Effectiveness of Motivational Interviewing on adult behaviour change in health and social care settings: A systematic review of reviews

Background The challenge of addressing unhealthy lifestyle choice is of global concern. Motivational Interviewing has been widely implemented to help people change their behaviour, but it is unclear for whom it is most beneficial. This overview aims to appraise and synthesise the review evidence for the effectiveness of Motivational Interviewing on health behaviour of adults in health and social care settings. Methods A systematic review of reviews. Methods were pre-specified and documented in a protocol (PROSPERO–CRD42016049278). We systematically searched 7 electronic databases: CDSR; DARE; PROSPERO; MEDLINE; CINAHL; AMED and PsycINFO from 2000 to May 2018. Two reviewers applied pre-defined selection criteria, extracted data using TIDIER guidelines and assessed methodological quality using the ROBIS tool. We used GRADE criteria to rate the strength of the evidence for reviews including meta-analyses. Findings Searches identified 5222 records. One hundred and four reviews, including 39 meta-analyses met the inclusion criteria. Most meta-analysis evidence was graded as low or very low (128/155). Moderate quality evidence for mainly short term (<6 months) statistically significant small beneficial effects of Motivational Interviewing were found in 11 of 155 (7%) of meta-analysis comparisons. These outcomes include reducing binge drinking, frequency and quantity of alcohol consumption, substance abuse in people with dependency or addiction, and increasing physical activity participation. Conclusions We have created a comprehensive map of reviews relating to Motivational Interviewing to signpost stakeholders to the best available evidence. More high quality research is needed to be confident about the effectiveness of Motivational Interviewing. We identified a large volume of low quality evidence and many areas of overlapping research. To avoid research waste, it is vital for researchers to be aware of existing research, and the implications arising from that research. In the case of Motivational Interviewing issues relating to monitoring and reporting fidelity of interventions need to be addressed.


Introduction
There is overwhelming epidemiological evidence that health behaviour such as smoking, substance abuse (drugs and alcohol), physical inactivity, and unhealthy eating are associated with increased morbidity and mortality. The cost to the UK NHS for diseases associated with poor diet, physical inactivity, smoking, alcohol and obesity are estimated to be in excess of £12 billion [1]. The challenge of addressing unhealthy lifestyle choice is complex and requires sustained behaviour change. The UK NICE (2014) guidelines [2] recommend a range of behaviour change approaches, guided by a taxonomy of interventions [3], aimed at changing health-related behaviour of individuals, communities or whole populations.
Motivation to change is a key component of the behaviour change process as it guides and maintains goal-related behaviour [4]. One approach to change motivation and subsequent behaviour is Motivational Interviewing, introduced by William Miller in 1983 to help people with alcohol problems change their drinking behaviour [5]. The approach was developed further in the 1990s into "A collaborative conversation style for strengthening a person's own motivation and commitment to change" [5]. Motivational Interviewing aims to explore and resolve ambivalence that people might have about health behaviour in favour of change. It encourages people to say why and how they might change and pertains both to a style of relating to others and a set of skills to facilitate that process. The four overlapping processes involve: 1) engaging in a working relationship; 2) focusing on a problem to change; 3) evoking the person's desire to change; 4) planning the change [5]. In 1997 an international organisation of trainers established 'The Motivational Interviewing Network of Trainers (MINT)' with an aim to improve the quality and effectiveness of counseling and consultations for professional delivering Motivational Interviewing. The organisation has grown to represent 35 countries and 26 languages, which demonstrates the global popularity of this intervention. Some reviews report positive outcomes for Motivational Interviewing and suggest it could be useful for a wide range of behavioural and health problems [6][7][8][9] whilst others are more cautious in their conclusions and recommendations [10][11][12].
Many different health care professionals and other groups are using behaviour change interventions including Motivational Interviewing to help people change or adapt their behaviour. However, it is unclear for which behavioural problems and populations Motivational Interviewing is most beneficial, or in some cases, where there is evidence of no effect or possible harm. This overview aims to identify, appraise and synthesise the review evidence for the effectiveness of Motivational Interviewing on health behaviour of adults in a wide range of health and social care settings to answer the following question; What is the strength and quality of the current evidence to support the use of Motivational Interviewing to change adult behaviours in health and social care settings?
This question is important to guide health care professionals, researchers and other stakeholders to the most effective and worthwhile interventions for patients.

Design
We conducted a systematic review of existing reviews (referred to as an overview [13]). An overview synthesises the evidence from more than one systematic review at a variety of different levels, including the combination of different interventions, different outcomes, or people from different populations with different conditions.

Search methods
We systematically searched the following electronic databases from January 2000 to 28th May 2018; Cochrane Database of Systematic Reviews (CDSR); Database of Reviews of Effects (DARE); PROSPERO (an international prospective register of systematic reviews); MEDLINE; CINAHL; AMED and PsycINFO. The search string was adapted for each database. (See Appendix 1 for Medline search). A comprehensive search combined key terms using Boolean operators (e.g. AND, OR) for: Intervention (e.g. "motivational interviewing," "motivational enhancement") and Review type (e.g. "systematic review," "meta-analysis, " "review literature, " "qualitative systematic review," "evidence synthesis" OR "realist synthesis", "qualitative AND synthesis", "meta-synthesis � OR meta synthesis � OR metasynthesis", "meta-ethnograph � OR metaethnograph � OR meta ethnograph � ", "meta-study OR metastudy OR meta study"). Truncated forms of these terms and alternative spellings were included. To be eligible for inclusion, reviews met the following criteria: Inclusion criteria.
• Reviews using structured, pre-planned methods to synthesise research studies addressing a clearly defined topic or research question (which could comprise either quantitative, qualitative or mixed methodology) • Published from January 2000 • Interventions described as Motivational Interviewing or Motivational Enhancement Therapy (MET) delivered in any format (e.g. face to face, online, group, text or telephone) • English language • Interventions focused on adults.
• Letters, commentaries, expert opinion, theoretical and "non-systematic" or unstructured reviews e.g. reviews without an aim that did not clearly describe the search strategy, selection criteria and quality assessment employed.
• Reviews focused solely on children and adolescents under the age of 18 years • Reviews focused on Motivational Interviewing intervention to change professional or organisational group behaviour.
• Reviews focused on combined psychological interventions e.g. Motivational Interviewing combined with Cognitive Behavioural Therapy.

Identification of studies
Members of the review team (PC / SM) ran the search strategy and then examined all titles to exclude clearly irrelevant papers. Two reviewers (PC and HF) independently reviewed the abstracts of all potential records identified from the electronic searches and excluded those not meeting the inclusion criteria. Inter-rater reliability was assessed for agreement of abstract screening. Two reviewers (PC and HF) independently assessed full papers for all potentially relevant reviews. Full text papers ranked as irrelevant by both reviewers were excluded at this stage of the screening process. The final selection of full text papers (judged as relevant or unsure) were discussed at a consensus meeting, with a third reviewer (MM or AP) as required.

Data extraction
Three reviewers (PC, HF and EC) independently extracted the following information: review question or aims; types of studies included; characteristics of participants and numbers included; interventions details. The TIDieR framework [14] was used to guide reporting of interventions components and comparators. Two reviewers (HF and PC) checked all the extracted data and discussion between the two reviewers resolved any disagreement; with assistance from a third reviewer (AP) when necessary. A data extraction form (excel) specifically developed by the overview author team was used to collate the data.

Categorisation of reviews
Two reviewers (PC and HF) categorised each review into one of four of the following domains depending on the focus of the review. Domain 1: Stopping or preventing an unhealthy behaviour Domain 2: Promoting healthy behaviour for a specific problem Domain 3: Behaviour change for multiple health related problems and /or multiple behaviour problems Domain 4: Behaviour change in specific settings Reviews in Domain 1 and 2 were then sub-grouped by HF and PC according to the main health behaviour or problem.

Assessment of quality of reviews
Two reviewers (HF and PC) independently assessed the methodological quality of included reviews using the ROBIS tool [15]. Any disagreement was resolved through discussion between the two reviewers. The tool covers four domains to detect bias in systematic reviews relating to study eligibility criteria; identification and selection of studies; data collection and study appraisal; synthesis and findings. The full result of assessment of bias aids transparency and aims to help researchers judge risk of bias in the review process, results and conclusions.

Meta-analyses data extraction
One reviewer (PC) extracted comparative data for individual and combined outcomes from any review that included meta-analyses. Data exploring effectiveness of Motivational Interviewing as the main intervention compared with any other intervention or control was extracted. One reviewer (HF) checked the data entry.
This included the following data: Number of trials and participants in the meta-analysis; Measure of effect (e.g. effect size, mean difference, standardised mean difference, relative risk); Measure of variability (95% confidence intervals) and Measure of heterogeneity (I-squared).
Three reviewers (AP, PC and HF) checked the quality assessment of individual studies reported in the reviews and considered the results when grading the evidence. We used the GRADE (Grades of Recommendation, Assessment, Development, and Evaluation) criteria to assess whether the quality of the evidence presented in the meta-analyses was high, moderate, low or very low [16] for all available comparator data within each review. This involved judgement of risk of bias relating to study design, imprecision, inconsistency, indirectness, and publication bias [17]. In addition, one reviewer (PC) extracted any data that included exploration of moderator variables and tabulated effect size for each comparator.

Meta-analysis synthesis
For reviews including a meta-analysis two reviewers (PC and HF) independently checked the overlap in studies within all the reviews and resolved any uncertainty through discussion. We excluded data superseded by a more up-to-date review (e.g. where a Cochrane review had been updated while we were conducting the overview), or in cases where an overlapping review was conducted with the same review question, we selected the higher quality review judged using the ROBIS quality assessment tool [15]. We tabulated the intervention, comparison, outcome, number of studies and participants' data relating to effectiveness and the GRADE of evidence [18]. Using the data relating to effectiveness we noted whether there was statistically significant evidence of benefit or harm for each outcome reported in the metaanalyses, or if there was no evidence of benefit or harm (no statistically significant effect).

Narrative review synthesis
For all systematic reviews without meta-analysis data (defined as narrative reviews), we summarised key findings. We systematically documented and explored the conclusions reported by the authors of the reviews. Where these reviews included overlapping aims and outcomes, we compared conclusions; where there was a discrepancy in conclusions, we focused conclusions of the most up-to-date and highest quality reviews (judged using ROBIS) [15]. We considered whether these were in agreement with the results of any related meta-analyses reported in other reviews and focused our conclusions on the most up-to-date and high quality data.

Results
The search identified 5222 records; we screened 2852 titles and removed 2363 obviously irrelevant records after removing duplications. Two reviewers screened 489 abstracts and 235 full text articles, excluded 131 reviews and extracted data from the remaining 104 reviews. The inter-rater reliability for abstract screening was 92%. The PRISMA flow diagram (Fig 1) shows the flow of literature through the searching and screening process.
Gambling behaviour. Three reviews focused on Motivational Interviewing and psychological therapies for gambling addiction [7,30,63]. Yakovenko et al (2015) [7] identified eight trials including longer term follow up, Petry (2017) [63] reviewed trials of psychological interventions but identified only 2 trials that included Motivational Interviewing as a stand-alone intervention.

Domain 2: Reviews focused on interventions aimed at promoting healthy behaviour for a specific problem
Oral hygiene behaviour. Five reviews focused on oral hygiene, 3 compared conventional oral hygiene advice with Motivational Interviewing interventions [64,65,68]. One compared periodontal therapy alone with Motivational Interviewing and periodontal therapy combined [66], and one included a meta-analysis of psychological treatment for people with poor oral health [67].
Eating disorders. Three reviews focused on eating disorders of mainly female participants e.g. Anorexia nervosa and bulimia nervosa [10,69,70].
Weight management behaviour. Three reviews focused on changing diet and physical activity for weight management in obese adults [71,72,74] and one investigated the management of weight gain during pregnancy [73].
Management of diabetes. Six reviews focused on the management of people with diabetes. They include reviews focussed on evidence for; improving health behaviour in the management of diabetes [75], promoting glycaemic control [77] and lifestyle modifications programmes for-metabolic risk [78]. Four other reviews categorised in Domain 3 (multiple health problems / behaviours) and Domain 4 (Behaviour change in specific settings) assessed the effectiveness of Motivational Interviewing for diabetes management alongside obesity and other health related problems [71,91,114,118].
Management of neurovascular disorders and cardiovascular disease (CVD). Three reviews focused on behavioural interventions for neurovascular disorders, but the reviews only included 11 trials in total evaluating the effectiveness of Motivational Interviewing. One review investigated Motivational Interviewing for the management of activities of daily living for stroke victims, identifying one study only [81]. Hildebrand (2015) reported one of 39 trials that incorporated Motivational Interviewing into interventions to support occupational therapy for stroke victims [82]. Lee et al (2016) [83] investigated lifestyle modification, physiological and psychological outcomes for people diagnosed with Cardiovascular disease. Overall there is insufficient evidence in this group to make firm conclusions about effectiveness of Motivational Interviewing. Sexual health behaviour. Five reviews focused on promoting safe sexual behaviours [84][85][86][87][88]. Two reviews focused specifically on sexual health in gay men [84,85]. One review focused on the effectiveness of Motivational Interviewing on contraceptive use in women [87].
Adherence to medication. Adherence to medication was assessed for different populations and health problems. Hu et al (2014) assessed interventions including Motivational Interviewing to increase medication adherence in racial and ethnic minority groups [94]. Five reviews assessed medication adherence for patients with HIV [90,94,96,97,120]. Two recent reviews with meta-analyses assessed the effectiveness of Motivational Interviewing to enhance medication adherence for adults with chronic diseases and health problems [93,95].
Engagement with interventions. Three reviews focused on engagement with a specific intervention [98][99][100]; one specifically on cardiac rehabilitation. Karmali et al (2014) assessed adherence to cardiac rehabilitation but only one trial of Motivational Interviewing was identified in this review [98]. A review with meta-analysis of outcomes relating to adherence by Lawrence et al (2017) [100] investigated individuals' uptake of mental health interventions. Miller et al (2017) [99] assessed the efficacy of Motivational Interviewing to improve health screening for various problems e.g. breast screening, uptake of colonoscopy.
In addition, two other reviews grouped in Domain 1 and 2 assessed the effect of Motivational Interviewing on adherence to drug management programmes in offender populations [29] and adherence to treatment for chronic pain [102].
Management of musculoskeletal problems. Two reviews focused on musculoskeletal problem [102,103] with some overlap of trial within the reviews. In the most recent review, Alperstein and Sharp (2016) identified 7 trials focused on pain outcomes and adherence to treatment in adults with various musculoskeletal problems e.g. low back pain, rheumatoid arthritis [102].
Management of irritable bowel disorders. One review explored the use of Motivational Interviewing to improve outcomes for people with irritable bowel disorders including quality of life measures [104].
Cancer care. One review focused on Motivational Interviewing to address various lifestyle behaviours and health problem associated with cancer such as fatigue, weight problems, and physical activity participation [101].
Of the 155 meta-analysis comparisons that were extracted, we found no high quality evidence. Twenty seven comparisons provide moderate quality evidence according to the GRADE criteria. Most of this evidence was categorised in Domain 1 (Stopping an unhealthy behaviour). Further details of the outcomes for the moderate quality evidence are reported in Table 6.
Seventy one comparisons provided low quality evidence and 57 provide very low quality evidence judged by the GRADE criteria. S3 Table summarises the comparisons that were judged as providing low or very low quality evidence. The key reasons for downgrading the evidence to low or very low quality primarily relate to; risk of bias of the review was unclear; heterogeneity was judged to be moderate to high, or confidence intervals very large; volume of evidence was judged to be insufficient to support a definitive conclusion and concerns about the quality of the trials included within the comparison judged by review authors. Table 6 summarises the 27 comparisons, which provide moderate quality evidence for Motivational Interviewing interventions judged from six reviews [12,49,56,58,84,111]. Eleven of these 27 comparisons (7% (11 of 155) of all meta-analyses' comparisons) provide moderate quality evidence for mainly short term (<6 months) statistically significant beneficial effects of Motivational Interviewing. The remaining 16 comparisons demonstrate no benefit or harm, compared with a control of usual care or other active interventions. Moderate quality evidence of a beneficial effect of Motivational Interviewing was available for;

Moderate quality evidence for effectiveness of Motivational Interviewing
Alcohol use. 13 comparisons from two reviews [49,58] explored the effect of Motivational Interviewing on outcomes relating to alcohol use in mixed populations. Eight of the 13 comparisons provide consistent evidence that Motivational Interviewing has a beneficial effect on outcomes relating to the frequency and/or volume of alcohol consumption, for short term outcomes (< 4 months), but the evidence relating to sustained (>4 months) outcomes is less consistence. Comparisons relating to risky behaviour and drink driving demonstrated no benefit (or harm) of Motivational Interviewing. There is evidence of beneficial effects from one review of young adults (<25 years), for reducing binge drinking, frequency, quantity of alcohol consumption and peak blood alcohol concentration [58]. The effects are small (See Tables 5 & 6). MI may be beneficial in preventing alcohol abuse for college students and pregnant women but more high-quality research is needed to confirm results.
Further high-quality research focusing on training and competency, fidelity, delivery and dose for specific groups is justified. Windows of opportunity to reduce alcohol/drug abuse should be investigated further e.g. in specific antenatal groups and during pregnancy, in offender populations and students at transition to university Barrio  Systematic review of reviews of the effectiveness of Motivational Interviewing  Table).
There is low quality evidence for small effects for men who have sex with men on some outcomes relating to unprotected anal sex but the evidence is inconclusive. (See S3 Table) Very low quality evidence of effect of contraceptive use in the short term for some women at risk of pregnancy (See S3  Table). for small effects on medication adherence e.g. adults with chronic diseases (See S3 Table) The most promising results are for adherence to HAART medication in people who are HIV-positive, but the higher quality reviews concluded that the methodology within the trials was poor. Further higher quality research required. Systematic review of reviews of the effectiveness of Motivational Interviewing MI is an effective approach to changing behaviour. It offers promise in improving cardiovascular health status.

(Continued )
Smoking cessation. One comparison from a review on smoking cessation was judged to provide moderate quality evidence. This review comparing Motivational Interviewing with usual care or brief advice, provides evidence of beneficial effects on abstinence from smoking, particularly when attention was paid to treatment fidelity [12].
Substance abuse (drugs). One comparison from a review of people with substance abuse dependency and addiction provides evidence of a benefit of Motivational Interviewing when compared with no intervention. The other four comparisons derived no benefit or harm when Motivational Interviewing was compared with usual care or any other treatment [56].

Implication for clinical practice and research (Interpretation of authors of overview)
Kohler et al (2015) [112] To examine changes in alcohol consumption after brief MI for young people with existing alcohol use problems, who were admitted to an emergency care unit alcohol positive, with an alcohol-related trauma, or with a history of elevated alcohol consumption 6 RCTs (2 specifically over 18) n = 1433 age [18][19][20][21][22][23][24][25] Young people in emergency care who screened positively for past or present risky alcohol consumption.
BMI/ standard care, including written information (e.g. alcohol-use risk handout, educational brochure.
Alcohol consumption, frequency and quantity MI was never less efficacious than a control intervention. Two trials found significantly more reduction in one or more measures of alcohol consumption in the MI intervention group.

M-A (UNCLEAR)
Narrative reviews support the meta-analyses suggesting there is no difference in outcome between professional groups who deliver MI. High quality research assessing competency and fidelity of MI interventions is needed to confirm if any benefits reported by Merz et al (2015) are sustained over 12 months. Knight et al (2006) [118] To identify the extent to which MI has been used in different physical health settings and appraise the effectiveness of MI Prognostic markers, disease endpoints, risk reduction behaviours; physical functioning and quality of life, substance abuse, patient adherence to medical advice and patient approach to change.
The emerging evidence for MI in medical care settings suggests it provides a moderate advantage over comparison interventions and could be used for a wide range of behavioural issues in health care.

M-A (UNCLEAR)
Merz et al (2015) [113] To identify evidence to reduce alcohol use and prevent alcohol related consequences in young adults (18-24 years old) admitted to the emergency department following acute alcohol intoxication.

RCTs (n = 618)
Young adults (18)(19)(20)(21)(22)(23)(24) Physical activity. Four comparisons from a review of Motivational Interviewing for promoting physical activity participation were judged to provide moderate quality evidence when Motivational Interviewing was compared with a control or usual care. One out of the four comparisons provide evidence of benefits. No benefit was found for the other three comparisons, including outcomes for people with cardiovascular disease and obesity [111].
Sexual health. Four comparisons from one review provide moderate quality evidence of no benefit or harm of Motivational Interviewing relating to changing high risk sexual behaviours in men who have sex with men [84] when compared with a control.

Exploration of moderator variables
Of the six reviews that provide any evidence judged to be of moderate quality, three did not report the results of any subgroup analyses [56,84,111]. The three reviews that contain moderate quality evidence and report subgroup analyses are: � Lindson-Hawley 2015 [12]-smoking cessation (  Systematic review of reviews of the effectiveness of Motivational Interviewing Table 5. Summary of reviews contributing data to comparison that provide moderate, low and very low quality evidence of effects of Motivational Interviewing (MI).

Sub-groups
Reviews contributing data to overview Reviews with data, but superseded by more up-todate or higher quality review judged by overview authors using ROBIS  ) as no treatment, assessment only or non-specified treatment as usual, but Motivational Interviewing is not beneficial when compared to other 'strong' interventions.
Generalisable conclusions relating to the most effective delivery of Motivational Interviewing (e.g. face-to-face or group), dose, or characteristics of provider or patient across behavioural domains are difficult to draw.

Results of narrative reviews
Of the 104 reviews included in this synthesis, 65 did not combine any data within meta-analysis. The main findings from the narrative reviews are summarised in Tables 1 to 4. The    [10,69,70]. One review in this category judged as low risk of bias suggests that Motivational Interviewing is important in psychiatric settings for reduction of substance use in the short term.

ALL BEHAVIOURS COMBINED
Narrative reviews of physical health problems include: cardiovascular problems (Motivational Interviewing for increasing physical activity) [83,110]; musculoskeletal health (adherence with intervention for back pain) [103]; diabetes self-management (effect of smoking, blood-glucose control, diet and weight management [62,75,76,[78][79][80]; oral health hygiene [64][65][66]68] (use of dental fluoride, increasing dental utilization and reducing sugar consumption); obesity (adherence to weight loss programmes); management of neurovascular disorders [82]. The most recent reviews report outcomes for the effectiveness of Motivational Interviewing for cancer care [101] and outcomes related to the treatment of irritable bowel disorder [104].

Quality of narrative reviews
In total 20 narrative reviews were judged as low risk of bias graded using the ROBIS tool [15] [11,20,25,35,42,44,47,51,53,54,59,65,83,89,92,97,98,[113][114][115]. Five of these reviews report positive effects of Motivational Interviewing. Rueda et al (2006) found beneficial effects of Motivational Interviewing for adherence to highly active antiretroviral therapy where there appears to be promising results for interventions delivered over 12 weeks or more [97]. Taggart et al (2012) found further support for benefits of Motivational Interviewing in achieving impacts around smoking cessation compared to other group education [115]. Cooper et al (2015) reported positive results for some but not all outcomes for reducing cannabis use [42]. Noordman et al (2012) conclude that Motivational Interviewing can be effectively delivered by physicians and nurses as a face-to-face communication-related behaviour change technique [114]. Reviews published since 2016 report mixed results. Kay et al (2016) suggest that Motivational Interviewing has potential for use in oral care [65]. Chatters et al (2016) report short term benefits for reducing cannabis use in younger adults [47]. However, most were unable to make firm conclusions about effectiveness of Motivational Interviewing [20, 44,59,89]. In a review of brief non face-to-face Motivational Interviewing interventions Jiang et al (2017) found promising evidence for telephone delivery in the treatment of substance abuse, but the results were not consistent for other alternative modalities such as text messages in groups or internet-based interventions.

Discussion
This overview is the first to integrate and systematically grade the quality of the evidence for the effectiveness of Motivational Interviewing interventions across a wide range of settings and populations for people with many different health problems and diseases. We have created a comprehensive map of all reviews relating to Motivational Interviewing to provide clarity relating to an intervention for which there have been multiple overlapping (and sometimes conflicting) reviews. Conflicting review evidence can create barriers and challenges to practitioners wanting to deliver evidence-based practice. This overview provides practitioners, policy makers and researchers with a summary of the quality and strength of the evidence for Motivational Interviewing. It signposts practitioners to the most up to date reviews, enabling them to efficiently access best review evidence to support clinical decisions. We found no high-quality evidence from the meta-analysis data within any review, mainly due to methodological flaws in the reviews and poor quality of the included studies.
Motivational Interviewing appears to be most effective for stopping or preventing unhealthy behaviours (categorised as Domain 1) such as binge drinking, reducing the quantity and frequency of drinking, smoking and substance abuse. For gambling behaviour, low quality evidence of short to long-term effectiveness suggests that further research on the effectiveness of Motivational Interviewing is warranted to address this significant public health problem [62]. For promoting healthy behaviour (categorised as Domain 2) where people may have little desire to change, most of the evidence is inconclusive or of low quality. For example, there is low quality evidence for the effectiveness of Motivational Interviewing for weight loss outcomes in obese and overweight adults. The exception in Domain 2 is physical activity promotion where there is moderate quality evidence of beneficial effects of Motivational Interviewing for increasing physical activity in people with chronic health conditions. However, the trials assessing adherence to physical activity participation were small and further high quality research in this field is justified to investigate the effectiveness of Motivational Interviewing in different populations, settings and context.

Mode of delivery
The exploration of moderator variables from meta-analysis data does not provide enough data to be confident about the effects of different modes of delivery for Motivational Interviewing. Reviews that focus on the mode of delivery report inconsistent results [45,51,95,105]. The TIDieR guidelines [14] capture some of the features that are relevant to intervention delivery but the mode of delivery is considered to be an important component of intervention and is not reported consistently in the literature [121]. Recent reviews have compared telephone [51] or technology-delivered Motivational Interviewing interventions (TAMIs) [105] and report inconsistent results or no beneficial effects. For example, Shingleton et al (2016) [105] found that TAMIs are feasible to deliver but there is limited evidence of effectiveness. For an intervention that relies on building and developing a relationship between client and provider it seems unlikely that this mode of delivery could be successfully adapted for Motivational Interviewing without considerable focus on training and fidelity measures.

Implication for clinicians and policy makers
The National Institute for Health and Care Excellence (NICE) guidelines [2] include Motivational Interviewing as a component associated with some effective interventions for behaviour change strategies. However, the NICE (2014) Programme Development Group (PH49) are cautious about making general recommendations due to lack of details of intervention components reported in this field of research [2]. This overview has identified clear gaps in the evidence in support of most of the interventions categorised in Domain 2 (e.g. weight loss programmes for obesity, oral health behaviour, management of diabetes and musculoskeletal disorders, adherence to medication and engagement with interventions). The high quality reviews on smoking cessation [12] and alcohol abuse [49] both recommend caution when interpreting results. However, the overall effect size reported by Lundahl et al [108] of 0.22 (95% CI 0.17 to 0.27) is similar to other complex behavioural intervention [122,123]. If applied to the 1 million smokers in the UK, or the millions of physically inactive people globally [124], it is plausible that the impact of Motivational Interviewing on health at a population level may be larger. Further rigorous research is required to support this assumption.

Training and fidelity
Many different health care professionals including nurses, counsellors, physicians, medical students, social workers, and physiotherapists deliver Motivational Interviewing interventions, but there is little information about their training. Reviews that compared different health care providers found either no difference between groups [114] or reported limited conclusions due to small sample size [12].
Details of the fidelity of training of professionals delivering the interventions were generally poor although this is not unique to reporting of Motivational Interviewing. Training issues are fundamental to the success of any complex intervention and Motivational Interviewing, like other surgical, therapy or other behavioural interventions, requires practice of skills and a basic level of competency. There is no formal requirement for training in Motivational Interviewing or evaluation therefore practitioners can claim to use the approach without assessment, and competency is likely to influence outcome.  suggest that investment in training would need to be large to impact on change in practice [125].
It is difficult to comment on the cost effectiveness of Motivational Interviewing as it was not the focus of this overview, however we identified very little health economic data. Where cost data was available from a trial of smoking cessation in the UK, no clear conclusions could be drawn as the sustained quit rates did not reach statistical significance [12].

Strengths and limitations of the overview
This overview is the first to synthesise systematic review evidence on the effectiveness of Motivational Interviewing from a wide range of populations and settings with an aim to provide information that informs practice and policy. It highlights the discrepancy between the widespread recommendations of Motivational Interviewing as a universal behaviour change strategy and the available evidence supporting this approach. We carried out a comprehensive search with an inclusive selection criteria and it is unlikely that we missed any reviews written in English prior to our initial search, but this overview is not exhaustive.
The conclusions of this overview are highly dependent on, not only the quality of the reviews but the studies within the reviews. We extracted data according to the TIDieR guidelines [14] but many intervention details were missing, making it difficult to draw conclusions with confidence. This problem needs to be addressed in future trials to facilitate data synthesis and provide clear recommendation to all stakeholders. Our assessment of review quality (ROBIS) [15] and evidence quality (GRADE) [17] are subjective judgements and we used these judgements to categorise the evidence, concentrating our conclusions on those judged to be moderate quality (or low bias for narrative reviews). Some may consider our methods overly critical, but authors of the higher quality reviews are equally cautious with their recommendations [11,12,49].

Recommendations and implication for future research
The established Network of Trainers (MINT) alone have delivered Motivational Interviewing around the world to millions of people [126] but many questions remain unanswered regarding effectiveness.
Recommendations for clinical practice. Many different health professional groups are using Motivational Interviewing but the evidence for training reported in the literature is limited. The 'Motivational Interviewing Treatment Integrity code' (MITI) has evolved over the last 10 years [127] with an aim to standardise the delivery of Motivational Interviewing interventions. Guidelines for the minimum intervention content and training requirements for Motivational Interviewing are available and should be followed to standardise intervention delivery [127,128].
Recommendations for future reviews. This overview has identified and brought together systematic reviews relating to Motivational Interviewing interventions; however further systematic reviews are warranted to inform clinical practice and future primary research in this field. Recommendations include, but are not limited to; 1. Research should address the fact that in clinical practice Motivational Interviewing is often delivered in combination with another psychological intervention. Systematic reviews exploring combined interventions were excluded from this overview; consequently, it is important to identify and appraise any existing systematic reviews relevant to this, prior to planning new reviews or primary research.
2. Future systematic reviews would benefit from the development of a taxonomy to ensure meaningful categorisation of the delivered intervention which considers the theoretical basis for Motivational Interviewing. Meaningful categorisation of Motivational Interviewing should be central to informing clinically relevant analyses and subgroup analyses.

3.
A systematic review to explore the cost-effectiveness of Motivational Interviewing as an intervention for those health conditions where there is moderate quality evidence of a beneficial effect of Motivational Interviewing on patient outcomes.

4.
A systematic review to explore the barriers and facilitators to delivery of Motivational Interviewing, focussed on those health conditions where there is moderate or high quality evidence of a beneficial effect.

5.
A systematic review of qualitative evidence to explore the acceptability and perceptions of this intervention to people who are offered Motivational Interviewing.
6. Stakeholder involvement should be conducted in future reviews of the Motivational Interviewing literature particularly relating to categorising interventions and outcomes.
7. The use of reporting templates, recognised guidance and best practice for the conduct of systematic reviews and primary research is essential. e.g. PRISMA [129] and TIDieR [14].
Recommendations for future primary research.
1. Exploration of the effect of Motivational Interviewing should consider long-term outcomes and cost-effectiveness. Subgroup analyses should explore the length of intervention delivery and time since the end of the intervention.
2. Investment in training would need to be large to impact on change in practice [130] and this along with other issues relating to sustainability of the intervention e.g. context, should be considered in future trials.
3. To ensure avoidance of research waste [131,132] it is essential that researchers are fully aware of existing reviews before embarking on further reviews, and that critical systematic reviews of evidence are completed prior to further primary research.

Conclusion
For the health problems that Motivational Interviewing was originally developed to address such as smoking cessation and alcohol misuse, the evidence provides some support for implementation particularly if fidelity of the intervention is prioritised. However, Motivational Interviewing has been implemented already for a wide range of other health and social problems where a "one size fits all" approach has been adopted with inconsistent effects.