Behavior change interventions and policies influencing primary healthcare professionals’ practice—an overview of reviews

Background There is a plethora of interventions and policies aimed at changing practice habits of primary healthcare professionals, but it is unclear which are the most appropriate, sustainable, and effective. We aimed to evaluate the evidence on behavior change interventions and policies directed at healthcare professionals working in primary healthcare centers. Methods Study design: overview of reviews. Data source: MEDLINE (Ovid), Embase (Ovid), The Cochrane Library (Wiley), CINAHL (EbscoHost), and grey literature (January 2005 to July 2015). Study selection: two reviewers independently, and in duplicate, identified systematic reviews, overviews of reviews, scoping reviews, rapid reviews, and relevant health technology reports published in full-text in the English language. Data extraction and synthesis: two reviewers extracted data pertaining to the types of reviews, study designs, number of studies, demographics of the professionals enrolled, interventions, outcomes, and authors’ conclusions for the included studies. We evaluated the methodological quality of the included studies using the AMSTAR scale. For the comparative evaluation, we classified interventions according to the behavior change wheel (Michie et al.). Results Of 2771 citations retrieved, we included 138 reviews representing 3502 individual studies. The majority of systematic reviews (91%) investigated behavior and practice changes among family physicians. Interactive and multifaceted continuous medical education programs, training with audit and feedback, and clinical decision support systems were found to be beneficial in improving knowledge, optimizing screening rate and prescriptions, enhancing patient outcomes, and reducing adverse events. Collaborative team-based policies involving primarily family physicians, nurses, and pharmacists were found to be most effective. Available evidence on environmental restructuring and modeling was found to be effective in improving collaboration and adherence to treatment guidelines. Limited evidence on nurse-led care approaches were found to be as effective as general practitioners in patient satisfaction in settings like asthma, cardiovascular, and diabetes clinics, although this needs further evaluation. Evidence does not support the use of financial incentives to family physicians, especially for long-term behavior change. Conclusions Behavior change interventions including education, training, and enablement in the context of collaborative team-based approaches are effective to change practice of primary healthcare professionals. Environmental restructuring approaches including nurse-led care and modeling need further evaluation. Financial incentives to family physicians do not influence long-term practice change. Electronic supplementary material The online version of this article (doi:10.1186/s13012-016-0538-8) contains supplementary material, which is available to authorized users.


Introduction
Approximately one in six Canadians aged 20 years or older suffer from chronic diseases such as diabetes, cardiovascular diseases, chronic respiratory diseases, arthritis, osteoporosis, mental illness, and cancer [1]. Combining direct medical costs ($38.9 billion) and indirect productivity losses ($54.4 billion), the total economic burden of chronic illness exceeds Canadian $93 billion a year [2]. Despite this enormous expenditure, 12 to 15% of Canadians feel they receive inadequate chronic disease care [3,4]. The major unmet needs include long waiting periods for medical services [5] and unavailability of essential services [4]. Compared with people in other developed nations, Canadians today are less satisfied with their access to and quality of care [6] and have worse health outcomes for several medical conditions [7]. The numbers of patients with chronic diseases and the existing gap in quality of care present a significant challenge for public health policy-makers [8,9].
With the objective of closing gaps in quality of care and managing patients with chronic diseases, the implementation of patient-centred treatment has recently gained attention from policy-makers [10][11][12]. Patient-centered medical centres may become the future backbone of the Canadian healthcare system [13]. These teams may include family physicians, physician assistants, nurses, pharmacists, social workers, mental health counselors/psychologists, dieticians, and midwives among others. To achieve efficient and effective patient-centered medical homes, some changes in the way healthcare is delivered will be required. To do so, effective behavior change interventions and supporting policies are required [14,15]. However, it is unclear which intervention(s) and policies are appropriate, sustainable, and sufficiently safe to support practice change and improve patient-relevant outcomes in primary healthcare settings. Despite extensive published literature including randomized controlled trials [16,17], observational studies [18,19], and systematic reviews [20][21][22], no recent comprehensive review classifying or evaluating the feasibility or effectiveness of interventions and policies in terms of patients' and professionals' outcomes exists. The objectives of this overview of reviews were to identify, classify, and critically appraise reviews evaluating behavior change interventions and policies influencing primary healthcare professionals working at primary healthcare centers.

Data sources and searches
The search strategy was developed and tested through an iterative process by an experienced medical information specialist in consultation with the review team. We searched MEDLINE (Ovid), Embase (Ovid), CINAHL (EbscoHost), and the Cochrane Library (Wiley). Strategies utilized a combination of controlled vocabulary (e.g., "Physicians", "Primary Care", "Physician's Practice Patterns", "Quality Improvement") and keywords (e.g., family practitioner, home clinic, policy adherence). Vocabulary and syntax were adjusted across databases. Results were restricted to the English language and the dates from January 2005 to July 2015 (Additional file 1). We used Dis-tillerSR (Version 2, Evidence Partners Inc. ON, Canada) for study selection, data extraction, and project management.

Study selection
We included (1) systematic reviews, overview of reviews, scoping reviews, rapid reviews, or health technology assessments that (2) evaluated behavior change interventions or policies on primary healthcare professionals (including general practitioners/family physicians, physician assistants, nurses, pharmacists, social workers, mental health counselors/psychologists, dieticians, and midwives) (3) working at primary healthcare settings (4) reporting any outcomes of primary healthcare professionals' practice change, and (5) published in the English language as full-text articles. Primary healthcare settings were defined as the provision of integrated, accessible health care services by clinicians who are accountable for addressing a large majority of personal health care needs, developing a sustained partnership with patients, and practicing in the context of family and community [23,24]. Considering the application of outcomes in the Canadian context, reviews that exclusively included studies conducted in either underdeveloped or developing countries were excluded.
The abstracts and titles of relevant citations were independently screened by two reviewers to determine eligibility. The same two reviewers independently assessed the eligibility of full-text reports of relevant citations using a standardized pre-piloted form outlining the inclusion and exclusion criteria. Disagreements were resolved by consensus or with the involvement of a third reviewer, if needed.

Data extraction and quality assessment
Two reviewers independently abstracted data from the included reviews using standardized piloted forms. The following data were extracted from each included review: review type, number and study designs that the review included, types of professionals evaluated, interventions, outcomes, therapeutic domains, and authors' conclusions.
All behavior change interventions and policies were classified into nine categories of interventions and seven categories of policies following the behavior change wheel framework proposed by Michie et al. [15]. This framework consists of a behavior system at the hub, encircled by nine intervention functions and then by seven policy categories. The nine behavior change interventions include (1) education (increasing knowledge or understanding): e.g., continuous medical education; (2) persuasion (using communication to induce positive or negative feelings or stimulate action): e.g., reminders; (3) incentivization (creating expectation of reward): e.g., payment for performance; (4) coercion (creating expectation of punishment or cost): e.g., punishment or fines; (5) training (imparting skills): e.g., communication skills training; (6) restriction (using rules to reduce the opportunity to engage in the target behavior): e.g., rules for prohibiting the use; (7) environmental restructuring (changing the physical or social context): e.g., shared decisionmaking; (8) modeling (providing an example for people to aspire to or imitate): e.g., local opinion leaders; (9) enablement (increasing means/reducing barriers to increase capability or opportunity): e.g., clinical decision support systems. While the seven policies include: (1) communication/ marketing (using print, electronic, telephonic or broadcast media): e.g., advertising media; (2) guidelines (creating documents that recommend or mandate practice): e.g., management guidelines; (3) fiscal (using the tax system to reduce or increase the financial cost): e.g., financial provisions from policy-makers; (4) regulation (establishing rules or principles of behavior or practice): e.g., rules and regulations; (5) legislation (making or changing laws): e.g., law amendments; (6) environmental/social planning (designing and/or controlling the physical or social environment): e.g., social support; (7) service provision (delivering a service): e.g., service or facilitation.
Two reviewers independently, and in duplicate, evaluated the methodological quality of the included reviews using the assessing the methodological quality of systematic reviews (AMSTAR) scoring system [25]. Conflicts were resolved by consensus or discussion with a third reviewer, if needed. Reviews with AMSTAR score ≥8, 4 to 7, ≤3 were considered as high, moderate, or lowmethodological quality, respectively.
We summarized the findings that emerged from the subjective judgment matrix, which was based on the authors' conclusions, qualitative data, quantitative data with statistically significant group differences in terms of patients' and primary healthcare providers' outcomes, and the methodological quality of included reviews [25][26][27][28]. The protocol for this overview of reviews has been developed prior to conduct the review and provided to the Primary Health Care Branch, Manitoba Health, Seniors and Active Living, Government of Manitoba, Canada. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines for reporting the systematic review were followed.
Modeling (providing an example for people to aspire or imitate) Two reviews [98,99] (n = 60 studies) evaluated modeling using local opinion leaders [98], or mental health workers [99] in primary healthcare settings. Evidence from moderate-to high-quality reviews demonstrated that involving local opinion leaders or subject experts to promote evidence-informed practices decreased the rates of consultations and prescriptions [98,99].

Multiple interventions
Several reviews were focused on how to better manage chronic diseases using any behavior change interventions.
To avoid misclassification, we classified these reviews under an umbrella term, multiple interventions. Forty-one reviews [31, (n = 1375 studies) of poor-to highquality focused on multiple interventions. The use of computer alerts within electronic medical records increased screening for sexually transmitted diseases [115].
Interventions in pharmacy services reduced suboptimal prescribing [117,127,133], and educational interventions improved primary healthcare providers' identification, assessment, prevention and/or management of obesity in children and adolescents to achieve weight loss [121]. No review focused exclusively on audit and feedback, but multifaceted audit/feedback, reminders, educational outreach visits, and patient-mediated interventions [31,116,118,119] were found to be effective in influencing health professionals' prescribing practice. Financial incentives combined with educational interventions and audit/feedback have been found to be effective in increasing the practice of generic prescribing [124]. Multifaceted interventions where educational interventions occurred at many levels may be successfully incorporated into established medical communities after addressing local barriers to change [120,123,130,153]. Advance practice nurse care [136], quality improvement strategies [137,[148][149][150][151][152], case management [138], collaborative care [140], evidencebased medicine practice strategies [144], midwife-led continuity services [145], comprehensive asthma care [146], and patient-centered medical home [125,147] have all been evaluated. Moderate-to high-quality reviews demonstrated improved safety, quality care, increased vaccination rate, and improved management of patient with depression and anxiety in primary healthcare settings [135-137, 139-142, 144, 147, 148, 150, 151]. Few reviews failed to provide any conclusive outcomes [122,126,129,131,134,143,154,155].
Policies (Additional file 1: Table S1) Service provision (delivering a service) Five reviews [156][157][158][159][160] (n = 44 studies) of poor-to highquality evaluated effects of consultation time [156,158], brief non-pharmacological interventions (computer-based cognitive-behavioral therapy) [157], and non-medical prescribing [159] (drug prescriptions by nurses, pharmacists, and allied health professionals) on behavioral change of primary healthcare professionals. While a health technology report [160] assessed evidence on specialized communitybased care and concluded that specialized communitybased care effectively improves outcomes in patients with heart failure, chronic obstructive pulmonary disease, and diabetes. Bibliotherapy, cognitive behavioral therapy-based websites, and cognitive behavioral therapy-based computer programs [157] found to be effective in improving management of patients with depression. Other reviews [156,158,159] were not found to be effective.
Communication (using print, electronic, telephone, or broadcast media) Three reviews [161][162][163] (n = 44 studies) of moderateto high-quality evaluated communication as an intervention reporting inconclusive results. One review [161] uniquely assessed whether patients benefit from improved communication between primary healthcare practitioners and nephrologists. The review found little evidence of benefit from enhancing the quality of letters from specialists to primary healthcare practitioners.

Guidelines (creating documents that recommend practice standards)
Two reviews [164,165] (n = 42 studies) of moderate-to high-quality evaluated the impact of guidelines on the improvement of healthcare professionals' practice. None of the interventions found to be effective method for increasing advance directive completion rates in the primary healthcare setting [164,165].

Discussion
In our overview of reviews, we identified, classified, and evaluated the behavior change interventions and policies influencing practice change of primary healthcare professionals who primarily manage patients with chronic diseases at primary healthcare centers. Interactive and multifaceted continuous medical education programs including training with audit and feedback, and clinical decision support systems were found to be of benefit in improving knowledge, optimizing prescriptions, increasing screening rate, enhancing patient outcomes, and reducing adverse events. Limited evidence on environmental restructuring and modeling were found to be effective in improving collaboration and adherence to treatment guidelines. Collaborative team-based approaches involving primarily family physicians, nurses, and pharmacists were found to be effective. Limited evidence on nurse-led care approaches were found to be promising and warrant further evaluation using better study designs for different chronic diseases. Evidence clearly does not support the use of financial incentives to family physicians, especially for long-term sustained behavior and practice change.
To the best of our knowledge, so far this is the largest comprehensive overview of reviews evaluating authors' reported efficacy of behavior change interventions and policies influencing primary healthcare professionals' practice change and classified according to the behavior change wheel proposed by Michie et al. [15]. Our outcomes support the inferences reported by other overview reviews [166] and review [167] focused on individual interventions. Grimshaw and colleagues [166] reported that educational outreach (for prescribing) and reminders were found to be most promising approaches. Multifaceted interventions targeting different barriers to change are more likely to be effective than single interventions. We reported that education intervention found to be effective, especially when used as multifaceted interventions to achieve primary healthcare professionals' practice change to improve quality of care and better manage patients with chronic diseases. Ivers and colleagues [167] reported audit and feedback generally leads to small but potentially important improvements in professional practice. We did not find any review exclusively evaluating audit and feedback on primary healthcare professionals; however, it was used with other interventions (e.g., education and training) and provided mixed results. With regards to financial incentives, Flodgren and colleagues have reported that financial incentives may be effective in changing healthcare professional practice [168]. In contrast, we found that financial incentives were not effective in practice change of family physicians working at primary healthcare centers.
This review did identify limited evidence on a few promising interventions, including nurse-led approaches and use of opinion leaders or specialists. Further, thorough evaluation in specific areas of interest should be performed before they are widely implemented in a healthcare setting.
To reduce the gap in quality of care and better manage patients with chronic diseases, behavioral interventions and supporting policies are essential. Through this overview of reviews, we attempted to provide an evidence to improve our understanding on which behavioral interventions and policies are effective to influence practice of primary healthcare professionals working in primary health care settings. This review is heavily weighted by evidence on family physicians, thus indicating the need for studies on other primary healthcare professionals. We excluded reviews that either evaluated these interventions and policies on specialists and hospital settings or included studies conducted exclusively in low-to middleincome countries, where the functionality of healthcare systems is different than Canada. Behavior change interventions or policies were classified based on the framework proposed by Michie and colleagues [15] and no other frameworks were explored or compared. Considering this is an overview of reviews and we have not performed a meta-analysis, we did not attempt to review individual studies from included reviews; there is a possibility of few studies might have been included by multiple reviews or might be a chance of over representation of outcomes. Evidence ranged from poor-to high-quality as well the high heterogeneity in interventions, study population, and outcomes prevented to generalize the conclusion to specific category of primary healthcare professionals or interventions and policies.

Conclusion
Behavior change interventions including interactive and multifaceted continuous medical education, training with audit and feedback, enablement through advanced information technology-based systems, and collaborative team-based interventions can effectively modify healthcare professionals' practice and patient outcomes. Limited evidence exists to support environment restructuring and modeling. Nurse-led systems of care warrant further evaluation. Financial incentives to family physicians do not influence long-term behavior and practice change.

Additional file
Additional file 1: Table S1. Outcomes and methodological quality assessment of included reviews. (DOC 662 kb)

Acknowledgements
We sincerely thank Kristin Anderson, Lindsay Story, and Nathan Hoeppner from Manitoba Health, Seniors and Active Living for their suggestions and comments on the subject, as well as the protocol of the overview.

Funding
None. Dr. Zarychanski is a recipient of the New Investigator Salary Award from the Canadian Institutes of Health Research.

Availability of data and materials
For the additional information on data and material presented in this manuscript, please contact the corresponding author.
Authors' contributions Dr. BFC played a role in the conceptualization of the project, wrote the protocol, led and coordinated this overview, screened citations, assessed studies for eligibility, extracted data, performed quality assessments, drafted and revised the manuscript, approved the final version of the manuscript as submitted, and agrees to be accountable for all aspects pertaining to the overview. Drs. MJ, ASM, and JL screened citations, assessed studies for eligibility, extracted data, and performed methodological quality assessments. Drs. KMS, AA-S, and RZ and played a key role in the conceptualization of the project and provided methodological expertise during the protocol development and conduct of the overview. They critically reviewed and provided expert comments on the manuscript and approved the final version of the manuscript. BS played an important role in designing and executing the search strategy, provided relevant comments on the manuscript, and approved the final version.