Factors Impacting on Development and Implementation of Training Programs for Health Professionals to Deliver Brief Interventions, with a Focus on Programs Developed for Indigenous Clients: A Literature Review

This paper reviews the literature on evaluations of brief intervention training programs for health professionals which address one or more lifestyle factors of chronic disease to identify factors impacting on development and implementation of programs. A search was conducted of the literature evaluating brief intervention training programs from 2000–2019 in the databases: Medline, CINAHL, Psychinfo, Academic Premier, Science Direct, Ovid (Including EMBASE and Healthstar), Web of Science and Informit. The content analysis and data extraction were aligned to the domains in the Consolidated Framework for Implementation Research (CFIR) to assist in the narrative synthesis. The search identified eight evaluations of programs targeting multiple risk factors, and 17 targeting single risk factors. The behavioural risk factor most commonly addressed was smoking, followed by alcohol and drug use. Programs consisted of face-to-face workshops and/or online or distance learning methods. Facilitators included availability of sustainable funding, adapting the program to suit the organisation’s structural characteristics and adoption of the intervention into routine client care. For Indigenous programs, the use of culturally appropriate images and language, consultation with Indigenous communities, and development of resources specific to the communities targeted were important considerations.


Introduction
The increasing prevalence of chronic disease is a global issue affecting all countries and populations. Tobacco use, foods high in saturated and trans fats, salt, and sugar (especially in sweetened drinks), physical inactivity, and the harmful consumption of alcohol cause more than two-thirds of all new cases of chronic disease [1]. Whilst there have been significant gains in the management of chronic diseases in recent decades in Australia, they continue to be the leading cause of mortality and morbidity [2]. The exposure to risk factors and the prevalence of chronic disease are not uniformly distributed. For example, Australian Aboriginal and Torres Strait Islander people (referred to as Indigenous Australians in this paper) suffer a higher burden of disease compared to other

Methods
This review searched electronic databases and other sources from 10 February 2018 to 07 March 2018 initially, and the search was updated on 27 May 2019. In this review we included all types of research studies: case-control, case studies, randomised controlled trials, qualitative and mixed methods studies.

Inclusion and Exclusion Criteria
Articles were included if they were: evaluations or studies of BI training programs that trained health professionals to deliver BIs to modify one or more health behaviours of general and Indigenous populations, and if they were published from 1 January 2000 to the search date. The review considered a range of locations where the training occurred, including hospitals and community clinics. Articles were excluded if they were: evaluations of delivery of BIs rather than of a training program, conducted as part of a university curriculum or on students (as the latter do not shed light on actual practice implementation), not published in the English language, dissertations, discussion pieces, editorials, and opinion pieces, systematic reviews, and meta analyses.

Search Procedures
To locate literature that specifically evaluated BI training programs for health professionals which targeted behavioural risk factors of chronic disease, a set of search terms was developed. Trial searches were conducted before finalising the search strategy. The search terms were grouped as follows: group 1 consisted of: evaluat* OR assess* OR analys* OR measur* apprais*, group 2 consisted of: brief intervention OR behaviour* intervention OR motivational interview* OR counsel* OR intervention OR interview*, group 3 consisted of: train* OR program* OR activity OR deliver* OR support OR research OR study, and group 4 consisted of four sections: (a) smok* OR tobacco OR nicotine OR smok* cessation OR smok* control OR tobacco control, (b) nutrition OR diet OR obesity OR weight, (c) physical activity OR exercise, and (d) alcohol OR grog OR substance OR drug*.
The search was conducted in two phases. Firstly, eight scientific electronic databases, Medline, CINAHL, Psychinfo, Academic Premier, Science Direct, Ovid (Including EMBASE and Healthstar), Web of Science and Informit, were searched. Secondly, to identify any grey literature, a web search was undertaken in the same period in the Cochrane database, and the websites and publications of the following organisations: Australian Institute of Health and Welfare (AIHW), The Lowitja Institute, Australian Indigenous Health Infonet, The National Aboriginal Community Controlled Health Organisation (NACCHO), Menzies School of Health Research, and health departments of Australian federal, state and territory governments. The search for grey literature was mainly limited to Australia due to constraints of time and resources.

Review Processes
The combined searches of all databases and the web search located 14,972 articles, of which 5776 duplicates were removed. The 9196 articles located from the literature search were screened through a three-step process. Firstly, the lead researcher (SS) assessed the articles by examining the title to verify its fit with the inclusion/exclusion criteria. From the screening of 9196 articles, 245 articles were identified to progress to the next step. Secondly, these 245 articles, based on a screening of their titles and abstracts, were checked against the inclusion criteria. A total of 217 articles were selected from this review process.
As the third step, full texts were retrieved for the articles selected in the second step, and the lead researcher (SS) and co-researcher (VM) conducted a review by reading all full text articles. Any disagreements between the lead researcher and the co-researcher on the inclusion/exclusion of articles were referred to a third reviewer (FC). From these processes, 41 articles were identified for inclusion in the literature review. Studies that were reviews of BI training kits, reports of training and development without details of the training component and training conducted as part of a university curriculum or on students were excluded. Sixteen studies were excluded in this process, resulting in the final count of 25 articles for the literature review. The search results are presented in the PRISMA flow diagram ( Figure 1).
To synthesise and analyse the literature to identify factors impacting on training program implementation, this review applied the five domains of the Consolidated Framework for Implementation Research (CFIR) [12]. CFIR's five domains are classified into 26 constructs and 13 subconstructs. The 26 constructs within the five domains of the CFIR are shown below in Table 1. CFIR, as a conceptual framework, can guide systematic assessment of multilevel implementation contexts to identify factors that might influence intervention implementation and effectiveness [13], and to understand potential barriers and facilitators in preparation for implementing an innovation [12]. The CFIR framework was selected because the domains of the CFIR broadly address key aspects of program implementation, and the framework has been extensively applied in other program evaluations [14,15]. Study implementation aspects were categorised according to the CFIR domains: the characteristics of the intervention, outer setting, inner setting, characteristics of the individuals, and implementation processes.

Results
Nearly half of the 25 retrieved studies (12) were conducted in the United States (US), eight in Australia and three in the United Kingdom. Table 2 displays the key characteristics of the studies included in this review. Most studies/programs (15/25) were conducted solely for research purposes, and not as operational programs. These studies were mainly descriptive. The operational programs, however, provided deeper analysis through measuring changes in trainees' skills and competence, and included discussions about the contextual factors that impacted on the implementation.   Eleven studies specifically looked at the broader aspects of program implementation by analysing the factors that supported or hindered the processes of training: the program's fit with the current practices of trainees, policies and protocols in place within health services to support the training program and the use of systems for recording and reporting BI delivery. Three studies evaluated client/patient satisfaction with the brief intervention they received.
Outcome measures in the different papers varied considerably, and they broadly examined the impact on the trainees in the areas of satisfaction with the training and improvements in confidence, motivation, skills and knowledge pre-and post-intervention, cost of the intervention, implementation of brief intervention with clients, trainees' perceptions of role adequacy and self-efficacy, clients' satisfaction with the brief intervention provided, reported changes in clients' engagement in the risk behaviour, access to resources, and exploration of policies and protocols within health services to support the activity. Three studies conducted audits of client clinical charts to determine the quantity and quality of chart recordings of brief interventions [23,31,36]. In one study, reductions in baseline screening test scores for alcohol and smoking were observed in a 3-month follow-up [21]. A high proportion of studies reported on programs that specifically targeted Indigenous populations (n = 4, 16%) and they were all from Australia (50% of the Australian studies). Evaluations of BI training programs targeting Indigenous populations that were available from other countries were not related to assessing change in behavioural risk factors of chronic disease, and hence were not included in this review.
Of the studies identified in this review, there were six programs that had more than 200 trainees (14,762, 1051, 1020, 519, 488 and 441), nine programs had between 51 and 200 trainees, 10 programs had under 50 trainees; while one program did not specify the trainee numbers. The program with the largest number of trainees, the 'Country-wide distance training for delivery of screening and brief intervention for problematic substance use' [21] conducted in Brazil trained 14,762 health workers through a distance learning program. The other three large programs were conducted in Australia with the common objective of training health professionals working with Indigenous people to provide smoking cessation advice. The SmokeCheck Program [33] in the state of Queensland trained 441 health workers and the NSW SmokeCheck Program [37] in the state of New South Wales trained 519 health workers. The Quitskills program (2019) implemented in all eight states and territories of Australia [30] trained 1020 health workers.

Factors that Impacted on the Implementation of the Program
The five domains and the 39 constructs and subconstructs of the CFIR were used to categorise the factors that impacted on the implementation of the different studies (Tables 3 and 4). In conducting the review, we rigorously checked if there were any factors that influenced the implementation of the intervention that did not come under any of the constructs of the CFIR. No factors were identified which could not be classified to one of the CFIR constructs.

The CFIR Domain of Intervention Characteristics
The domain of intervention characteristics relates to the features of the intervention that impact on its adoption and application. These features are classified into eight constructs: Intervention Source, Evidence Strength and Quality, Relative Advantage, Adaptability, Trialability, Complexity, Design Quality and Packaging, and Cost.  [19] Wales, UK.    A construct of this domain that was most frequently addressed in the different studies was 'Adaptability', which signifies the ability of an intervention to be tailored to meet local needs. For example, programs that targeted Indigenous clients developed culturally appropriate resources, and programs that trained a mix of health workers made adjustments to the curriculum to suit the trainee population from different disciplines and cultural backgrounds, and with varying levels of education [21,30,33,37]. The findings of these studies suggest that adaptability of the programs was perceived to have supported their adoption by different health services, and the benefit of such adaptability in the application of the programs was highlighted in the studies.
Programs that used resources developed by educational institutions or universities [19,25,40] were perceived to be of high credibility by the participants. Although many of the programs were supported by government funding, none of the studies reported specifically applying government guidelines in the development of resources. Four training programs incorporated online components, either as standalone or to supplement face-to-face workshops [19,21,26,36]. For the conduct of large-scale programs, particularly those that spanned large geographical areas, web-based programs were found to be beneficial [21] in facilitating learning although they were mainly used as an additional component to face-to-face training [16,19,36].

The domain of Outer Setting
The domain of outer setting incorporates four constructs: patient needs and resources, cosmopolitanism, structural characteristics, and external policies and incentives. The construct that was most often mentioned, in ten studies [18,21,23,27,28,30,31,33,36,37] was 'Patient Needs and Resources' that is defined as the extent to which patient needs, as well as barriers and facilitators to meet those needs, are accurately known and prioritized.
The influence of the external policies was evident in the development of some of the Australian programs, which were based on a number of national and state level strategies that broadly advised that any intervention designed to address smoking must be tailored to disadvantaged groups and must encourage behaviour change. The materials and resources used for these programs were perceived to be culturally appropriate to Indigenous populations [30,37].
The effect of external supports in the form of incentives and funding on the maintenance of a program was discussed in a few studies by [9,30,33,37]. The Queensland SmokeCheck Program was funded through the state health department, and funding availability was a major enabler for the NSW SmokeCheck Program. The availability of funding supported the conduct of training at multiple sites and also covered the costs of travel [37]. While the funding for the SmokeCheck Programs was not continued from the early 2010s, the South Australian Cancer Council's Quitskills Program [30] has had continued funding from the Australian Government. The study by Carneiro et al. [21] on another large program, the 'Country-wide distance learning program' in Brazil, does not explicitly state the funding source, but the support of the government is implied. This program developed a significant quantity of course material that could be disseminated through online platforms. In summary, the sustainability of funding and the extent to which patient needs are addressed are important considerations in the development and implementation of BI training programs.

The Domain of Inner Setting
The domain of Inner Setting includes five constructs: Structural Characteristics, Networks and Communications, Culture, Implementation Climate, and Readiness for Implementation. For operational programs conducted over large geographical areas, distance was a major structural characteristic that impacted on the implementation. In the SBIRT (screening, brief intervention, referral and treatment) training in the state of New Mexico in the United States [9], two clinical supervisors were engaged to travel around the state of New Mexico, to train behavioural health counsellors. For the Brazilian distance learning program in 'Screening and BI' [21], the number of professionals proposed to be trained in the program was large, and most sites were in small towns, distant from main health centres. Consequently, web-based and distance learning program resources were used, which suited the structural characteristics of the multiple organisations that were spread across the country. However, as these methods were new to most participants, the project employed strategies such as: constant telephone support, online discussion forums and use of call centres to support participants.
Another large operational program, the Queensland Smokecheck program [33], was conducted through one day face-to-face workshops in 31 locations covering urban, regional and remote areas of the state of Queensland in Australia. The NSW Smokecheck program [37] consisted of participation in a one-day training workshop, conducted in 46 locations for the health professionals, and another 17 workshops for managers, spread over the entire state of NSW in Australia. This project collaborated with health services so that they were involved in the organisation of training workshops. The role of Aboriginal Health Workers as mediators between service providers and communities was acknowledged in this training program [37].
The construct 'Readiness for implementation' was referred to eight times. This construct consists of three sub-constructs: Leadership Engagement, Available Resources, and Access to Knowledge and Information. Some of the challenges experienced in the recruitment of training participants for different programs, as reported, for example, in the evaluation of the NSW SmokeCheck Program, were: small staff numbers for some organisations; some organisations did not feel they fitted in with the program; the reluctance on the part of managers and senior health professionals to attend the program; and the lack of managerial support for the implementation of the program [37]. Barriers to implementation of the programs (post training) included differing perceptions of managers about managing the intervention, particularly adoption of the intervention into routine client care, gaps in health professionals' knowledge and skills impacting on their ability to deliver brief interventions, lack of clear instructions, and a single framework to follow [37]. Carneiro et al. (2017) observed that lack of time, infrastructure or support from service managers, as well as competition with other work activities limited the availability of many professionals to participate in the study processes.

The Domain of Characteristics of Individuals
Out of all five of the CFIR domains, the influence of the characteristics of the trainees on the implementation of the programs has been the least evaluated area. This domain consists of five constructs: Knowledge and Beliefs about the Intervention, Self-efficacy, Individual Stage of Change, Individual Identification with Organization, and Other Personal Attributes. Several programs trained multiple health professionals: nurses, health workers, general practitioners, specialist doctors, and alcohol and community workers. However, none of the studies evaluated how the program's impact varied according to the professional or personal characteristics of the participants.

The Domain of Implementation Processes
This domain consists of the following constructs: Planning, Engaging, Executing, and Reflecting and Evaluating. The construct of 'Planning' was referred to seven times and 'Engaging' was referred to five times. The study by Gonzales et al. [9] describes the benefit of utilising the services of a non-profit organisation (SDCCHP), as they were able to work directly with a network of service providers across the state with minimal bureaucratic obstacles, acting as an organisational champion. However, there were no studies in this review that specifically examined the use of individual opinion leaders or evaluated the impact of formally or informally appointed champions other than brief mentions in two studies [17,38]. The study by Harris et al. (2005) [28] notes the importance of the establishment of local planning for the successful implementation of the program. However, they also state that due to this study partnership not being at the most senior levels of the local area health services, the planning of referral pathways in some areas was not as effective as in others [28]. The incorporation of a training session for managers was adopted by a number of interventions in this review on the presumption that implementation, as described above, is influenced by senior staff and systems [27,37].

Discussion
This review identified that the evaluations of large operational programs such as the Queensland and NSW Smokecheck Programs, the South Australian Cancer Council's Quitskills Program, Brazil's Country-wide distance training program and the New Mexico SBIRT project had the potential to probe into the multiple factors that impacted on the implementation of these programs [9,21,30,33,37]. These evaluations and two small operational studies [18,27], and three non-operational studies [28,31,38] provided insights into development of their programs and their resources, and how collaborations were entered with other organisations. They also discussed some of the barriers and facilitators to implementation. The review has shown the importance of: local planning for the successful implementation of the program, ensuring that program partnerships are at the most senior levels of the local area health services, the presence of an opinion leader or a champion, clear instructions and a common framework, adoption of the intervention into routine client care, ensuring sustainability of funding, involvement of reputed organisations in the development of the programs and flexible training methods such as face-to face training and distance learning methods including online learning to meet the characteristics of the organisations.
We know from previous studies that the characteristics of the 'Inner Setting' significantly influence implementation efforts. Although professional training in the delivery of brief interventions is usually delivered by external providers, its implementation requires active collaboration with different health services. The size and spread of the health service, the structure and duration of the training program, the materials and resources developed, and the program's requirement to cater for a mix of health professionals are important considerations in its implementation [41,42]. We also know that systemic barriers of lack of time of health services and personnel, competition with other pressing healthcare needs, and lack of policies and systems in place within the participating organisations are barriers to implementing BI training programs [10,43]. Attitudes towards the behaviour targeted, lack of structural and organisational support, unclear role definition as to responsibility in addressing the behaviour with clients, and fears of damaging professional/ client relationships are also known to be barriers to intervention implementation [42,43].
This review too, based on the number of references to the different CFIR domains, confirmed that the characteristics of the health service (in CFIR terms, the inner setting) play a significant role in the implementation of the program. Besides establishing partnerships at senior levels with the participating organisations from the start, the importance of extending the program or tailoring the training to suit the managers and supervisors of the organisations was identified as helping to ensure better adoption by the health services [37]. This review also touches on the sustainability of the skills gained by health workers through BI training. To address this issue, the review has identified efforts such as: flexible ongoing training methods, in-person site visits, training materials designed specifically for the unit settings, monthly calls or in-person booster sessions to enhance ongoing implementation efforts [39]. For programs that are developed for Indigenous populations, the review affirms the importance of addressing specific client needs and consultation with Indigenous communities [33,37]. Besides considering the characteristics of the client population in the development of the course content and curriculum, it is also important to tailor the course to each community as much as is practically possible [30,33].
In the area of delivery of brief interventions by health professionals to their clients, the availability of resources (learning materials, assessment tools and referral guides) and how they were managed within the organisation, and the program's fit with the structural characteristics of the organisation were important [39]. BI delivery needs to be developed as a systems-based practice, with BI and referral to treatment services ideally being active processes.
In addition, a review of previously conducted systematic reviews highlighted similar themes and identified the importance of follow-up training sessions and/or training events spread over a longer time period [44]. Whilst the use of objective observational tools for evaluating the fidelity and quality of brief interventions is identified, there is also consensus that formal learning is only one aspect of learning BI skills and that informal learning that occurs in everyday counselling in clinical practice is equally important. The need for further qualitative research was noted in several studies [41,43]. Studies also identified the association between implementation effectiveness (material utilisation, screening, and BI rates) and the intensity of the implementation effort, i.e., the amount of training and/or support provided [42].

Strengths and Limitations
The application of the CFIR in this review has been highly beneficial in guiding the assessment of multilevel implementation contexts to locate factors that influenced implementation and effectiveness of the interventions, particularly in identifying the non-explicit descriptions of barriers and facilitators from the different articles. The studies included in this review differed significantly in terms of the settings, the size and nature of trainee populations, modes of interventions, program content, and outcome measures, and thus provided a comprehensive picture of the intervention environment.
However, most studies extracted in this review were about small programs with few participants, delivered across few sites (usually one site), with a very short period of operation, and usually targeting one health behaviour, and hence conducted little evaluation to gain insights into the factors that impacted on the implementation of the programs. Similarly, although BI programs for multiple health behaviours were reviewed in eight studies, none of the studies specifically identified or described the challenges of conducting multi-behaviour change training or implementation of BIs for multiple risk factors. Whilst key factors that impact on the implementation of brief intervention training programs for the general population were well identified, there was limited information on factors specific to the Indigenous populations. Only a small number of broad factors such as: consulting the communities prior to development of the program, tailoring the program to match the participant population and ensuring the cultural suitability of the programs and resources were identified. Limited research has examined the most effective methods for delivery of BI training to Indigenous health workers who work with Indigenous clients. Similarly lacking is an understanding of the extent of decline in the skills and knowledge that trainees acquired from a BI training program, although two studies reported that participants felt increasingly less legitimate and confident as time elapsed [22,33]. We also have little understanding of the nature and frequency of refresher or ongoing programs to address such gaps. The funding issues associated with refresher programs and the practical challenges of implementing them have also not been covered.

Implications for Future BI Training
The Final Report on NSW SmokeCheck stated that 'the results point to the need for organisational development on the part of health services to ensure that smoking cessation becomes a standard component of Aboriginal Health Workers' and other health professionals' clinical and health promotion roles. The report points to the need for health services to provide organisational support to trainees, highlighting the role of health services' managerial staff in supporting the implementation of BI training programs. This suggests the need for all future BI training programs to incorporate specific training components for managers. The health practitioners' perceptions about their knowledge and their right to work in the identified area(s) are important in the ongoing support for the program, and the resultant delivery of BIs. Hence, the design of the programs needs to consider the trainees' fit with the program: their educational and professional backgrounds and build on their role legitimacy and adequacy. This is particularly important consideration for managers, as the responsibility for preventive care is often perceived to be on front line health professionals rather than on the whole health service.
To increase the accessibility of the program to health workers located in rural and remote locations, the need for the availability and sustainability of funding is highlighted. It is also important to consider alternate modes of delivering the training such as web-based modules and distance learning. Gaps in health professionals' knowledge and skills continue to exist even after participation in the training program, and erosion in motivation and skills is likely to happen over time, impacting on their ability to deliver brief interventions. To address these concerns and to address the high turnover of staff members, routine repeat courses and refresher programs suitable to the trainee population need to be available. In the context of BI delivery for Indigenous clients, the role of Aboriginal Health Workers as mediators between service providers and communities needs to be acknowledged in the training program [37]. Whilst it is important to develop culturally appropriate resources and learning materials, it is also important to tailor the course to each community as much as is practically possible, particularly to be inclusive of regional cultural and language differences.

Conclusions
The identification of common contextual factors that impacted on the implementation of the program has been difficult due to the varied nature of the studies. However, the descriptions of the development and implementation of the programs do provide some information on the factors that favoured or hindered the different processes. The most commonly discussed CFIR domain in the different evaluations was the influence of the inner setting or the characteristics of the participating organisations. The application of the CFIR has been particularly beneficial in identifying and teasing out the not so obvious descriptions of the contextual factors from the different articles. A highlight of this review is the generation of the knowledge that the literature on BI training programs for health professionals is generally deficient in understanding the complex interplay of contextual factors that impact on the implementation of the programs.
Key messages from the review for the development and implementation of future BI training programs include: ensuring supports from management and engagement of workers from the start and incorporating the intervention as a standard component of health professionals' roles. The availability of ongoing funding and adapting the program to meet the structural characteristics of the organisations are also important. Subsequent to the conduct of the training program, ongoing implementation of the intervention needs to be incorporated into routine practice aligning with the operational procedures of the organisation. This review highlights that healthcare providers can attain brief motivational interview knowledge, skills and confidence relatively quickly and these can be sustained over time, however, further research is needed to determine if BI skills are long lasting. Finally, to comprehensively understand the factors that impact on the implementation of BI training programs, conduct of in-depth qualitative studies to evaluate specific training features and methods that work best is recommended. It is expected that the results of such a study would provide deeper insights into facilitators and barriers of organisational implementation of BI training programs and its ongoing delivery.

Conflicts of Interest:
The authors declare no conflict of interest.