How To Support Smoking Cessation In Primary Care And The Community: A Systematic Review Of Interventions For The Prevention Of Cardiovascular Diseases

Introduction

Cardiovascular diseases (CVDs) are the leading cause of mortality worldwide. In 2016, they were responsible of 17.9 million of deaths, 31% of global mortality. Over 75% of deaths due to CVDs take place in low- and middle-income countries.¹ In the last 20 years, mortality in low- and middle-income countries shifted from infectious diseases to CVDs. Current age standardized mortality rates in low-income countries are higher than those in developed countries.² Additionally, global mortality by CVDs is expected to grow. By the year 2030, non-communicable diseases are projected to account for more than 75% of deaths worldwide, the majority being the result of CVDs.²

CVDs can be prevented by acting on modifiable risk factors. Tobacco consumption is a major risk factor for CVDs.^1,3 The benefits of stopping smoking have a broad evidence base. Smoking cessation has been proved to be effective in reducing CVD-related morbidity and mortality.^4,5,6,7 Smoking cessation is one of the most effective preventive measures: in secondary prevention it could reduce the risk of a new cardiovascular event by almost half.⁴

Considerable efforts have been made in recent years to control the epidemic in tobacco use. In 2003, the WHO adopted the WHO Framework Convention on Tobacco Control. It targeted national and regional organizations, providing the guiding principles required to reduce the prevalence of tobacco use. In 2008, the WHO proposed the MPOWER, a practical tool to implement this framework: monitor tobacco use and prevention policies, protect people from tobacco, offer help to quit, warn about the dangers, enforce bans on advertising, promotion and sponsorship, and raise taxes. Practical actions were price and tax measures, packaging and labelling regulations, smoking restriction in public places, advertising restrictions, provision of educational campaigns and cessation services.^8,9 In 2019, 136 countries have implemented at least one of the key interventions to reduce tobacco demand.¹⁰

Despite recent policies, smoking is still a concern worldwide. 1.1 billion people are active smokers in 2019, 80% of them living in low- and middle-income countries. Total tobacco-attributable deaths increased from 5.4 million in 2004 to more than 8 million in 2019, representing the leading cause of death, illness and impoverishment.^10,11 Tobacco use contributes to poverty via the direct cost of purchasing tobacco, the health care costs for tobacco-related diseases, as well as the lost human capital that results from morbidity and mortality.^10,11 In high-income countries, the number of smokers is still high. In Europe, in 2016, 28.7% of the overall population smoked and tobacco was the leading cause of preventable death and disease in the UK.^12,13

SPICES (Scaling-up Packages of Interventions for Cardiovascular disease prevention in selected sites in Europe and Sub-Saharan Africa) is an international implementation research project on CVDs non-pharmacological primary prevention. Different settings are involved in high and low- middle-income countries: France, United Kingdom, Belgium, Uganda and South Africa.

SPICES interventions will target the individual and the community level. Concerning smoking, this is the less implemented measure of the MPOWER: in 2019, only 23 countries were providing cessation services at the best-practice level.¹⁰

First step was to identify non-pharmacological effective interventions for CVDs primary prevention.¹⁴ The aim of this work was to identify evidence-based non-pharmacological effective smoking cessation interventions and their implementation strategies for use in a primary care setting and/or at a community level.

Materials And Methods

A systematic review of international CVD prevention guidelines and national guidelines of each participating country was carried out between September 2017 and January 2018 following the PRISMA statements criteria and using the ADAPTE procedure.^15,16 The ADAPTE procedure provides a systematic approach to ensure quality standards for guideline development, evaluation and implementation.¹⁷ TRIP (Turning Research Into Practice) and G-I-N (Guidelines International Network) databases were searched for international guidelines. A purposive search for national guidelines was carried out using national health authority websites: the Haute Autorité de Santé for France; the Tijdschrift Huisarts and Nederland Huisartsen Genootschap guidelines for the Flemish part of Belgium; the National Institute for Care and Excellence (NICE) for United Kingdom; the European Society of Cardiology (ESC) within European countries and the WHO for Uganda. The South African researchers ensured that ESC guidelines were used in South Africa. Databases were searched for guidelines focusing on CVDs prevention and/or management of CVD risk factors, such as diabetes, hypertension, smoking, sedentary lifestyle, unhealthy diet, excess weight and obesity. Detailed research equations for each database are available in Annexure 1.

Guideline titles and summaries were analyzed for eligibility. Inclusion criteria were guidelines focusing on CVD primary prevention and/or management of CVD risk factors in an adult general population and to be applied in primary care or in a community setting. Only those published after 01/01/2012 were included. If different versions of the same guideline were found, only the most recent revision was included.

Exclusion criteria included; exclusive focus on specific populations (geriatric, infantile, socially deprived, etc.) or on secondary or tertiary prevention, exclusive focus on cardiovascular risk assessment, on pharmacological or surgical interventions, or on management of acute disease. Guidelines on specific conditions (eg, type 1 diabetes and familial hypercholesterolemia) were excluded as were those published before 01/01/2012 and not updated. The consortium agreed on this date in order to focus on recent guidelines with the most recently updated data. The research team considered that guideline authors would have reviewed and included literature published before 2012, so that no relevant data would have been missed.

Only guidelines published in English, French or Dutch were included. The lack of free full-text availability was an exclusion criterion as the research team assumed that a clinical practice guideline should be freely available. Two researchers completed this search independently, merging at the end.

Following the ADAPTE procedure, screened guidelines were eligible for quality evaluation with the AGREE II instrument (Appraisal of Guidelines for Research & Evaluation).¹⁸ AGREE II is a tool that assesses the quality of guidelines by evaluating their methodological rigor and transparency. It consists of 23 key items organized within 6 domains followed by a global rating (overall assessment [OA]). Each item is rated on a 7-point scale (7 being the best possible quality).

Full-text screened guidelines were evaluated independently by at least two researchers working blind and merging at the end. After discussion within the SPICES consortium, final inclusion was based on OA scores. We included guidelines with all OA scores superior or equal to 5 were excluded those with all OA scores below 5 or with OA scores around cut-off (one score of 4 and one of 5). Discrepant OA scores (more than 1-point difference and one score above 4) were discussed between appraisers and inclusion was based on a consensual decision taken with the researchers and a scientific committee of three senior researchers.

Data Extraction And Analysis

The included guidelines were analyzed, following the ADAPTE procedure, and a matrix of the extracted recommendations was created in order to facilitate the analysis.

Recommendations were included if they had an A or B level of evidence or 1++, 1+, 2++, 2+ for the NICE grading system and/or “Class 1” or “Strong” (regardless of the level of evidence). Bibliographic references which supported the included recommendations were listed to make further details on implementation strategies accessible. For guidelines without recommendation grades, references were included if they referred to randomized or non-randomized controlled trials, interventional cohort studies, or systematic reviews of such studies. Where few strategies had been described in the guidelines or where the description was too concise, the analysis of the full-text articles, from which the guidelines had been issued, was added. The aim was to fully describe the interventions used in such articles.

For these articles, inclusion and exclusion criteria were defined. Studies with a non-pharmacological lifestyle intervention for the primary prevention of CVDs in a general, adult population were included. Main outcomes had to be pertinent to primary care or commonly used for CVD risk assessment. Interventions had to be implemented in a primary care or community setting and had to be effective with a significant statistical difference. The objective of this work was the identification of effective interventions. Non-interventional studies and those on specific populations and on secondary or tertiary prevention population were excluded, as were studies in a hospital/clinic or university context or where patients were constantly monitored or supervised because that was deemed not feasible in a primary care context. Publication date was not an exclusion criterion. When different articles dealt with the same intervention they were identified, and duplicates were removed. Full-text articles were obtained on PubMed or via the University Web Library. If not available, the article in paper format was ordered via the library of the Université de Bretagne Occidentale.

Interventions and implementation strategies used in the included references were searched through the articles, especially in the method section, and described in detail and then summarized. Information was collected on the frequency and duration of the intervention, on the setting (eg, workplace or public event, individual or group sessions), on the material used (informative material, leaflets or videos), on the psychological model used and if mass media were involved. The status and background training (if provided) of those carrying out the interventions were noted. All data were collected in an excel matrix file.

Results

Guidelines selection process is shown in the PRISMA Flowchart (Figure 1).

Figure 1 PRISMA Flowchart showing the guidelines selection process.

Forty-seven guidelines were screened for quality evaluation with AGREE II tool. Full results are available in Annexure 2. Twenty-six guidelines on CVD primary prevention were included based on their AGREE II overall score (Table 1). They were to be considered among the best quality and the most pertinent for primary CVD prevention in a primary care or community setting.

Table 1 Included Guidelines For CVD Prevention

Guideline Analysis

Sixteen guidelines provided no recommendations on smoking. They focused on other risk factors, such as excess weight, sedentary lifestyle and unhealthy diet.^{19,20,23,26,28,30,35,36,38,39,41,43,31,32,33,34}

Other guidelines recommended smoking cessation to prevent CVDs. It was recommended for all health care providers to advise smokers to quit, to provide information about methods to aid smoking cessation, including counselling services. Physicians were recommended to prescribe nicotine replacement therapy when indicated. No details, or very few, were found in the guidelines on how to advise smokers.^{21,25,27,29,37,42} Two guidelines focusing on smoking provided more detailed recommendations.^22,40 The following recommendations were identified:

• It was recommended to assess smoking status in all patients and their families and to use the Fagerstrom test (2 or 6 questions) to assess tobacco addiction (grade A).
• If smoking cessation is not attainable, smoking reduction should be the target, as an intermediate step towards quitting completely (grade B).
• Health care professionals were recommended to provide consultations dedicated to giving up smoking. The use of different techniques was recommended, such as counselling, psychological support, cognitive-behavioral therapy or motivational interviewing. Goals should be agreed with the patient and self-help documents should be provided. Past experiences, individual preferences and bio-psycho-social status should be taken into consideration (grade A).
• Intervention should be matched to the behavioral stage of change of the smoker concerned (according to the Prochaska and DiClemente model⁴⁴) and have an initial, intensive phase (weekly consultations) followed by a maintenance phase lasting 3 to 6 months (monthly sessions) (grade B).
• Health care professionals should appraise themselves of the level of knowledge found within the community regarding beliefs held about smoking and other CVD risk factors.
• Phone and/or internet follow-up and self-help, using eHealth technologies, could be alternatives even though they may be slightly less effective than face-to-face interventions (grade B).
• It was recommended to involve media and public educational campaigns (grade A) and schools (grade B) as part of multicomponent strategies to promote healthy lifestyles. School personnel and health care providers were recommended to set the example and to avoid smoking at work (class 1, grade A).

ESC guidelines provided public health recommendations which dealt with law making (eg, ban smoking in public places and set higher taxes on tobacco) and were consequently excluded because they were considered infeasible on a primary care level.⁵ NICE PH49 provided general recommendations on behavioral approaches (such as social support) and on health care system organization but no detailed interventions for stopping smoking.²⁴

References Analysis

One thousand eighty-one articles were identified of all 26 included guidelines for CVD prevention. After inclusion and exclusion criteria had been applied, 310 articles were included. Sixty-three included targeting smoking in their intervention. Seventeen studies were excluded from this analysis either because they presented no smoking-related outcomes (N=5) or they showed no significant differences for such outcomes (N=12). They were first included because the intervention was effective for other outcomes. Finally, 46 studies providing an effective smoking cessation intervention were included for analysis.

Interventions Description

Interventions and implementation strategies were described in detail and then summarized.

Fifteen studies provided a multicomponent intervention (targeting multiple risk factors) including an effective smoking cessation component and 31 studies targeted smoking only.

All interventions targeted a modification of the risk behavior related to smoking. They can be classified as behavioral change communication strategies.

Although interventions were heterogenous among included articles, some types of strategies were identified: self-help strategies alone, brief advice (BA), multiple sessions behavioral change strategies and community-based interventions.

Self-help strategies are presented in Table 2. These articles studied self-help strategies alone without face-to-face counseling.

Table 2 Included Articles (Alphabetical Order) With A Self-Help Intervention

Self-help documentation focused on the need to make a personal decision, the steps to successful quitting, the frequent need for repeated efforts, and the importance of setting a specific quit date and using substitutes for smoking. Self-help approaches were less effective compared with face-to-face counseling.⁴⁷

BA was defined as one or two short individual consultations to provide information on the potential harm caused by smoking, advice on how to quit and on where to search for help. BA strategies are presented in Table 3.

Table 3 Included Articles (Alphabetical Order) With A Brief Advice Intervention

BA was more effective than usual care (including very brief advice): results for RCT with intention to treat analysis (ITT) showed 4% to 11% more continuous abstinence at 12 months of follow-up. The measure of exhaled carbon monoxide showed it to be slightly, but non-significantly, more effective than BA alone, while BA plus lung age estimation with a portable spirometer was more effective than BA alone.^49,54

Multiple session behavioral change strategies (Table 4) were implemented as a series of meetings with participants to raise awareness of behavioral risk, to help them change and prevent relapse.

Table 4 Included Articles (Alphabetical Order) With A Multiple Session Intervention

The study from Hollis et al was included in both the self-help and the multiple session strategies as it implemented both interventions effectively.⁴⁵

Based on RCTs with ITT results, multiple session behavioral change strategies were more effective than usual care (+3 to 9% on 12 months continuous abstinence) but also than BA (+4 to 5%).

Eight studies focused on workplace (Table 5) for the implementation of multiple session strategies. The workplace was specified in order to link individual strategies with interventions, targeting all workers, such as workplace information campaigns (posters), newsletters, smoking restricted areas, local media campaign and social support among employees.^{78,79,80,81,82} Trained employees were successfully involved in the organization and intervention.^80,81

Table 5 Included Articles (Alphabetical Order) With A Multiple Session Intervention At The Workplace

Three studies were linked in financial incentives but they did not seem very effective when compared with behavioral strategies.^78,82,83 Three studies provided a BA intervention at the workplace, but with no or less involvement of the workplace structure and management.^50,51,60

Community-Based Interventions

These were public interventions such as public campaigns carried out in the community and/or in workplaces (Table 6). They targeted the global CVD risk and were multi-component (diet, sedentary lifestyle, smoking).^{85,86,87,88,89}

Table 6 Included Articles (Alphabetical Order) With Community-Based Interventions

The efficacy of such interventions was unclear: none of the included studies was an RCT and none showed any statistical difference on continuous abstinence but only on point prevalence abstinence (last 7 days reported consumption).

Discussion

The objective of this work was to identify effective non-pharmacological interventions for smoking cessation to be used in a primary care or a community setting. A systematic review of guidelines, following the ADAPTE Process, included the best quality CVD prevention guidelines adapted to such settings. Health care professionals were recommended to screen for smoking, to give advice and support and to prescribe NRT when indicated. Guidelines globally failed to provide clear and detailed strategies to implement such recommended advice and support. Plus, they were centered on health care professionals. Recommendations for changes to law and to taxation were deemed infeasible in a primary care or community context and were excluded. Two guidelines only provided more detailed recommendations, such as the use of behavioral change strategies with an intensive and a maintenance phase, as well as the use of techniques such as motivational interviewing or the Prochaska stage of change theory.^22,40

References analysis made it possible to specify implementation strategies for these recommendations. Although the included articles were seen to be heterogeneous in terms of the interventions, the settings and the outcomes, four main behavioral change strategies were identified.

Self-help materials showed a small positive effect in this review and in the literature, but they were less effective than face-to-face counseling.^{47,63,76,78,90} BA and multiple session behavioral change strategies were found to achieve better results than usual care or very BA (ie, 30 s to a couple of minutes).^45,52,73 Relative risk (RR) for physician advice has been estimated to be 1.66 for smoking cessation.⁹¹ Such interventions seemed even more effective when a defined behavioral change technique was used (motivational interviewing, “Five A’s”, Prochaska model or cognitive-behavioral therapy).^52,68,73 This was consistent with evidence in the literature summarized in two recent Cochrane reviews.^92,93 More intensive multiple session strategies were more successful than BA.⁶⁶ Similar results were found by other systematic reviews either for group or individual sessions: RR for abstinence was 1.22 to 1.57 for more intensive interventions compared with simple advice.^91,94,95,96 It should also be taken into consideration that the more intensive the intervention it is, the more it impacts on time and resources.

Several effective workplace interventions were found in this review. Despite this, the role of the “workplace” itself in this effect was still unclear. Best quality studies, implementing effective workplace interventions in our review included multiple session behavioral change strategies. Implemented interventions might be as effective whether offered in the workplace or elsewhere.⁹⁷ Spirometric lung age estimation was effective when added to BA in one study, but lung age did not differ among those who quit and those who did not. Smokers with poorer lung age were just as likely to quit as those with normal lung age. Other studies failed to show spirometry or lung age effective in promoting smoking cessation.⁹⁸ Lung age biofeedback effect on smoking is still uncertain and concerns may be raised on the cost-effectiveness ratio.

Progressive and abrupt cessation showed similar long-term results.⁹⁹ Evidence in the literature was insufficient to advocate aversive smoking techniques.¹⁰⁰ Accordingly, only one article included in this review studied overt and symbolic aversion, but due to a small sample and a per protocol analysis, its results need to be interpreted with precaution.⁶¹ Money incentives were not very effective and their effect was time-limited.¹⁰¹ They did not seem a pertinent solution.

Diverse health care professionals were effective in leading such behavioral change interventions. This review showed evidence to indicate that trained community workers can successfully lead these interventions.^52,66,78,82 This is important for settings where access to health care professionals is difficult.

Community-based public interventions, such as community events, mass media involvement and information campaigns, were identified in this review. They showed positive but small effects in non-controlled or non-randomized controlled trials which limited their strength. A systematic review found that community interventions had no effect, or a very limited effect, on quitting rates among adults. Among the 37 included studies, only 4 used random assignment of communities and largest and best conducted studies failed to detect a significant effect.¹⁰² Similarly, exclusive mass media interventions had an unclear effect mainly due to the lack of good quality studies.¹⁰³ As smoking is also determined by the social context, the community approach still should be part of a smoking cessation project.

This review did not identify any intervention involving social media. The use of social media has been shown to be feasible and acceptable, although data are lacking regarding its effectiveness.¹⁰⁴

Recent efforts to reduce smoking focused on anti-tobacco national policies, mass media communication and law-making. Less was done on the individual level. Cessation services were the least implemented component of MPOWER and physician interventions were not routinely provided, even in high-income countries.^10,105,106 This review highlighted the efficacy of different behavioral change strategies that can be implemented in primary care and/or in the community. These approaches are highly cost-effective. BA seems more suitable for general practitioners or practice nurses because of the short duration of their consultation (5 to 20 mins in most countries).¹⁰⁷ Pharmacists and dentist could play a major role as well.

Multiple session strategies, using a structured behavior change model (Prochaska Transtheoretical model, or motivational interviewing techniques) were seen to be more effective. Since they demand more time, they do not seem suitable for general practitioners worldwide.

This review showed that diverse health care professionals can successfully lead such interventions including lay health community workers.^52,66,78,82 Training lay people to provide such behavioral change interventions could be a solution for public health authorities that deal with settings where access to care is difficult, such as low-income countries and regions in developed countries where physician availability is limited. Furthermore, involving lay people could strengthen the community approach of community-based smoking cessation projects.

Strengths And Limitations

International guidelines were reviewed following a validated and systematic protocol including their quality evaluation. Guidelines were selected, where pertinent for primary care or community context, by researchers who are primary care health care professionals (general practitioners and nurses).

This review might carry a selection bias. Studies were included if they showed a significant difference on primary outcomes. This could imply a bias concerning the effectiveness of these interventions. Since the references included arose from high-grade recommendations of the best quality international guidelines, these interventions were considered effective and this bias was therefore limited.

Multiple outcomes were found for smoking cessation interventions and the relevance of some outcomes is debatable and could involve a bias. Seven-day point prevalence abstinence was the most frequently recurring outcome in this review. Biochemical validation was frequent. An expert consensus stated that smoking cessation should be evaluated by 7-day point prevalence abstinence and prolonged abstinence at 6 and 12 months because shorter time intervals carry a very high probability of relapse.¹⁰⁸ In this review, this bias concerned community interventions.

A publication bias was possible because this review protocol did not include a grey literature search. Nevertheless, different systematic reviews found similar results for the interventions described in this review.^92,93,94

Conclusion

Multiple session behavioral change strategies including follow-up and self-help materials were the most effective interventions for smoking cessation in primary care or community context.

These were based on motivational interviewing, Prochaska and DiClemente stage of change theory or cognitive-behavioral therapies. BA was slightly less effective. Nonetheless, BA involves less time and fewer resources and it is more suitable for primary care. These interventions could be provided by physicians, nurses and other health care professionals. Adequately trained lay health workers were also shown to be successful in providing these interventions. Community-based and workplace public interventions without behavioral change strategies were not effective. Evidence for community-based interventions effectiveness was weak due to the lack of robust studies.