Using Consolidated Framework for Implementation Research to investigate facilitators and barriers of implementing alcohol screening and brief intervention among primary care health professionals: a systematic review

Background Alcohol screening and brief intervention (SBI) is recommended to be implemented in primary care settings to intervene against hazardous/harmful drinking. However, studies showed that the uptake rate was low in many regions/countries. This systematic review presented current findings on the facilitators and barriers of SBI implemented by health professionals in primary care settings using the Consolidated Framework for Implementation Research (CFIR). Methods We included qualitative, quantitative, and mixed-method studies identified through four electronic databases (PubMed, MEDLINE, PsycInfo, and Web of Science) from inception to June 2020. Included articles had to address barriers and facilitators of SBI implementation and provide sufficient details that the CFIR domains could be identified and data were abstracted using a standardized extraction form. Results A total of 74 studies published from 1985 to 2019 were finally analysed and summarized. The most common facilitators were knowledge and positive beliefs about SBI (characteristics of the individuals) and available resources (inner setting). In contrast, the most common barriers were cost related to implementing SBI (intervention characteristics), negative beliefs about SBI (characteristics of the individuals), and lack of self-efficacy in implementing SBI (characteristics of the individuals). It could be observed that factors related to the inner setting and characteristics of individuals were extensively studied whilst the process received the least attention. Conclusions Most of the facilitators and barriers are modifiable. Additionally, most literature focused on various kinds of available assets to implement SBI. To promote the spread of SBI implementation, more high-quality studies on the implementation process are needed. This systematic review could serve as a reference framework for health authorities to devise strategies for improving the implementation of SBI in primary care settings. Trial registration This systematic review was registered in PROSPERO (CRD42021258833). Supplementary Information The online version contains supplementary material available at 10.1186/s13012-021-01170-8.


Introduction
Harmful use of alcohol is a known risk factor for more than 200 types of diseases and injuries [1]. In light of this, alcohol screening and brief intervention (SBI) techniques have been developed in a bid to rectify the situation that excessive alcohol use causes harm. SBI measures a person's level of alcohol consumption and provides brief interventions based on their drinking level [2].
The World Health Organization (WHO) stressed the need to increase coverage of SBI in order to enable early identification and intervention against hazardous/harmful drinking behaviour before serious consequences happen [3]. Across countries, primary care professionals are in a unique and privileged position to identify and intervene against hazardous/harmful drinking [4,5]. A meta-analysis of 34 randomized controlled trials (RCTs) showed that stand-alone SBI in primary care settings had a significant and moderate effect in reducing alcohol consumption among hazardous/harmful drinkers, as compared to no or minimum intervention [6]. Early identification and secondary prevention of alcohol use disorder using SBI in primary care settings are strongly recommended by the WHO [2] and other national health authorities [7][8][9].
However, there was a significant gap between the actual implementation and what is recommended for SBI [10]. Screening rates in European countries were low [11] and less than half of individuals engaged in hazardous drinking were identified by their general practitioners (GPs) [12]. In view of this low uptake of SBI, a number of studies have identified facilitators or barriers to practising SBI in primary care settings. The enabling factors include training [13,14], proven efficacy of SBI [15,16], financial incentives [17,18], and support from government policy [15,19]. On the other hand, a broad variety of barriers were reported, such as lack of time [20,21], lack of counselling skills [16,22], low availability of screening or intervention tools [23,24], and fear of harming their relationship with the patient [25,26]. Nevertheless, the previous studies did not use an implementation science lens to look at facilitators and barriers to implementing SBI, except for one study. That study used Greenhalgh's conceptual framework for dissemination of innovations to explore facilitators and barriers of SBI implemented by different health professionals in the USA [27].
To our knowledge, two systematic reviews published in 2011 and 2017 summarized the facilitators and barriers to implementing SBI in various settings (e.g. trauma centres, in-patient settings, primary care settings, etc.) [28,29]. They did not give a separate discussion on primary care settings where SBI was suggested to be implemented by the WHO. The number of studies conducted in primary care settings was relatively small in the two reviews (n = 31, n = 14). One of them only included qualitative studies. Furthermore, both systematic reviews did not use the lens of implementation science to synthesize or discuss the findings. The present systematic review included quantitative, qualitative, and mixed-method studies which focused exclusively on primary care settings. It also used an implementation science framework to synthesize the findings.
Due to the lack of application of implementation science framework in this research area, in the present systematic review, the Consolidated Framework for Implementation Research (CFIR) was used to analyse and summarize the facilitators and barriers to implementing SBI in primary care settings. The CFIR has been used to guide the systematic assessment of multi-level implementation contexts to identify facilitators and barriers that might influence implementation [30,31]. It provides a comprehensive and standardized list of implementation-related constructs that can be applied across the spectrum of implementation research [30,31]. The CFIR consists of five domains, including intervention characteristics (features of an intervention that might influence implementation), inner setting (features of the organization that might influence implementation), outer setting (features of Keywords: Consolidated Framework for Implementation Research, Alcohol screening, Brief intervention, Facilitators, Barriers, Primary care, Health professionals external context or environment that might influence implementation), characteristics of individuals (individuals involved in implementation that might influence implementation), and implementation process (refers to the plan of implementing a given innovation, to the contents and quality of the plan and how it has been adhered to during the actual implementation process). Making use of the CFIR is helpful to generalize the findings across contexts [30].
The present systematic review included quantitative, qualitative, and mixed-method studies which focused exclusively on facilitators and barriers to implementing SBI in primary care settings. It also used the implementation science framework, the CFIR, to synthesize the findings in a systematic manner, which could serve as a reference framework for health authorities to devise strategies to improve the implementation of SBI in primary care settings.

Methods
This systematic review was conducted according to a preregistered protocol in PROSPERO (CRD42021258833) and the PRISMA guideline [32].

Search strategy
Articles were identified by searching the electronic databases PubMed, MEDLINE, PsycInfo, and Web of Science, covering the publication periods from inception to June 2020. Implementing SBI in primary care settings started in the 1980s as a result of the WHO international collaborative project on SBI [33]. The guideline/recommendation and practice of SBI in primary care settings have not changed significantly since then. The Boolean operator was used in the search strategy conducted, using "OR" and/or "AND" to link search terms. The asterisk "*" was used as a wildcard symbol appended at the end of the terms to search for variations of those terms (Additional file 1).

Selection criteria
The articles included in this review were original, quantitative, qualitative, or mixed-method studies published in peer-reviewed journals. The studies examined the facilitators and/or barriers of alcohol screening or alcohol brief intervention implemented by healthcare professionals (physicians, nurses, and other health workers) in primary care settings. In this review, barriers refer to obstacles that hinder health professionals from performing SBI and facilitators refer to enabling factors for health professionals to perform SBI. We excluded articles that focused on: (1) efficacy of alcohol screening and brief intervention; and (2) alcohol screening combined with other drugs' screening.

Quality assessment and analysis
The information from the included articles was assessed by the Mixed-Method Appraisal Tool (MMAT) version 2011 with detailed descriptions of the rating [34]. A data extraction form was used which included reference ID, first author, publication year, title, country, study design, participants, sample size, facilitators, barriers, CFIR domains, and constructs. We used a thematic approach that the facilitators and barriers were coded on the CFIR framework. A six-step data synthesis process of the facilitators and barriers was developed: (1) Two reviewers extracted facilitators and barriers from each article independently; (2) After extraction, they discussed to achieve a consensus on the facilitators and barriers identified in each article. In some cases, wordings of the same facilitator or barrier were slightly different in different studies. The wordings were revised and the same description for the same facilitator or barrier was used after discussion by the two reviewers; (3) Each facilitator or barrier was coded under the domains/constructs of the CFIR by the two reviewers independently; (4) After finishing the coding independently, they discussed the results, and any discrepancies were resolved through discussion; (5) The revised coding results were read and checked by the two reviewers independently to ensure all facilitators and barriers were mapped to the CFIR constructs correctly; (6) All information in the codebook was adapted to make Tables 1 and 2.

Identification of studies
The initial search returned 4078 citations, of which 1285 were excluded due to duplicates (Fig. 1). After that, we further removed 2681 articles after screening for titles and abstracts. We performed full-text screening on 112 articles, of which 38 articles were excluded because they did not meet the selection criteria. Figure 1 presents the PRISMA flow chart of the selected studies.

Practice of SBI
Although the analysis of the practice of SBI was not the main aim of this review, we tried to extract related information from the included studies and give a preliminary analysis in this area. Among the included studies, participants were asked about their current practice of SBI in 15 studies, and these studies were conducted in nine countries, i.e. Finland, the UK, Germany, the USA, France and South Africa, Sweden, Sri Lanka, and Canada (Additional file 3). The participants in these studies were all physicians, except in one study in which nurses were also included. There were two ways of measuring their practice: 1) had ever performed SBI; 2) performed SBI on a regular basis. The percentage of participants who reported that they had ever performed screening or brief intervention ranged from 45.0% to 100%. However, the percentage was much lower when it was on a regular basis, which ranged from 9.4% to 40.0%, except for one study with 75.0%. Regarding whether they had performed SBI, the highest rate was found in South Africa (100%), the UK (98.0%), the USA (95.0% and 84.0% in two studies), France (94.1%), and Germany (84.2%), whereas the lowest rate was found in Finland (45.0%). For regular basis, the highest rate was found in Canada (75.0%) and the UK (40%) whilst the lowest rate was found in Finland (9.4%) and Sri Lanka (15.0%).

Facilitators and barriers of SBI implementation based on the CFIR
The following results are presented according to Table 2.

Intervention characteristics
Evidence strength was considered by primary healthcare providers when implementing SBI. About 74-81% of physicians in the UK and Poland agreed that the proven efficacy of early alcohol intervention was a facilitator of implementing SBI [15,60,93], whilst doubt about the effectiveness of brief interventions was cited as a barrier to implementing SBI by physicians or nurses in the USA, the UK, and Finland [21,23,26,46,56,90]. For adaptability, physicians, nurses, and other health professionals (e.g. social workers, psychologists) in the USA, Catalonia, the Netherlands, Poland, Sweden, and the UK suggested that SBI could be adapted or refined to meet special needs, such as using computer-based methods for screening, targeted rather than universal screening (e.g. new patient registrations, general health checks, and particular types of consultations) [26,56,74]. Perceiving SBI

Characteristics of individuals
Self-efficacy: self-efficacy

Intervention characteristics
Cost: workload or lack of time; causing management problems

Characteristics of individuals
Knowledge and beliefs: belief that it was not a rewarding task
For cosmopolitanism, available referral services were reported by physicians, nurses, and other health professionals as facilitators of implement SBI, such as the provision of addiction care and specialized treatment for alcohol problems [15,18,19,24,46,50,52,57,59,60,72,78,80,86,93]. For instance, about 59-94% of physicians and 57-83%% of nurses in Poland, Sweden, the UK, Sri Lanka, and the USA reported that access to local community alcohol teams, general support services (e.g. self-help or counselling), were facilitators of implementing SBI [15,24,46,50,52,57,59,60,72,80,93]. Regression analyses in one study conducted in Norway also showed that having places to refer patients to was significantly associated with physicians' screening or brief intervention activity [18]. On the contrary, the lack of referral services was also cited as a barrier to implementing SBI [16, 38, 54-56, 83, 89, 90, 92, 27]. For instance, about 52-76%% of physicians in Canada and the USA mentioned that wait-lists were long and treatment services were limited [89,90]. Moreover, univariate analyses showed that physicians in Sweden who infrequently addressed alcohol issues were more likely to be uncertain where to refer the patients [59]. Support from external policy was cited as an incentive to implement SBI [15,16,19,60,93]. About 65-82% of physicians in the UK and Poland suggested that implementation of SBI as part of a national strategy and more public health education campaigns make society more concerned about alcohol were enablers of SBI implementation [15,60,93]. However, lack of support from government policy was usually mentioned [15,16,22,60,82]. For instance, government policy that did not support preventive medicine was pointed out by 56-98% of physicians in South Africa, the UK, and Poland [15,22,60].
One construct, peer pressure, was not covered by previous studies.

Inner setting
For structural characteristics, teamwork or interprofessional cooperation in the delivery of SBI was suggested as a facilitator [38,59,74,81]. In contrast, lack of staff, specialist support, or multidisciplinary team in primary care settings were cited as barriers [26,41,51,92].
Concerning compatibility, doubt about the appropriateness of screening all patients and such activity causing interruptions of the natural course of consultations were reported by physicians in South Africa, Canada, and some European countries such as Norway, Slovenia, and the UK [42,62,78,82,86].
High prioritization of alcohol issue was cited as an enabler by physicians, nurses, and clinic managers in South Africa, and the USA [74,81] whereas a low rating of the importance of alcohol matter, competing priorities, or patients with multiple problems were reported as barriers by physicians, nurses, social workers, psychologists and other health professionals in Brazil, France, Germany, Italy, Spain, and the UK [38,58,83].
Two types of organizational rewards and incentives reported as facilitators, including provision of financial reimbursements/ salary improvements, and training for early alcohol intervention recognized for quality assurance credits [15-19, 60, 93]. About 35-84% of physicians in Australia and some European countries considered that provision of financial reimbursements/salary improvements for carrying out early alcohol intervention, recognizing SBI training for continuing medical education (CME), or recognizing SBI for quality assurance credits were facilitators [15,60,93]. On the contrary, a lack of organizational incentives for SBI implementation was reported across countries [15,16,22,26,60,69,82,92,93]. For instance, about 31-90% of physicians in Poland, South Africa, the UK, and the USA reported that there was a lack of contractual incentives or the government health scheme did not reimburse their time spent on preventive medicine [15,22,60,69,93].
Regarding access to knowledge and information, easy access to clear guidelines related to implementing SBI was suggested [13,47,59]. Support calls responding to questions or problems that arose during SBI implementation were demonstrated to be effective in two multi-country studies [14,40]. However, a lack of simple guidelines was reported by physicians and nurses in Finland and Slovenia [20,82].
Four constructs in this domain were briefly discussed, i.e. culture, learning climate, goals and feedback, and leadership engagement. Regarding culture, one study conducted in Brazil reported that the organizational culture of alcohol use (e.g. often using alcohol during workrelated celebrations) in the primary care service of the Military Police would hamper treating problem drinkers [38]. For learning climate, more chances to try and observe SBI were reported as facilitators by nurses and clinic managers in South Africa [81]. A lack of understanding of the goals of SBI was mentioned in one study in the USA [27]. A decrease in institutional support due to changes in leadership was also reported in one study conducted in Brazil [38].
In this domain, two constructs, networks and communications, and tension for change, were not studied.
For other personal attributes, many demographic characteristics were found to be positively associated with SBI implementation. These personal characteristics included male patients [18,44,66,71], younger physician age [18,55], smaller number of patients seen by GPs in an average week [60], GPs having longer average practice consultations [66], GPs in solo practice [66]. In contrast, other personal characteristics were found to be barriers, such as old patients aged 60-69 years [66], physicians having alcohol drinking habits or problems [60,82], some nurses worrying more or having lower self-efficacy than physicians [61]. Nevertheless, there were mixed results that some studies showed that female healthcare provider was a facilitator [44,55,72,84], whilst it was considered a barrier in one study [86].
In this domain, two constructs, i.e. individual stage of change, and individual identification with the organization were not covered by previous studies.

Process
There is a lack of studies in this domain. For executing, physicians reported that there was a lack of a systematic strategy in the clinic in one study conducted in Canada [86]. For reflecting and evaluating, the feedback provided by the SBI trainers during their visits at the clinics was found to be helpful in one study conducted in South Africa [81].
Two constructs, planning and engaging, had not been studied.

Facilitators and barriers not covered by the CFIR
Facilitators that were not covered by the CFIR included medical documentation of patients' alcohol drinking, reminders for providers to do SBI, and reminders for patients about their medical condition. Barriers that were not covered by the CFIR included stigma-related issues, conflicting messages in the society that drinking alcohol was acceptable and even beneficial to health, alcohol counselling involving family and wider social effects, and providers' struggles in prevention versus treatment.

Discussion
In this systematic review, we sought to identify facilitators and barriers of SBI implemented by health professionals in primary care settings. We used the CFIR framework to analyse and summarize the facilitators and barriers.
We found that the practice rate of SBI was low in most countries on a regular basis. Among these countries, Finland, England, and France have participated in the WHO Phase IV implementation project (World Health Organization Collaborative Project on Identification and Management of Alcohol-Related Problems in Primary Health Care) [94]. Although the practice rate has increased since the commencement of this project, e.g. Finland [87], their practice rate is still below the satisfactory level. This may be due to social contextual factors. For instance, in Finland, alcohol problems used to be handled mostly by social welfare authorities and the police, and the disease model was adopted only later [21]. For future studies, researchers should investigate the specific contextual facilitators and barriers, and devise targeting strategies to help health professionals improve the practice of SBI in primary care settings. Nonetheless, the analysis of the practice of SBI was not the main aim of this review. Future work on the methodology of searching for articles specifically studying the current practice of SBI is recommended and a separate review and in-depth analysis will bring a more thorough discussion in this area.
The most common facilitators were knowledge and positive beliefs about SBI (characteristics of the individuals) and available resources (inner setting), whilst the most common barriers were cost related to implementing SBI (intervention characteristics), negative beliefs and the lack of knowledge (characteristics of the individuals), and the lack of self-efficacy (characteristics of the individuals). Notably, knowledge or lack of knowledge was cited as the most common facilitator or barrier. Knowledge as a facilitator contains several aspects, such as education on how alcohol influences health, familiarity with screening instruments, and counselling or intervention skills [24,59,85]. On the other hand, health professionals expressed difficulty in defining what is healthy drinking in different studies [13,20,78,82]. It was argued that many primary healthcare professionals still seemed to have the outdated dichotomous model of alcohol drinkers, i.e. those with alcohol dependence and moderate drinkers [20]. They should be informed that a more complex spectrum dividing alcohol problems into more detailed subgroups has been suggested, such as hazardous and harmful drinking in addition to alcohol dependence [20]. Training was the most common facilitator in the construct-available resources. Training is important for health professionals to acquire knowledge and practice skills required to perform screening and brief intervention. Under the construct cost, time cost appeared to be a prominent barrier. In most primary care settings, most physicians and nurses have to carry out a predetermined number of consultations, tasks, and assessments per day. They have limited time for each patient. They need to set priorities for screening conditions that they have tools and knowledge, or based on patients' requests [21,59]. Lack of self-efficacy is another important barrier. One of the main reasons may be the lack of guidelines about what they should do to alcohol drinkers [20,82]. Allowing the professionals to apply simple guidelines creatively and select what was the most appropriate for their patients was related to higher self-efficacy [20,46].
Most of the factors of SBI implementation identified in this systematic review are modifiable. There are a host of implementation strategies developed by implementation researchers for modifying factors to implement an intervention. For instance, Expert Recommendations for Implementing Change (ERIC) comprises a list of 73 implementation strategies [95][96][97]. The CFIR-ERIC Matching tool was developed which provides a list of implementation strategies to consider based on the CFIR-based barriers/facilitators [98,99]. For instance, if the barrier, lack of self-efficacy, is selected, strategies such as conducting ongoing training, making training dynamic, providing ongoing consultations will be provided. For another example, if adaptability is selected, strategies such as conducting local needs assessment, identifying early adopters, and making tailoring strategies will be provided. Moreover, when looking at how factors were studied according to the constructs in the CFIR framework, we observed that factors related to the inner setting and characteristics of individuals were extensively studied whilst the process received the least attention. Constructs in the inner setting and characteristics of individuals mainly focus on the assets of the organization and individuals, respectively. Researchers have put lots of effort into studying how the assets, such as available resources, knowledge, and training, affected the implementation of alcohol screening and brief intervention. Nonetheless, the role of the dynamic process of implementing SBI is generally neglected in previous studies, such as the process of planning, how different types of leaders engage in the implementation, as well as how to execute and evaluate the implementation process. It is conceivable that the consistently low uptake rate of SBI might be due to the lack of understanding of factors related to the implementation process. Therefore, future studies should examine the execution and processes of the SBI implementation.
To our knowledge, the CFIR was used by one systematic review to synthesize factors of implementing enhanced recovery pathways [100]. Similar to our findings, that systematic review suggested that more effort should be put into the implementation process in future studies. This conveys a message that the neglect of the implementation process may be one of the main reasons for the low uptake rate of many evidence-based practices in real settings. Therefore, using the CFIR to do systematic reviews of facilitators and barriers can help us gain a thorough understanding of implementation research .
There are three limitations in this systematic review. Firstly, we could not investigate whether there is a difference in facilitators or barriers between different types of health professionals, such as physicians and nurses, due to a relatively small number of studies involving nurses in this review. Secondly, different health systems or settings might induce different facilitators or barriers. Nevertheless, there is not enough information in the studies for stratification into different health systems or to the public or private sector. Since administration, culture, and management might be very different in different systems or sectors, it is possible that factors affecting the implementation of SBI might also be different. Thirdly, although the CFIR is comprehensive, it might not be able to cover some facilitators/barriers (e.g. stigma).

Conclusions
Most literature focused on various kinds of available assets to implement SBI. To promote the spread of SBI implementation, more high-quality studies on the implementation process are needed. This systematic review could serve as a reference framework for health authorities to devise