Questionnaires that measure the quality of relationships between patients and primary care providers: a systematic review

Background International guidance on models of care stress the importance of good quality, continuous patient-provider relationships to support high quality and efficient care and hospital avoidance. However, assessing the quality of patient-provider relationships is challenging due to its experiential nature. The aim of this study was to undertake a systematic review to identify questionnaires previously developed or used to assess the quality of continuous relationships between patients and their provider in primary care. Methods MEDLINE, PubMed, Cumulative Index of Nursing and Allied Health Literature (CINAHL) and SCOPUS databases were searched for English language studies published between 2009 and 2017. Key terms used identified studies conducted in the primary care setting examining relationships between patients and providers. Studies that focused on the conceptualisation, development, testing or review of a questionnaire, or studies that used a questionnaire for assessing the quality of continuous relationships between patients and providers were eligible. Studies that did not assess quality via a questionnaire, only assessed single aspects of relationships, only assessed single encounters, assessed transitions between settings or assessed relationships using an index were excluded. Information on validity testing of each relevant questionnaire identified from articles was reviewed to inform recommendations for future research and evaluation. Results Twenty-seven studies met the eligibility criteria, including 14 unique questionnaires. The questionnaires were diverse in length, scope, focus and level of validity testing. Five questionnaires were considered not feasible for future use due to size and lack of development work. Three questionnaires were considered strongest candidates for use in future work based on being relevant to the topic and primary care setting, freely available in English and not needing additional pilot work prior to use. These three questionnaires were the Care Continuity Across Levels of Care Scale, the Nijmegan Continuity Questionnaire and the Patient-Doctor Depth of Relationship Tool. Conclusions This study provides an overview of 14 unique questionnaires that have been used to assess the quality of continuous relationships between patients and primary care providers. The decision to use one of the questionnaires in future work requires careful consideration, including the scope, length, validation testing, accessibility of the questionnaires and their alignment with the initiative being evaluated.


Background
The focus of primary care is changing in many countries, with the aging of populations and growing need for continuity rather than episodic care. Structural reforms such as the 'Patient-Centered Medical Home' (North America) and 'Health Care Home' (Australia & New Zealand) are helping services move away from transaction-based care towards care that is patient-centred and continuous [1,2]. This change in focus emphasises the importance of all patients having a high-quality relationship with a primary care provider that continues over time [2]. 'Continuity of Care' has seemingly been associated with improved clinical outcomes, but the critical elements in play remain undescribed. Concepts such as therapeutic alliance [3], working alliance [4], continuity of care [5], relational continuity [6] and relationship-based care [7] describe the positive outcomes that occur when a patient has a sense of affiliation, collaboration and trust with a single provider that is ongoing in nature [8]. These high-quality relationships have been shown to result in positive patient experiences, greater patient satisfaction, increased treatment adherence and improved patient outcomes [3,4,9]. Supporting these continuous, high-quality relationships is clearly warranted.
Assessing the quality of relationship between patients and providers is challenging due to its experiential nature. There is no universal agreement about the definition of quality relationships or the components that underpin the concept, making it challenging to develop valid and reliable assessment tools (questionnaires). Furthermore, the quality of relationships between patients and providers is thought to be influenced by demographic factors of the patient and provider, role of medical receptionists and other staff, and organisational factors of general practice clinics [5]. It is therefore not surprising that quality of relationship is one of the least commonly evaluated aspects of care and there is no recommendation on how to evaluate relationship quality within the reforms happening to general practice [10,11].
A systematic review has previously been conducted to identify questionnaires that can be used to assess the quality of relationships between patients and doctors across all health care settings [12]. The search was conducted in 2009 and nineteen tools were identified, with variable levels of validity testing to support their development. The review methodology provided a wide reach of measures to consider, but none of the questionnaires were developed for use in the primary care setting where the majority of patients and families experience ongoing care. As a result, there is still no best approach recommended for primary care and the feasibility of these reviewed measures, whilst important, is unknown.
The aim of this study was to conduct a systematic review of the body of evidence for studies that measure the quality of continuous relationships between patients and primary care providers. The review will identify questionnaires developed or used since the previous systematic review [12] and will also appraise the questionnaires on their validity and feasibility for use in the primary care setting. The review will inform evaluation strategies for health care homes.

Methods
Overview A systematic review was conducted to identify measures of continuous quality relationships between patients and providers in primary health care. For the purpose of the review, 'relationships' referred to an ongoing sense of affiliation and collaboration with a provider in primary care, typically a General Practitioner (GP) [8]. The systematic review was conducted in accordance with the Preferred Reporting Items for Systematic reviews and Meta Analyses (PRISMA) statement [13].

Literature search
A systematic computer-based literature search was conducted between March and June 2017. Databases searched were MEDLINE, PubMed, Cumulative Index of Nursing and Allied Health Literature (CINAHL) and SCOPUS. Medical subject headings (MeSH), were used in the execution of PubMed and MEDLINE database searches. Boolean connectors AND and OR were used to combine search terms. Three categories of search terms were used; (i) terms relating to the setting: 'primary care' , 'primary health care' and 'general practice' , (ii) terms relating to relationships: 'relational continuity' , 'continuity of care' , 'physician-patient relations' , 'professional-patient relations' , 'therapeutic alliance' , 'patient participation' , and 'patient empowerment' , and (iii) terms relating to the methodological focus of the study: 'tool' , 'instrument' , 'scale' , 'survey' , 'questionnaire' and 'measure'. Google Scholar and PUBMED were used to obtain additional articles identified by journal hand searching. All database search results were imported into EndNote and duplicates removed prior to screening.

Eligibility criteria
Studies were included in the review if: 1) they focused on the conceptualisation, development, testing or review of a questionnaire for measuring the quality of continuous relationships between patients and a primary care provider; or 2) they used a questionnaire for assessing the quality of continuous relationships between patients and a primary care provider. Studies were also included if the authors' interpretation of "relationships" related to patients having a sense of affiliation, collaboration and trust with a single provider that is ongoing in nature, including phrases such as therapeutic alliance, working alliance, continuity of care and relational continuity. All study designs were considered relevant, including observational, descriptive, intervention and theoretical methodologies. Studies needed to be available in full-text, English and published between the years 2009-2017. This time period was chosen because the literature search in the previous systematic review related to this topic occurred in 2009 [12]. The focus on health care homes as an approach to primary health care reform has also occurred since this time [1].
Studies were excluded if they described the importance of high quality relationships without measuring or assessing these (i.e. via a questionnaire). All remaining studies that utilised a questionnaire were excluded if the questionnaire assessed: (1) single encounters only (rather than continuous care), (2) single aspects of relationships (such as communication), (3) transitions between health care settings (such as attending primary care after hospital discharge; informational continuity) or (4) assessed the quality of relationships between providers in a multidisciplinary team. Finally, studies that used a formulaic index to assess relationships (such as the number of different providers seen in a year) were also excluded due to the inability to assess the quality of relationships using this approach.

Study selection
The study selection process is illustrated in Fig. 1. A quality control training procedure was conducted to ensure consistency of coding between reviewers. Three reviewers independently read the abstracts of the first 100 articles identified in the search and coded them as 'retrieve full text' if the article met the inclusion criteria; 'exclude' if the article did not meet the eligibility criteria or 'unsure' if the reviewer was not able to make a decision. Agreement between all reviewers was obtained for 62/100 abstracts (62%), and at least one reviewer coded 'unsure' for the remainder of articles. Where the coding differed, consensus was achieved through group discussion. Another 50 abstracts were then reviewed and coded independently, with agreement for 46/50 (92%) abstracts obtained. Following another group discussion, the remaining abstracts were divided between the three reviewers for independent, duplicate coding.
Full manuscripts were retrieved for those studies coded by two reviewers to meet the inclusion criteria or where more information was required in order to make a decision. Disagreements between duplicate reviewers were considered by the third reviewer and resolved via group discussion. Reference lists from all systematic review articles retrieved but not included were cross-checked to identify

Data extraction
Data from all included articles were extracted using an electronic spread sheet developed specifically for this review. Information extracted included authorship team; year of publication; country; stated aim; participant characteristics (age, sex and relevant health conditions); phrases used to indicate its relevance to the review topic; and relevant tools used in the study. For each questionnaire, information about the name, author, number of items, answer format, target respondents and validation activities were extracted into a separate electronic spread sheet. Hand searching was conducted on each questionnaire to identify information not provided in the reviewed article, including validation work.

Risk of bias and data analysis
Quality assessment of included articles was not undertaken for this review as it does not draw conclusions from the findings of the articles. Rather, information on validity testing of questionnaires was extracted, covering internal consistency, construct validity, test-retest reliability, face validity, and test discriminate validity. This information was analysed by two reviewers using an iterative process of data extraction, discussions and contacting authors of questionnaires where required. Decisions about the appropriateness of questionnaires to primary health care were made in group meetings after considering the focus, length, validation and availability.

Results
The initial database search identified 3629 articles for screening as outlined in Fig. 1. Within this group, four systematic reviews were screened and although none met the inclusion criteria, their reference lists identified an additional three articles for consideration. After removing duplicate copies of articles, the main reasons for excluding articles were: the topic not being relevant (n = 2015); the study using a formulaic index to assess relationships (n = 21); the study examining single encounters only (n = 6), the study examining single aspects of relationships such as communication (n = 3); or the study not being available in full-text in English (n = 4). This left 27 studies eligible for inclusion. Table 1 outlines the characteristics of the 27 studies included in the review. Of the 27 studies, eight studies focused on the conceptualisation, development, testing or review of a questionnaire [14][15][16][17][18][19][20][21] and 19 studies used a questionnaire in a study investigating the quality of relationships between patients and primary care providers [6,19,[22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38]. The following phrases were used in the studies to refer to "quality of relationships" and considered synonymous with the topic of this review: care continuity, continuing relationship, continuity of care, interpersonal care, long term relationships, longitudinality of care, patient-doctor relationships, patient-provider relationships, personal continuity, quality of care, relational continuity, relations, therapeutic alliance, therapeutic relationship.
Some questionnaires were used in several of the studies. Therefore, although 27 studies were included in the review, only 14 questionnaires were used. Table 2 summarises the 14 relevant questionnaires used in the studies. Three of the questionnaires (Primary Care Assessment Survey; Primary Care Assessment Tool; Primary Care Evaluation Tool) were large instruments investigating multiple components of quality care, with only a very small section (e.g. one subscale) examining quality of relationships. These questionnaires were considered unfeasible for future use because most of the data would be irrelevant to the topic. Also, they would require substantial time (e.g. 45 min for the Primary Care Assessment Survey) to complete the questionnaire. Two of the questionnaires (neither with a name) were developed only for use in the reviewed study and were not pilot tested for the purpose of others' utilising the questionnaires in work [22,28]. These two questionnaires were considered unfeasible for future use as there was no evidence to support their validity. This left nine questionnaires that were examined further. Table 3 provides information on the feasibility of using the nine remaining questionnaires. The questionnaires are diverse in length; the shortest being the Therapeutic Bond Scale (6 items; 30 s to complete) and the longest being the Care Continuity Across Levels of Care Scale (73 items, up to 15 min to complete). Some of the questionnaires focus entirely on assessing quality of relationships, such as Patient-Doctor Relationship Questionnaire (100% relevant items). However, for other questionnaires, relationship quality is not the only focus, such as the Generic Measure of Continuity Scale (34% relevant items) and Nijmegan Continuity Questionnaire (28% relevant items). Seven of the questionnaires are freely available for use, whereas the Generic Measure of Continuity Scale and Therapeutic Bond Scale requires payment prior to use. All the questionnaires are in the English language, with several also translated to other languages.

Discussion
This study systematically reviewed the body of evidence to address the lack of understanding on how to best measure the quality of continuous relationships between patients and primary care providers. Fourteen relevant questionnaires were found in the 27 studies included in the review. Of the 14 questionnaires, nine were considered as potentially feasible for future use, including three that were considered strongest candidates based on being relevant, freely available in English and not needing  additional pilot work prior to use. These three questionnaires are the Care Continuity Across Levels of Care Scale (CCAENA), the Nijmegan Continuity Questionnaire and the Patient-Doctor Depth of Relationship Tool. The decision to use one of the reviewed questionnaires in future work requires careful consideration. Some of the questionnaires focussed solely on assessing quality of relationships and did not examine any other topics, for example the Patient Doctor Relationship Questionnaire (PDRQ-9) [39]. However, for this questionnaire, no evidence of pilot testing was found that confirmed the content was relevant and sufficiently comprehensive to assess the experiential nature of relationship quality. Questionnaires with a broader focus could be interpreted as less relevant, such as the Nijmegan Continuity questionnaire (28% relevant items), however this questionnaire has confirmed construct validity and test-retest validity, demonstrating its appropriateness for future use in research [20,40]. Researchers and primary care workers are encouraged to consider several factors that may impact on the use of these questionnaires in their work, including their scope, focus, length, availability and validity testing.
Caution is needed when interpreting the level of validity testing undertaken for the questionnaires used in the studies in the review. Diverse terms were used to describe the same type of validity testing (such as internal consistency and construct validity) [15,20,31,40]. Furthermore, only two studies assessed test-retest reliability, the Nijmegan Continuity Questionnaire and Patient-doctor depth of relationship tool. Confirming test-retest reliability is considered essential for evaluations of interventions in order to be confident that any changes seen in results over time is due to a change in service rather than natural variation of results [41]. Undertaking validity testing does not guarantee that a questionnaire is "valid". For example, the authors of the Generic Measure of Continuity Scale conducted a pilot study to investigate its correlation with pre-identified indicators of continuity and found very low correlation [15]. No Table 1 Characteristics of included studies assessing the quality of continuous relationships between patients and providers in primary care grouped by inclusion criteria and in alphabetical order of first author (Continued)  Internal consistency of subscales was tested (Cronbach's alpha ranged from 0.8-1.0). changes were made to the scale to ameliorate the low correlation, which hampers its use without further development work. Conversely, the Patient-doctor depth of relationship tool has undertaken the most comprehensive pilot testing work of all the reviewed questionnaires and has confirmed good face validity, high internal reliability and strong test-retest reliability, indicating its appropriateness for future use [42]. This is a comprehensive review which identified six questionnaires that were not captured in the previous review [12]. Two of the questionnaires were included in the previous systematic review (Patient-doctor depth of relationship tool and Patient doctor relationship questionnaire (PDRQ-9)) and continue to be used in studies [21,36]. The remaining questionnaires have been developed or refined since this time, indicating an increasing focus on this aspect of health care evaluation. However, this review has also identified some notable limitations regarding questionnaires assessing quality of relationships between patients and primary care providers. None of the questionnaires consider providers' perspectives relationships, or the association between patients' and providers' perspectives on their relationship. Furthermore, none of the studies investigated whether the quality of relationship predicted patient outcomes, warranting future work to confirm the notion that quality of relationships is associated with improved outcomes.

Conclusions
This study provides an overview of 14 unique questionnaires that have been used to assess the quality of relationships between patients and primary care providers. This area is of increasing importance with the growing focus on patient engagement as a critical element in the prevention and management of chronic disease and unhealthy lifestyle choice. The selection of a questionnaire for future work should be based on its scope, focus, length and feasibility for use in the setting in which it will be applied.  No evidence of establishing reliability of construct validity, which may preclude its use in evaluation work