were 253 unique records retrieved in the search, and twenty studies were deemed relevant. Ten measurement tools were identified. None of the tools measured therapeutic relationship as a single entity; however, six tools measured constructs considered part of patient‐provider relationship (eg trust, communication, working alliance). There has been little validation testing of these tools in haemophilia patient populations. Conclusions There is a need for a validated tool for measuring therapeutic relationship in the care of patients with haemophilia. This review provides a foundation for future research in this area.


| INTRODUC TI ON
Haemophilia is an inherited bleeding disorder caused by a deficiency of a clotting factor in the blood. Patients are at a lifelong risk of bleeding into joints and muscles. Recurrent bleeding often results in chronic impairment of musculoskeletal structures and function, leading to pain and disability. 1 Prevention of this process is a priority in the improvement of health and quality of life of patients with haemophilia. This is accomplished through regular encounters and monitoring by an interdisciplinary haemophilia treatment clinic (HTC), which consists of physicians, nurses, physical therapists and social workers. 2 Successful management of haemophilia requires that patients actively participate in their care with the HTC. This purposeful partnership of patient and health-care providers from the HTC is described as the "therapeutic relationship." Therapeutic relationship has been consistently associated with treatment outcomes in health research. 3 Therapeutic relationship has been the topic of expert narrative reviews exploring the evolution of patient-provider relationships and patient autonomy, 9,10 and ideas about therapeutic relationship in contemporary haemophilia care. [11][12][13] Findings from qualitative studies suggest that patients and health-care providers consider aspects of patient-provider relationship to be a key component of haemophilia treatment. 12,[14][15][16][17] At present, there is an emphasis in haemophilia research on understanding the factors that influence patient's degree of adherence to treatment, which is important because patient adherence to treatment is linked with positive outcomes such as reduced pain and improved joint health. 18,19 Further, preliminary research suggests significant associations between patients' degree of adherence to factor replacement therapy and certain dimensions of therapeutic relationship. 20 Specifically, patients reporting a higher degree of trust in their haemophilia physician have higher rates of adherence to treatment. 20 Similarly, a good relationship with a haemophilia healthcare provider has been positively correlated with adherence levels. 21 As interest in this area of haemophilia research grows, it becomes important to establish a validated and standardized approach to measuring therapeutic relationship. A high-quality measurement tool will improve the validity of research into the processes and mechanisms by which therapeutic relationships impact outcomes, such as pain, joint health, and quality of life for patients with haemophilia. A standardized approach to measurement will also facilitate comparisons between studies of interventions aimed at improving therapeutic relationship.
Given the importance of a validated tool, and the relevance of studying therapeutic relationship in this population, we conducted a scoping review to provide a comprehensive overview of the research in the area of measurement of therapeutic relationship in the care of patients with haemophilia. Although we focus on research applications of measurement, this review also has implications for evaluating quality of care and assessing the patient's experience of care.
The objectives of our scoping study were to: 1. Locate and inventory the studies that assess therapeutic relationship in haemophilia, and describe the nature and extent of this evidence.

2.
Identify the measurement tools that were used, and examine the literature associated with each tool.

3.
Summarize the characteristics of the tools that are relevant to researchers when selecting an appropriate measure of therapeutic relationship.

4.
Identify knowledge gaps in this area and directions for future research.

| Design
We adopted the Arksey 23 Briefly, Levac et al 23 emphasize the need for an iterative and team approach to study design, K E Y W O R D S blood coagulation disorders, inherited, health-care surveys, patient outcome assessment, patient participation, patient-reported outcome measures, professional-patient relations, validation studies establishing inclusion and exclusion criteria, searching and selecting relevant articles, and identifying key variables for data extraction. We incorporated these recommendations into the methods of this study.

| Search strategy
We identified studies that were relevant to our research question through online searches of relevant health databases from their inception to April 2017. These searches were performed with the assistance of an experienced health research librarian at the University of Alberta. The following electronic data-

| Study selection
Two members of the research team independently screened the titles and abstracts of the publications retrieved in the database search. Full texts of the potentially relevant articles were acquired and appraised in reference to our study selection criteria. We included peer-reviewed articles that described the development, testing or use of a measurement tool in a research study to assess or measure therapeutic relationship or related construct, focusing on a population of patients with inherited bleeding disorders and the health-care providers (from any discipline) involved in their treatment. We included an article if it measured a subcomponent of therapeutic relationship (eg trust, empathy, communication) or a construct that may be considered to contribute to therapeutic relationship (eg patient-centredness, satisfaction with care, shared decision making). We included self-report questionnaires (patient or health-care provider perspective), observer-rated scales and coding schemes, all modes of administration (eg paper and pencil, computerized or interview). Any discrepancies between reviewers that arose during the review process were resolved through discussion. We used a kappa coefficient to quantify inter-rater reliability between reviewers.
As the reviewers became familiar with the literature, they noticed that therapeutic relationship was often conflated with other constructs related to clinical encounters and that authors often did not clearly define the construct being assessed. This made it difficult to determine the content of the measurement tools. To address the issue, we added an item content analysis step to our methods (described in the "Data analysis" section below), similar to methods used by Eveleigh et al. 24 This iterative approach to methods is an advantage of scoping study methodology for an emerging research area like therapeutic relationship, where little is known about the literature prior to starting the study. 23 A second challenge encountered during study selection related to the definition of "patient satisfaction with care." This term might refer to patient satisfaction with interpersonal aspects of care, satisfaction with the specific intervention or satisfaction with the F I G U R E 1 Flow chart of the methods used for data analysis outcomes of treatment. We addressed this challenge through discussion within the research team, which resulted in a clearer definition and common understanding to only include studies assessing satisfaction with interpersonal aspects of care.

| Charting the data
A single reviewer extracted relevant study features, which were determined based on our research question and objectives. We obtained a copy of each measurement tool that was identified.

| Data analysis
To describe the nature and extent of the evidence, we calculated descriptive statistics (frequencies and percentages) for the key characteristics of the studies included in the review. Figure 1 shows the flow of the methods of data analysis.

| Measurement properties
A useful measurement tool should meet two standards of comprehensiveness. 25 First, a tool should be accurate and precise through the full range of the variable being measured (eg from poor to strong therapeutic relationships) within the target patient population. It is therefore important to examine the evidence concerning the tool's measurement properties, that is, reliability and validity, in the context of the intended target population. 25 Second, the content of the tool should adequately represent all the multiple dimensions or components of a health construct. 25 Therefore, we conducted a second search of the literature to find all published work associated with each measurement tool identified. We searched reference lists, MEDLINE, and the search engine Google, using the name of the tool, any known synonym and abbreviations. We extracted information related to the development and testing of the tool, the measurement properties reported and the theoretical basis of the tools from the articles retrieved in the second search. We examined the extent of the validity evidence for each of the measurement tools identified. We used the COSMIN 26 (COnsensus-based   Standards for the selection of health status Measurement Instruments) taxonomy and definitions for measurement properties as a guide. We summarized the characteristics of the tools in table form.

| Content analysis
The content of the tools was analysed using the framework of therapeutic relationship in physiotherapy developed by Miciak. 27 Therapeutic relationship has not been conceptualized in the haemophilia literature, and Miciak's framework has qualities that made it appealing for use in this study. The framework was developed using rigorous qualitative methods, it is comprehensive in scope, and it is sufficiently detailed to provide a clear understanding of the fundamental components of the therapeutic relationship. 27 The three components of therapeutic relationship are as fol-  27 We used the themes and the components in the therapeutic relationship framework to describe the content of the tools and to systematically distinguish the tools that primarily measure a component or subcomponent of therapeutic relationship. We termed these "relational tools," which we operationally define as a measurement tool that assesses attitudes, intentions, behaviours or feelings between a health-care provider and a patient. A general patient satisfaction questionnaire is not a "relational tool" as we have defined it. Although it may contain a small proportion of items that address patient-provider relationship, patient satisfaction questionnaires also typically assess organizational-or system-level health services and processes. We examined the content of a tool using the items as the unit of analysis. We coded each item in reference to the component of therapeutic relationship that it measured (if any). Items that did not fit the therapeutic relationship framework were coded as either "satisfaction with care" or "not interpersonal." Examples of the item appraisal are included as Appendix 2. For each tool, we calculated the proportion of items in each category (ie relationship, satisfaction or not interpersonal). We distinguished the relational tools based on the proportion of items that measured therapeutic relationship. Finally, we checked whether the tool addressed each of the three themes Miciak identified in therapeutic relationship (personal and professional, body as central and mutuality). Appendix 3 contains the findings of the content analysis. One member of the research team conducted item analysis, and a second member reviewed the results, with any discrepancies resolved through discussion. We summarized the content, function and validity evidence of each relational tool to provide a comprehensive overview of the relational tools used in haemophilia for researchers selecting a measurement tool.

| Patient and public involvement and engagement
The aim of patient and public involvement and engagement (PPIE) in this study was to plan, conduct and interpret findings of the research in a manner that was meaningful to patients and their health-care providers. One patient partner was involved throughout the study as a member of the study team (JH). He is a person with haemophilia and a Master's student at the University of Alberta . He helped design the study, refine the research question and scope, interpret F I G U R E 3 Flow chart of the article search and selection stages results and critically review written reports. This was accomplished through meetings with the lead researcher, electronic communications and informal conversations at related scientific gatherings.
Health-care providers were also consulted during project planning and after the literature search.

| RE SULTS
The search and selection process is summarized in Figure 3. The initial search of electronic databases returned 416 records. After 163 duplicates records were removed, two reviewers screened 253 titles and abstracts for potential inclusion. Inter-rater reliability between reviewers was high in the screening process (Kappa = 0.81). Forty-nine articles were retrieved for full-text appraisal. Thirty articles did not fit the criteria for inclusion. One of the articles was a systematic review, which was excluded from further analysis after a search of its reference list for relevant publications. Subsequent to the search, one article was located through the professional networks of the research team. Twenty articles were selected for inclusion, and inter-rater reliability was good (Kappa = 0.76).

| General description of the included studies
The main characteristics of the included articles are summarized in Table 1 With regard to the relationship construct measured in each study, there were no articles that measured therapeutic relationship as a single entity. One study assessed working alliance-a concept originating in the psychotherapy literature. 30 Seven studies assessed socio-emotional elements in therapeutic relationship, for instance, patient trust in the physician, empowerment, collaboration or provider receptiveness. Task-focused communication-communicative "acts" of the patient or provider-was assessed in four studies. Eight studies evaluated patient satisfaction with health services (n = 8). Six of these articles assessed satisfaction with the services of a HTC, and two assessed satisfaction with other health services (genetic testing, pain therapy). Table 2 contains an inventory of the studies included in this review.
The aims of the studies were grouped into three categories: Six studies were assigned to category 1, and these explored the associations between patient and provider characteristics, environmental factors and outcomes of treatment. The four studies in category 2 sought to evaluate an intervention, for example, a new application of a technology or service delivery model. Finally, the four studies in category three aimed to describe health services for patients with haemophilia. The remaining six studies aimed to develop a measurement tool. We identified shared decision making and adherence to treatment as two topics that were frequently studied relative to subcomponents of therapeutic relationship.
Five studies were conducted for the purpose of understanding elements of shared decision making. Four studies were aimed at understanding the factors related to adherence to treatment in haemophilia.

| Description of measurement tools
Thirteen of the twenty articles described a standardized measurement tool. Ten unique tools were identified: the "Specialist/ Nurses" subscale of Hemo-SAT, 31  An additional 27 articles were found that reported on a tool's development or testing. Within the associated literature, we found evidence for all tools regarding content validity and interpretability.
Additionally, we found that internal consistency (coefficient alpha) had been reported for all the self-report questionnaires. The measurement properties of six tools were tested in a haemophilia patient population. Key characteristics of the tools and their associated literature are described in Table 3.

| Content comparison of the tools
We did not identify any tools that comprehensively measured the full scope of therapeutic relationship. Based on item content analysis, we distinguished six tools that measure a relationship construct as a primary domain: WAI-CC, WFTPS, QUOTE, Veritas-PRO, TMDD and UOvS. Three of the four other tools measured satisfaction with care.
The results of our item content analysis showed the WAI-CC most comprehensively covers the components of therapeutic relationship framework, with 9 of 11 subcomponents represented, missing the subcomponents of "body as a pivot point" and "present." The scope of the UOvS content was broad as well, capturing 7 of the 11 subcomponents. The WFTPS measured elements of the bond (trust, caring) and the conditions of engagement (receptive, genuine, committed). The items in the Veritas-PRO, TMDD and QUOTE tools measured subcomponents of ways of establishing connections.
In terms of the three themes of therapeutic relationship, five of the six tools addressed the relationship as a mutual endeavour, and four of six tools addressed the body is central theme. A single tool attended to the personal aspect of therapeutic relationship (UOvS), while all tools examined professional aspects of therapeutic relationship. We compare the six relational tools in terms of functionality, content and measurement properties in Table 4. The conception and design of the study and the scope of our research question were guided by informal discussions with healthcare providers working in HTCs. In addition, a peer-review panel consisting mainly of clinicians from HTCs reviewed the project at the proposal stage, and we incorporated their feedback into the project design.

| D ISCUSS I ON
The purpose of this study was to provide an overview of the measurement of therapeutic relationship in the care of patients with haemophilia. We did not find any studies that measured the full scope of therapeutic relationship. From this, we concluded that no tool for the measurement of therapeutic relationship has been validated in  This tool was not named in the studies; therefore, in this paper we have referred to it by the institution at which it was developed.
TA B L E 2 (Continued) therefore, further validation of these tools will be required to ensure the results from studies using these tools are valid.
We identified six tools that measure constructs that are part of therapeutic relationship. The features of each tool must be . 48 It has shown good measurement properties in various outpatient settings. 49 The Healing Encounters and Attitudes Lists (HEAL) is a 57-item measure of the "patient-provider connection". 50 The HEAL measure has the advantage of being developed using item response theory, which offers greater flexibility and efficiency of measurement. 50

| Strengths and limitations of this study
We presented a robust overview of research and measurement tools and situate measurement of therapeutic relationship within the broader context of health service research in haemophilia. Also, Warren-Findlow, 79 Fassaert, 80 Briggs, 81  Muller, 87 Hall, 88 Bachinger, 89  Besley, 45 Hall, 43 Burns, 44 Morris, 46 Horvath and Greenberg 60 This tool was adapted from the "Working Alliance Inventory" in collaboration with its we identified the knowledge gaps and directions for future research.

TA B L E 3 Description of the measurement tools identified from studies involving patients with haemophilia
Some key strengths of our study are that we used a systematic and reproducible search and selection strategy, and we assembled a research team with content and methodological expertise. Further, we clearly reported our approach to data analysis using a robust theoretical framework of therapeutic relationship.
There were two main advantages to using the framework. First, it added structure and transparency to the analysis of the tools' content. The framework was justified given the rigorous methods with which the framework was developed and that therapeutic relationship has not been conceptualized in the care of patients with haemophilia. Second, the framework helped identify a clear distinction between patient-reported relationship scales and patient satisfaction scales. In an effort to include all available evidence of evaluation of therapeutic relationship, we included patient satisfaction with care as a measurement construct in this scoping study. It was important that we used a method that could distinguish the two constructs, because the use of patient satisfaction questionnaires to evaluate the quality of therapeutic relationships is generally not supported. 52 In part, this is because general satisfaction questionnaires often fail to include items that assess emotional constructs in the proportions that reflect patients' true priorities in their care. 52 A potential limitation of the study is that the framework of therapeutic relationship was developed in the context of physiotherapy for patients with musculoskeletal impairments, and the generalizability of the framework from physiotherapy to other health-care disciplines has not been established. Physiotherapists typically focus on the body and physical condition, and parts of the framework might be more pertinent to physiotherapists (eg "body as pivot point"). However, haemophilia is a haematologic condition that often manifests in the musculoskeletal system. During clinical encounters, health-care providers from all disciplines will be concerned with asking about physical symptoms, addressing issues related to the physical condition (eg experience of pain, joint bleeding), and how the patient experiences and is impacted by these physical problems.
Therefore, the therapeutic relationship framework used is likely relevant to the care of patients with haemophilia by health-care providers from all disciplines. Furthermore, the framework converges with the therapeutic relationship literature in haemophilia.
Qualitative studies in haemophilia addressing a patient-centred care model 12  Another potential limitation is the method of appraisal of the content of the items. The process involved the subjective judgement of the researchers, and it is possible that items in each measure would be classified differently by a different set of researchers.
In addition, the choice of therapeutic relationship framework could impact the results of the content analysis of the measurement tools. Therapeutic relationship is a complex construct that can be conceptualized and organized differently, thereby impacting the classification of tools as relational. For instance, some frameworks are focused on concepts such as bonding, 53 empathy, 54 trust, 55 or communication, 56 and working alliance, 30 while others are more broad, including contextual factors such as the health-care environment, 57 patient or health-care provider factors such as the prerequisite knowledge and qualities of the health-care provider, or patient expectations for care. 58,59 Despite these limitations, the results of the content analysis suggest that our method was suitable as there was a clear delineation between the tools classified as relational (proportion of relational items was 0.84 and above) and nonrelational tools (0.38 or lower). A final limitation is that one single researcher conducted the data extraction and content analysis steps; however, these were verified by another researcher.

| Reflections on PPIE
The degree of PPIE in health research can range from a consultationtype involvement to research that is completely led by the public. We

| CON CLUS IONS
In this scoping review, we sought to answer the question: "What validated measurement tool(s) exist for measuring the therapeutic relationship in the care of patients with haemophilia?" We did not find any measurement tools that have sufficient validity evidence to be used to measure therapeutic relationship in haemophilia care.
We identified six tools that were used to measure aspects of therapeutic relationship, but were not comprehensive in scope. There is a need for a conceptually sound measurement tool of the therapeutic relationship to be validated in the care of patients with haemophilia.

ACK N OWLED G EM ENTS
Thank-you to Maria Tan, health research librarian, for her help conducting the search.

CO N FLI C T O F I NTE R E S T
The authors have no conflict of interests to declare.