Transforming Tobacco Cessation Counselling in Primary Care: A Mixed-Methods Evaluation of the Shared Decision-Making Model

Abstract

Background

Smoking cessation aids are under-prescribed in primary care. General practitioners (GPs) might prescribe more smoking cessation aids if they focused on the choice between treatments as the default, rather than using motivational interviewing that focuses on whether to make a quit attempt.

Objectives

Explore the self-reported adoption and impressions of shared-decision making (SDM) for smoking cessation treatments using a decision aid amongst GPs who received intervention or control trainings.

Design

Mixed quantitative-qualitative convergent analysis of self-reported behaviours from GPs in a pragmatic, cluster-randomized trial conducted between 2021-2023 in Switzerland and France.

Methods

The intervention training encouraged GPs to offer a choice of smoking cessation aids as the default choice using a decision aid (sustained intervention). The control training presented cessation aids without encouraging GPs to change how they discuss smoking cessation. Implementation was measured using questionnaires before and after the trainings, then at 6- and 12-month follow-up. Self-reported behaviours were quantitatively analysed using logistic regression adjusting for baseline responses. Experiences using SDM and default choice were explored through semi-structured interviews.

Results

A total of 43 GPs completed questionnaires (26 women (60%), age range 30-59 years) and 12 an interview. The intervention group reported more sustained interventions (OR 8.2 (CI 1.9-50)), use of the decision aid (OR 19.7 (CI 4.1-135)), and offer of treatments as the default choice than the control group (OR 6.4 (CI 1.4-40)) at 12-month follow-up. Small sample size bias was addressed by using Firth’s methodology. In interviews, GPs in the intervention group highlighted the utility of the decision aid and the importance of adapting the default approach to the context of individual patients.

Conclusion

GPs in the intervention group reported more often initiating treatment discussions with patients who smoke using a decision aid, but not more prescriptions. The default approach was not appropriate for all patients.

Introduction

Tobacco and its associated diseases are a leading cause of preventable mortality.^1,2 Interventions delivered by general practitioners (GPs) have proven effective in promoting tobacco cessation, increasing life expectancy ³ and decreasing costs. ⁴ In Switzerland, the health program Vivre sans Tabac (Living without tobacco), part of the national health prevention program, Pepra, from the Swiss Medical Association (FMH), has been teaching GPs how to assist their patients in quitting tobacco for 20 years. The counselling aspect of the training program is based on motivational interviewing (MI). ⁵ This initiative aligns with the World Health Organization Framework Convention on Tobacco Control (FCTC), which in Article 14 mandates that member states design and implement effective programs for promoting tobacco cessation. ⁶ However, real-world quit rates remain inconsistent, highlighting the need for novel approaches to tobacco cessation. ⁷

Different approaches to smoking cessation have been developed and evaluated in primary care, ⁸ the shortest being brief interventions (BI), a 1-to-5-minute reminder that tobacco is a health risk and help is available for quitting. BI results in small improvements in cessation rates, suggesting the need for more intensive approaches. MI is a complex intervention that helps patients explore and resolve their ambivalence regarding the need to change a certain behaviour. ⁹ Through a series of discussions, providers bring the patient forward in the stages of change. Ideally, MI leads to the next approach, the sustained intervention, whereby GPs offer more intensive behavioural and pharmacological help to quit tobacco with close follow-up. ¹⁰

However, the focus of MI on ambivalence may contribute to an underuse of pharmacological and behavioural interventions for smoking cessation, because these approaches are not proposed before the patient indicates clearly their intent to stop smoking. A 2008 study concluded that although most junior physicians in Switzerland asked about smoking, they often failed to offer cessation advice, ¹¹ a problem that persisted in a study from 2025. ¹² Another 2010 report ¹³ showed that specific help for smoking cessation was seldom provided during medical visits. MI is also underused, as documented in other countries, with lack of time being a primary factor.^14,15 A recent systematic review found the evidence to support MI in aiding tobacco cessation is actually quite weak. ¹⁶

An alternative to MI is shared decision-making (SDM) which is focused on comparing available treatments for quitting. SDM is a 3-step process where healthcare providers present more than one reasonable alternative, describe the options, and support patients in making a choice while exploring their values and preferences ¹⁷ . The aim of SDM is to increase patients’ autonomy by shifting the doctor-patient relationship from paternalistic to one of equals. ¹⁸ This shift establishes a collaborative partnership where both parties actively explore tailored options. ⁷ MI and SDM have similarities and are not mutually exclusive. The main difference being that with SDM the practitioner may directly address the question of treatment, while with MI treatments are typically not proposed before patients say they are ready. SDM could provide a more direct use of the sustained intervention. ¹⁹

A recent observational study found that a structured SDM model was correlated with a substantial increase in treatment acceptance, prescription rates, and patient adherence, even among smokers who were initially ambivalent or unmotivated. ²⁰ Navigating the expanding diversity of modern cessation methods is highly facilitated by structured decision aids, which clarify the potential risks, benefits, and personal suitability of available treatments. ⁷

SDM focused on comparing treatments, supported by an encounter decision aid (DA), was tested in the FIRST study, a pragmatic cluster randomized trial conducted in Switzerland and France. ²¹ The trial intervention, a three-hour training, encouraged GPs to compare treatment options for smoking cessation with a SDM approach with all their patients who smoke. The hypothesis was that the GPs would more often discuss treatment options, encourage quit attempts, and prescribe quit aids. Patient-reported outcomes from the trial have been published previously. While the intervention did not increase smoking abstinence at 6 months follow-up, it increased the number of quit attempts, use of nicotine replacement therapies, and patient involvement in consultations. ²²

The current study sought to understand to what extent GPs in the intervention group reported adopting the study intervention, as compared to GPs in the control group who presumably maintained their usual approach to smoking cessation MI. In parallel, we conducted semi-structured interviews to explore their impressions of the new approach, the accompanying DA and the subsequent barriers to this method.

Method

Design and Setting

We performed a mixed quantitative-qualitative convergent analysis of GP adoption and impressions of SDM for smoking cessation in the FIRST trial. The FIRST trial was conducted in the French-speaking cantons of Switzerland and Lyon region of France between 2021 and 2023. ²² The primary outcome was patient self-reported, 7-day point prevalence smoking abstinence at 6 months. Pre-specified secondary outcomes included the self-reported acceptance of the default choice (DC), use of the DA and prescription of medications by GPs. Ethics approval was obtained from the Canton of Vaud Ethical Review Board (2020– 02898) for the study protocol version 1.4 on the 15.02.2022. All GPs provided written informed consent.

GPs were blindly randomized at enrolment to either the control or intervention group trainings. ²² The control group received a 45-minute update on tobacco cessation treatment options without trying to modify their usual counselling routine. Usual counselling for the GPs from the study region was MI, based on what they learned during undergraduate or postgraduate training. The intervention group received the same update on tobacco cessation treatment options plus an additional training on the concept and use of SDM with a paper or electronic encounter DA, https://www.howtoquit.ch. The DA is intended for use during consultations and helps GPs compare smoking cessation aids that can be prescribed or currently recommended in Switzerland and France. ²³ The DA was developed by Selby K et al and was not based on a preexisting validated model or Option Grid. They were encouraged to offer a choice of smoking cessation aids as the default, ²⁴ that is to begin discussing treatment as the automatic option unless the patient actively opposes.

Participants

Inclusion criteria for the trial were GPs practicing in private practice in the Cantons of Vaud, Geneva, Jura, Neuchâtel, Fribourg or Valais, or metropolitan France who had a consultation frequency of at least 80 patients per month. The exclusion criteria were GPs who had completed an intensive smoking cessation course (i.e. at least half a day of training) within the last 2 years and GPs who had plans to retire or relocate outside of Switzerland or France in the next 12 months. While GPs in France were included for the quantitative data collection, the qualitative interviews were conducted only with GPs in Switzerland. One GP was present for the control training and completed the pre and post questionnaires but then proceeded to not sign the trial contract nor include patients in the trial; the responses from this GP have been included in this analysis though they were excluded from the main trial results. ²²

Data Collection

Both control and intervention groups completed paper questionnaires before and after the training, and electronic questionnaires at 6 and 12-month follow-up (supplemental materials). The questionnaires were newly developed for this study and pilot-tested, along with the intervention training, on two groups of 12 GPs in training. The pre-training questionnaire focused on the type of counselling they used most often prior to the training, described as BI, MI or SI. The post-training questionnaire focused on which type of counselling they intended to use in the future. The 6 and 12-month questionnaires focused on the counselling they had used since the last questionnaire. Also explored, in both the before and after training questionnaires, was the prescription of tobacco cessation aids. We used REDCap software, a secure web application hosted on Unisanté (University Center for Primary Care and Public Health). Source data was from paper questionnaires (exclusively pre and post training surveys), electronic questionnaires, and audio recordings of qualitative interviews, depending on the data collection point. The surveys at 6 and 12-months post training were directly completed online by the GPs.

Qualitative interviews were conducted a minimum of 6 months after the training program so that GPs had time to use the SDM approach and DA in their practice. In total, 12 semi-structured interviews were conducted with a purposive sample of the GPs in both the intervention and control groups, with an emphasis on having a balance between genders and ages. As of 12 interviews theme saturation was reached with a repetition of the themes previously identified. The interviews were mostly conducted over the telephone (2 conducted in person). The interview guides (see supplemental materials) were developed to explore the questions of practicality, appropriateness, and feasibility of SDM and the DA. Questions were adapted after two initial interviews. No previous model or assessment grid was used to develop the questions.

Outcomes

Quantitative outcomes focused on the self-reported uptake of key behaviours by GPs, notably: 1) their main approach to giving quitting advice, 2) the frequency of offering prescription quit aids, 3) the appropriateness of the default choice approach within the primary care context, and 4) use of the DA during consultations with tobacco smoking patients.

To increase statistical power, all quantitative outcomes were dichotomized into binary variables, with neutral responses considered as non-adoption:

Primary Cessation Approach

GPs selected their main approach from three options: brief interventions, motivational interviewing (MI), or sustained interventions (SI). Because the study training specifically promoted SI, defined as direct and frequent treatment discussions with all current smokers, this outcome was dichotomized into sustained intervention versus other.

Prescription Frequency

GPs rated how often they suggested smoking cessation aids (e.g. nicotine replacement therapy, varenicline, bupropion) using a Likert scale. Responses were dichotomized into <10% of cases (never/rare) versus ≥10% of cases.

Appropriateness of the Default Choice

GPs evaluated the appropriateness of a “default choice” approach, defined as directly proposing cessation aids without first exploring the patient’s motivation to quit. Likert scale responses were dichotomized into Agree/Totally agree versus Disagree/Strongly disagree/I don’t know.

Decision Aid (DA) Utilization

GPs reported whether they used the DA during consultations. Responses (yes, no, or I don’t know) were dichotomized into Yes versus No/I don’t know, as

the neutral “I don’t know” option was interpreted as an indicator of an inactive stance regarding DA implementation.

Qualitative outcomes focused on the experiences of the GPs when implementing the intervention during 6 months of routine care. We explored facilitators and barriers to offering smoking cessation aids to patients in primary care, and how default choice or a DA influenced their practice.

Data Analysis

Quantitative data were extracted from REDCap and analysed using R version 4.4.2, R Core Team (2024). Descriptive statistics were used to present the characteristics of GPs in the intervention and control groups. We used logistic regression, controlling for baseline responses, to compare the trial arms in terms of the proportion of GPs who self-reported having adopted key behaviours at each time point. Each time point was analysed in a separate model. A methodology by Firth ²⁵ was used to address separation problems due to small sample sizes. To compensate for the limited sample size and to ensure categorical clarity, scale responses were dichotomized into binary formats as described in the previous paragraph. P-values of less than 0.05 were considered significant. Missing responses were excluded from single time points. In our sample, attrition peaked at the 6-month and 12-month follow-ups, with up to 19% of control GPs (4/21) and 4.5% of intervention GPs (1/22) missing from specific analyses (Supplemental table 1). Complete-case analysis in logistic regression has been shown to provide unbiased estimates in a variety of situations beyond Missing Completely At Random, including when the chance of being a complete case depends only on the outcome or only on the exposure. While we acknowledge that we cannot definitively verify that our situation falls into one of these categories, imputation is unlikely to provide any additional information.

The interviews were transcribed and analysed using a mix of inductive and deductive thematic coding using the MAXQDA 2020 software as a platform. Prespecified deductive themes were feasibility, practicality, appropriateness and barriers to implementing study intervention in routine practice. To ensure data reliability, 20% of the sample was subjected to dual independent coding, where two researchers analysed the data separately to verify consistency and minimize observer bias. Data and investigator triangulation were used to verify findings and reduce bias. During the interpretation phase, the quantitative and qualitative datasets underwent a side-by-side cross-analysis, allowing for the convergence of results from different sources to validate findings. This integrative approach enabled the researchers to identify complementary patterns and resolve any discrepancies between the two research strands.

Results

Characteristics of Participants

A total of 51 GPs were randomized and assigned a training date. Eight withdrew prior to consent, 4 in each group (Supplemental Figure). Consequently, 43 GPs (84%) completed the training (Table 1). There were more women GPs (60%) with an average age of 40-49 years (47%). The main country of practice was Switzerland, mostly in urban areas and within group practices of 4 or more physicians. Most GPs estimated that between 10-30% of their patients smoke daily, and the majority (70%) had never smoked or vaped themselves.

Table 1.

Characteristics of Participating General Practitioners

GP characteristics	Control (n=21)	Intervention (n=22)	Total (n=43)
Age
30-39 years	7 (33%)	7 (32%)	14 (33%)
40-49 years	11 (52%)	9 (41%)	20 (47%)
50-59 years	2 (10%)	6 (27%)	8 (19%)
Gender
Female	13 (62%)	13 (59%)	26 (60%)
Male	8 (38%)	9 (41%)	16 (37%)
Country of practice
Switzerland	16 (76%)	17 (77%)	33 (77%)
France	5 (24%)	5 (23%)	10 (23%)
Location of practice
Urban	13 (62%)	17 (77%)	30 (70%)
Rural	7 (33%)	5 (23%)	12 (28%)
Country medical diploma obtained
Switzerland	10 (48%)	10 (45%)	20 (47%)
France	5 (24%)	11 (50%)	16 (37%)
Country within European Union	4 (19%)	1 (5%)	5 (12%)
Country outside European Union	1 (5%)	0	1 (2%)
Number of GPs in same practice
1	1 (5%)	4 (18%)	5 (12%)
2	1 (5%)	2 (9%)	3 (7%)
3	6 (29%)	2 (9%)	8 (19%)
≥ 4	12 (57%)	14 (64%)	26 (60%)
Do you smoke or vape?
Yes, cigarettes	0	1 (5%)	1 (2%)
Electronic cigarettes (vape)	1 (5%)	1 (5%)	2 (5%)
Heated tobacco	0	0	0
Another tobacco product (Snus)	0	1 (5%)	1 (2%)
I used to smoke/vape	4 (19%)	4 (18%)	8 (19%)
I have never smoked or vaped	15 (71%)	15 (68%)	30 (70%)
How many half days do you work per week?
≤ 5 half days	7 (33%)	1 (5%)	8 (19%)
Between 6 and 9 half days	12 (57%)	14 (64%)	26 (60%)
≥ 10 half days	1 (5%)	7 (32%)	8 (19%)
How many hours/days of training on tobacco cessation have you received in the last 5 years?
0 to 2 hours	15 (71%)	15 (68%)	30 (70%)
2hrs to 1 day	5 (24%)	7 (32%)	12 (28%)
What percentage of your patients smoke daily?
10 - 20%	7 (33%)	7 (32%)	14 (35%)
20 - 30%	10 (48%)	7 (32%)	17 (42%)
>30%	2 (10%)	4 (18%)	6 (15%)
I don’t know	1 (5%)	4 (18%)	5 (12%)

Quantitative outcomes focused on the uptake of key behaviours targeted by our intervention are shown in the figures with the number of respondents in each category presented in Supplemental Table 1. In Figure 1, GPs in the intervention group were more likely than GPs in the control group to intend to use sustained intervention in their future practice post training (OR 17.0 (CI 4.0-94.0), P<0.001), and this difference was maintained over time (12 months: OR 8.2 (CI 1.9-50.3), P=0.004).

Figure 1.

The proportion of GPs mainly using sustained interventions when counselling current smokers, as opposed to motivational interviewing or brief advice (n=43). Brackets represent 95% confidence intervals

In Figure 2, the frequency of offering quit aids did not significantly differ between arms post training (P=0.067) nor over time (6 months P=0.35, 12 months P=0.59). However, both groups demonstrated important increases in offering quit aids after the training, which is probably a reflection of the control group participating in the study, receiving training about the prescription possibilities and thus being motivated to help their patients change their habits.

Figure 2.

Proportion of GPs reporting they prescribe quit aids to patients who smoke (n=43). Brackets represent 95% confidence intervals

In Figure 3, the proportion of GPs agreeing with the appropriateness of the default choice was significantly larger in the intervention group post training (OR 59.4 (CI 5.7-8222.1), P<0.001), and that difference persisted over time (6 months: OR 6.6 (CI 1.5-39.8), P=0.012, 12 months: OR 6.4 (CI 1.4-40.2), P=0.014) thus showing that this practice was indeed incorporated into the common practice of the GPs in the intervention group compared to the control group. However, the confidence interval was wide (CI 5.7–8222.1), indicating uncertainty regarding the size of the effect.

Figure 3.

Proportion of GPs reporting it is appropriate to offer patients quit advice regardless of their level of motivation to quit smoking (n=43). Brackets represent 95% confidence intervals

In Figure 4, the use of a DA was more frequent in the intervention group post training (OR 12.1 (CI 2.2-126.6), P=0.003) and throughout the follow-up time points, 6 months (OR 11.3 (CI 2.5-71.5), P=0.001) and 12 months (OR 19.7 (CI 4.1-134.8), P<0.001). The trend seen in the intervention group is also present for the control group although at a lower level (Figure 4).

Figure 4.

Proportion of GPs reporting the use of the decision aid during consultations with tobacco smoking patients (n=43). Brackets represent 95% confidence intervals

Qualitative outcomes focused on the appropriateness, practicality and barriers of the intervention. The themes and sub-themes identified during the semi-structured interviews are described in Table 2.

Table 2.

Thematic Analysis of Interviews With General Practitioners (n=12 Interviews)

Themes	Sub-themes supporting	Sub-themes opposing
Using pharmacological aids to quit smoking:	- Patient past experiences determine the type of prescription given - Motivated patient - It’s the patient who makes the decision to take a treatment and which type	- Patients’ lack of motivation Nicotine replacement therapy: - Cost to patients (not reimbursed by insurance in Switzerland) - High price hinders use by patients Medications (varenicline/bupropion): - Link between medicine and illness - 12 hours between taking the medicine, fear of cravings - Unavailability of medications
Implementing study intervention into GP practice	- Practical - Efficient (time) - Proactive - Generates patient interest - Establishes a relationship with the patient - More prescriptions - Talking more about stopping, even at less usual times	- Fear of offending the patient if approach the subject too directly - Don’t rush the patient - Depends on the situation
Default choice	- Suitable for primary care - Efficient (time) - Generates discussions - Pragmatic - GPs can be more direct because they already have a strong link with their patients.	- Could create conflict if the patient is not ready to quit - Inappropriate in certain situations
Encounter decision aid	- A good comprehensive summary - Suitable for primary care - Efficient (time) - Practical - Tool that opens and generates discussions with patients - Tool actively shared with colleague - Topic of tobacco more actively addressed - Happy patients	​

Theme: Implementing Study Intervention Into GP Practice

Implementing the study intervention into general practice was overall seen as positive with feedback such as,

It allows you to get straight to the point. (MT03, Pos. 23)

It's very rich because it's quite effective in that sense [talking directly about tobacco cessation]. (MT05, Pos. 115)

…considering the patient as a partner. (MT17, Pos. 39)

I found that the new method was a little less worried about the fact that it didn't hesitate to start, even with drug treatments. (MT04, Pos. 19)

However, some GPs were wary of the intervention, raising such concerns as,

I feel like I'll be offending [patients] if I'm too direct. (MT10, Pos. 27)

There are certain situations where it's not appropriate to talk about [tobacco cessation]. (MT17, Pos. 80)

Theme: Using Pharmacological Aids to Quit Smoking

When it comes to pharmacological prescription one of the main themes was,

It really depends on the patient. (MT11, pos. 19)

This is in reference to the patients’ past experiences with tobacco cessation, whether they were open or not to the discussion, if it was a new patient, or if follow-up was in place for several years.

The financial strain on patients needing to pay for pharmacological aids was not lost on GPs, who commented,

The problem with [NRT], of course was reimbursement, and that really posed a problem. (MT11, Pos. 43)

This being in reference to the fact that NRT is not reimbursed by Swiss basic insurance.

Theme: Default Choice

GPs gave a mean rating of 8 (out of 10) regarding the appropriateness of using the default choice approach when counselling smoking cessation in general practice. However, their comments were mixed. These comments were given in favour of default choice,

And frankly, I like it in the end. In fact, we put aside all the guilt that may be involved when we explore patients' smoking habits in detail. (MT17, Pos. 174)

Default choice makes it easier to talk about quitting smoking (MT41, Pos. 117)

As a result, the subject isn't really addressed [during motivational interviewing], whereas DC, I think it address the subject and the patients…are happy. (MT04, pos. 107)

The GPs’ main opposition to default choice was the concept of using this technique in all circumstances when meeting a tobacco smoking patient.

I mean …, it has to be an appropriate time. (MT17, pos. 146)

Interviewer: So one obstacle to using this kind of method in general practice would be an emergency…where we know less… MT05: Yes, where we don't know the patient (MT05, Pos. 110-111).

In reality it's perhaps a bit more complicated, because I think we still have to take into account where the patient is at the moment he's in front of us, because otherwise I have the impression that it's going to be a total failure, or else be counter-productive. (MT10, pos. 131)

Theme: Encounter Decision Aid

The encounter DA was very popular with the GPs, who gave a mean score of 9.5 (out of 10) regarding its suitability to general medicine. One GP had already actively shared the tool with another colleague.

So I'd really say having a tool (DA). Well, I think it's helpful to have a tool to bring along, it brings in a ‘third party’, it allows the two of us to concentrate on the same thing and not be in a simple discussion which is sometimes, difficult for the patient to materialize. (MT03, Pos. 89)

It [DA] allows you to interact with the patient. (MT04, Pos. 95)

I think it's practical, it's an easy approach to use in general practice. (MT41, pos. 133)

I was looking for, trying to find like a document, something on the internet, but in the end, each time, it was scattered information, here we have something clear on an index card, everything is well summarised… and it opened the discussion with some people. (MT11, Pos. 117)

it's very clear, summarized. The fact that it's concise is very useful. (MT05, pos. 127).

So I strongly recommended it. I gave it to… a colleague with whom we were talking about it (tobacco), and then I said I've got this. So he was very happy. (MT11, pos. 109)

We're no longer talking about the problem, but about how to get out of it, and that's all very appropriate. (MT17, pos. 194)

You focus the attention in a common way. I think it helps to tackle the subject (MT03, pos. 91)

A mixed-methods triangulation, integrating quantitative longitudinal data with qualitative insights from semi-structured interviews, was employed to validate the intervention’s impact and ensure robustness. The triangulation of quantitative and qualitative data corroborated the trends identified across both sources, suggesting that the standpoints captured during interviews were representative of the broader GP sample. Quantitative results revealed a distinct shift toward sustained intervention in the intervention group compared to traditional motivational interviewing (MI) (Figure 1), a finding closely aligned with the positive sub-themes identified in the qualitative analysis (Table 2). While both groups showed a comparable increase in prescribing quit aids (Figure 2), the interviews clarified this trend, attributing the control group’s progress to a ‘study effect'—where participation and the refresher training energized their clinical approach. Furthermore, while the intervention group significantly increased advice-giving regardless of patient motivation (Figure 3), the qualitative data provided a critical nuance: although this direct approach was valued for its efficiency, it also revealed professional concerns regarding potential strain on the patient-provider relationship. Finally, the high uptake and acceptance of the Decision Aid (DA) (Figure 4) were directly validated by the enthusiastic feedback from the GP interviews.

Discussion

This mixed-methods study was embedded in a pragmatic, cluster-randomized trial that trained intervention GPs to perform a sustained intervention where they offered smoking cessation treatment as the default choice using a DA. The quantitative results showed a significant increase in the use of sustained interventions and the DA in the intervention group, suggesting that GPs adopted the SDM approach focused on treatment and maintained it during 12 months of follow-up. This was further confirmed through the qualitative data with the DA especially being highly noted by the GPs. The default choice approach, although controversial in the interviews, did demonstrate a significant change for the intervention group. GPs described how they adapted what they were taught to fit the individual context of their patients. Following this, it is interesting to note that the default choice approach which is essentially an opt-out method of care is put forward by Richter et al as being a more appropriate and barrier lowering method of care. ²⁶

The use of quit aids increased similarly in both groups, suggesting that the control training also had a strong effect. Although this was not an aim of our study it does highlight previously demonstrated concepts that insufficiently trained GPs are less liking to give counselling advice. ²⁷ Although there was not a significant increase in the use of quit aids, it has been shown in previous studies that for tobacco cessation, like for other addictions, it takes multiple attempts to quit. ²⁸ The fact that GPs talk about tobacco use helps patients to question their consumption, seek help and move closer to quitting.^29-31

The WHO Framework Convention on Tobacco Control stipulates that countries must ensure their healthcare facilities, like primary care practices in Switzerland, provide effective treatment for tobacco dependence. ⁶ The qualitative results showed how sustained interventions with the treatment as the default choice were considered a more direct, and possibly more feasible, approach to tobacco cessation than MI. The literature has stated multiple times how difficult it is for GPs to discuss tobacco cessation for multiple reasons, one of the principal ones being a lack of time.^14,32,33 The improper implementation of MI can reinforce this difficulty, as the discussion stagnates on the patient’s ambivalence taking a significant amount of consultation time before any action is taken. ⁹ GPs must not only encourage their patients to change their behaviour but may have to sidestep the reason for the consultation to discuss a subject not brought up by their patient. Default choice has the benefits of being a simple and direct method of opening the discussion about tobacco cessation, therefore not taking excessive amounts of consultation time but also bypassing the question of ambivalence to focus on a proactive and encouraging discussions. The default choice was used in a range of different primary care practices, including French group practices where GPs were able to refer patients to public health nurses. As demonstrated by our study, GPs then have the capacity to adapt this technique to the patient and the situation in front of them. In addition, the study conducted by Boesch et al demonstrates how satisfied and positive patients were to the sustained intervention approach without causing any perceived strain on the therapeutic alliance, a concern that was brought up multiple times by the GPs in the interviews. ³⁴

Though the FIRST trial did not increase smoking cessation after 6 months follow-up, it did increase the number of quit attempts, use of smoking cessation aids and patient involvement in decision-making. ²² The current study adds to those findings by suggesting that GPs in the intervention group were able to adapt the concepts from the training to their daily practice and appreciated the DA in the clinical setting, therefore rendering it more realistic to primary care. The intervention group was overall very satisfied with sustained intervention, they were more relaxed, confident and readily available to discuss tobacco cessation with their patients.

Strengths and Limitations

In the strengths of our study, we highlight the longitudinal follow-up 12-months after the baseline training, allowing us to see how behaviours were maintained over time. Having an active control group was a strength to be able to compare two working groups. The combination of qualitative and quantitative data gave a more complete and multidimensional aspect to the results obtained. In the limitations, there was the small sample size of GPs, leading to large confidence intervals around our estimated effect size, and no external evaluation when it came to the interviews. We did have some missing data, with a small imbalance between intervention and control groups, especially at the 6-month follow-up. Our findings are based on GPs’ self-reported behaviours; given the significant desirability bias associated with smoking cessation, they may have overstated how often they actually implemented these actions. The desirability may have been accentuated by the fact that the same study team worked on the RCT FIRST as well as this study. There could have been a confusion of terms when GPs were answering questionnaires, as there were many concepts taught at the same time, SDM, default choice and sustained intervention, moreover the concepts are linked in practice. This is especially pertinent as the questionnaires were not previously validated. It should be acknowledged that the effect size of the intervention is known with very low precision due to the small sample size of GPs, as the cluster-randomized trial was powered to detect a change at the level of patients. Furthermore, while the trial protocol specified self-reported uptake of the intervention by GPs as a secondary endpoint, ²¹ the statistical analysis was chosen later. However, the systematic superiority of the intervention group and the large number of significant differences confer confidence in the overall ability of the intervention to change practice.

Conclusion

Participating GPs reported that SDM supported by an encounter DA was appropriate and practical in routine care and helped them deliver sustained interventions to more patients. They found that the tool made tobacco cessation discussions clearer, simpler, and more direct. Although the FIRST trial did not demonstrate an increase in patient-reported smoking cessation, the current article clearly shows there was a change in GP behaviours towards a more proactive offering of quit aids and a focus on treatment choices. Future research should focus on implementing shared decision making as a tool for smoking cessation in primary care.

ORCID iDs

Michael Amiguet https://orcid.org/0000-0001-8814-2043

Kevin Selby https://orcid.org/0000-0002-9096-0720

Data Availability Statement

The data underlying this study are not publicly available due to the inclusion of sensitive participant information. De-identified (coded) data from participants who provided explicit consent for data sharing may be made available upon reasonable request to the corresponding author, subject to approval by the study investigators and in accordance with applicable ethical and legal requirements.*