Background
Current guidelines recommend that general practitioners (GPs) routinely ask patients about smoking, advise them to quit, assess their motivation to quit, assist them with quitting, and arrange follow-up quit smoking support (the 5-A Model) [
1],[
2]. However, GPs report difficulties when translating these guidelines into practise [
3]-[
7], resulting in a substantial gap between evidence and practise. A study in Dutch general practise showed that 79% of all smokers and 40% of smokers who discussed smoking with their GP did not receive stop-smoking advice [
8]. The development of strategies that facilitate the implementation of guideline-recommended smoking cessation care may result in more patients being advised to quit and being provided with evidence-based quit-smoking support and, ultimately, giving up smoking [
9]-[
11].
Strategies to facilitate the implementation of evidence-based clinical guidelines often focus on influencing the behaviour of the health-care professionals [
12]-[
15]. Efforts to change the clinical behaviour of health-care professionals often involve didactic modes of delivery aimed at educating these professionals [
13]-[
15]. However, this approach implies a lack of knowledge and assumes that additional knowledge will change the behaviour of health-care providers, neither of which may necessarily be true. In fact, enhancing knowledge alone may not be the best, or even an adequate strategy, to influence the clinical behaviour of health-care professionals [
16]. Similarly, the motivation and/or the beliefs of GPs to routinely adopt evidence-based guidelines are not always a reliable predictor of the routine implementation of these guidelines [
17].
Psychological theories may provide a basis for identifying the predictors of GP behaviour and of behaviour change [
16]. Clinical practise is a form of human behaviour that is sensitive to theory-based strategies that have proven effective in patient samples [
18]-[
22]. However, a systematic review showed that only a minority of the 235 interventions that previously aimed to facilitate guideline implementation by health-care professionals actually used those theory-based strategies [
12].
One of the well-established theory-based strategies (albeit in other populations) is the self-formation of ‘conditional plans’, such as action plans and coping plans [
23],[
24]. Action plans link a situational cue to behaviour in order to promote behaviour change and habit formation, e.g. ‘
if X occurs (if the patient visits me because of a cough more than three times a year), then I will do Y (I will advise the patient to quit smoking)’. Coping plans anticipate potential barriers to behaviour change which impede action plans from working. Such plans aim to bridge the gap between the individual’s intention to perform the behaviour and the actual performance of that behaviour [
25],[
26].
The mechanisms that underlie the effectiveness of action and coping plans involve a heightened accuracy and speed of detecting the contextual cue for performing the intended behaviour [
27]-[
30]. Plans that are more specific are suggested to result in a greater improvement of the intended behaviour compared to incomplete or vague plans [
31],[
32]. In addition, studies have shown that individuals who act according to their formulated action plans (i.e. plan enactment) are more likely to benefit from their plans, e.g. enacting an action plan to remove all tobacco products results in a higher likelihood to actually quit smoking [
33],[
34]. The effects of plan specificity and enactment on behaviour are strongest among those individuals who are the most motivated to change the intended behaviour [
31]-[
33],[
35].
It has been shown that planning predicts the clinical behaviour of GPs in various clinical conditions [
36]-[
38]. Moreover, an intervention study showed that incorporating planning in postgraduate education increased the use of a practitioner-guided procedure among mental health professionals [
35]. However, to our knowledge, no studies have examined whether planning improves the provision of evidence-based smoking cessation care by GPs.
The present study incorporates action planning within a training session for GPs, aimed at increasing their provision of smoking cessation tasks as recommended in clinical guidelines, including asking patients about smoking, advising them to quit, and arranging follow-up quit smoking support for smokers who are motivated to quit. Because GPs often indicate that patients’ lack of motivation to quit may act as a barrier to the provision of guideline-recommended smoking cessation care [
5],[
39],[
40], GPs also formulated a coping plan to address this potential barrier.
Based on the reported positive effects of action planning in patient samples [
41]-[
43], we hypothesised that GP action planning would improve the performance of these smoking cessation tasks. Secondly, we hypothesised that formulating a coping plan for smokers who are not motivated to quit provided GPs with a solution for this type of barrier, thereby increasing the provision of smoking cessation care for this group [
38],[
44]-[
48]. Since the present GP training includes multiple behaviour change strategies, we also examined the nature of action planning including plan specificity and plan enactment. In line with the previous findings on plan specificity and self-reported plan enactment [
31]-[
35], we hypothesised that GPs who formulated a highly specific plan and reported a high level of plan enactment would be more likely to provide smoking cessation care post-training. Finally, we hypothesised that these effects would be most evident among GPs with high intention to routinely implement smoking cessation care prior to the training.
Methods
Design and intervention
The present paper reports the results of a two-group cluster randomised controlled trial in general practise. GPs were randomly assigned to either the intervention or control condition. The intervention entailed a 1-h individual training session for GPs in the delivery of smoking cessation care. The training was based on behaviour change techniques related to methods that underlie the current Dutch guidelines for treating tobacco addiction (the 5-A model [
2],[
49]): 1) GPs’ implementation barriers were identified, 2) GPs were provided with a state-of-the-art evidence about the effectiveness of smoking cessation care, 3) GPs’ motivation to routinely implement the guideline was identified and improved using motivational interviewing techniques, 4) GP instruction was provided and tailored to the identified implementation barriers, and 5) GPs were given the opportunity to receive additional feedback support. Previously, the effects of the multicomponent training on GPs’ provision of smoking cessation care were tested and reported elsewhere [
50]. Action planning was one of the components of the GP training, and our initial RCT did not provide insight into the effects of this single behaviour change technique. Therefore, the present study focuses on a further examination of the effects and nature of action planning among the trained GPs.
Participants
During the study period (January–August 2011), 25 GPs received a 1-h training programme that incorporated action planning. At baseline (pre-intervention), these 25 GPs saw 1,002 patients, of whom 195 (19.5%) were smokers. During post-intervention, the same GPs saw a different group of 630 patients, of whom 98 (15.6%) were smokers. In the control condition, 24 GPs and 1,769 patients (baseline: 1,066, post-intervention: 703) were included, of whom 384 (21.7%) were smoking patients (baseline: 238 (22.3%), post-intervention: 146 (20.8%)).
Measurements
GP intention
Six weeks prior to the training programme, GPs rated their intention to implement guideline-recommended smoking cessation care on a 4-point scale (‘
no intention to routinely implement smoking cessation treatment within 6 months’ (0), ‘
intention to routinely implement smoking cessation treatment within 6 months’ (1), ‘
intention to routinely implement smoking cessation care within 1 month’ (2), and ‘
already routinely implemented smoking cessation treatment’ (3). To facilitate testing of the hypotheses, we used a
post-hoc categorisation in line with the principles from the health action process approach [
51] to classify GPs into three groups depending on their response to the question about their intention: 1) ‘GP pre-intenders’ (answer category 0; 4 GPs, 393 patients), ‘GP intenders’ (answer category 1 and 2 combined; 14 GPs, 2,211 patients), and ‘GP actors’ (answer category 3; 7 GPs, 797 patients).
Patient-reported provision of smoking cessation care
During the 3 weeks prior to and after the GP training programme, all patients completed a questionnaire immediately after their GP consultation in which they rated their GP’s smoking cessation activities during that consultation. This questionnaire included the following items: ‘Did your GP ask you about smoking during the consultation?’, ‘Did your GP advise you to quit during the consultation?’, and ‘Did your GP refer you to any kind of follow-up quit smoking support during the consultation’? For each item, patients could answer ‘Yes’ (1) or ‘No’ (0).
Action planning
During the GP training programme, action planning was assessed based on the separate plans formulated by the GP for a) identifying smokers and b) advising smokers to quit. GPs wrote down
who was going to perform the activity,
when the activity was going to be performed, and
how the activity was going to be registered in the patient’s electronic health record. In addition, GPs formulated an action plan for c) arranging follow-up for smokers who are motivated to quit and a coping plan for d) arranging follow-up for smokers who are not motivated to quit. In these plans, GPs formulated the
what,
who, and
how of each plan. This method is comparable to that used in similar studies with patient samples [
31].
Specificity of GP plans
The degree of specificity of each of the components of the GPs’ plans (
who,
when,
what, and
how) was assessed using a rating method based on previous studies [
31],[
32],[
34]. The
who component of the plans was rated as
not completed (0) or
completed (1). The
when,
what, and
how components of the plans were rated on a 4-point scale; components were rated as
not completed (0) if GPs did not write down any plans, and components were rated as being
general (1) when GPs described them in rather general terms, e.g. ‘
I will ask my patients about their smoking during the consultation’. Components that were specified with moderate precision were rated as being
moderately specific (2), e.g. ‘
I will ask my patients about their smoking, routinely once a year’. A component was rated as being
highly specific (3) when GPs specified their future action with a sufficient amount of precision, e.g. ‘
I will ask my patients about their smoking when they present with smoking-related complaints during the consultation’.
Analyses of the when component showed that GPs specified either a particular moment (e.g. during the consultation), or a particular type of patient (e.g. patients with smoking-related complaints), or both; therefore, we decided to rate both these types of specifications. As a result, the total specificity score for the first two action plans (asking about smoking and advising to quit) ranged from 0 to 10, and for the third action plan (dealing with smokers who were motivated to quit) and the coping plan (dealing with smokers who were not motivated to quit), scores ranged from 0 to 7 (Appendix 1).
Two researchers independently rated the specificity of all components of the GPs’ plans. Kappa statistics were used to estimate the inter-rater agreement; this resulted in a high level of agreement between the two researchers for the total specificity scores of the GPs’ plans: i.e. for asking about smoking 0.998 (95% CI 0.995–0.999), for advising to quit 0.940 (95% CI 0.864–0.973), for arranging follow-up for smokers who are motivated to quit 0.945 (95% CI 0.850–0.978), and for arranging follow-up for smokers not motivated to quit 0.962 (95% CI 0.907–0.984). These high kappa coefficients are probably due to the type of rating method used. Disagreements were discussed until consensus was achieved. For analyses, the GPs’ total plan specificity scores were categorised into low (1) and high (2) scores, using the mean score as a cut-off.
Enactment of GP plans
After the GP training, we were interested in providing the GPs in the intervention group with their self-formulated if-then plans and ask them if they had the opportunity to enact them. Therefore, 6 weeks after the GP training programme, via a postal questionnaire, the GPs were asked to report the extent of plan enactment (response rate 76%; n = 19). In this questionnaire, each GP was provided with the four plans that they had previously formulated. GPs were asked to rate the extent to which they had enacted each plan using a 5-point scale: ‘plan not enacted, not intending to enact in the future’ (0), ‘plan not enacted, intending to enact within 1 month’(1), ‘plan not enacted, intending to enact within a week’ (2), ‘plan partly enacted’ (3),‘plan fully enacted’ (4). For missing data, a negative scenario was applied which assumed that GPs who did not complete the questionnaire did not enact their plans (score 0). For the analyses, scores for plan enactment were categorised into low (1) and high (2) scores using the mean score as a cut-off.
Statistical analysis
Descriptive statistics were used for the characteristics of the GPs and for scores on specificity of the GP plan and on plan enactment. To test our hypotheses, we linked GP the data with patient data and analysed these using two-level logistic regression analyses (generalised estimating equations), including the data at the GP and patient level.
In our model, data at the GP level included scores on plan specificity and plan enactment as independent variables. To examine the main effects of these variables on GPs’ provision of smoking cessation care (patient-reported), all patients were classified into three categories, i.e. patients who had a consultation with a GP who had formulated a highly specific plan/reported a high level of plan enactment (2), patients who had a consultation with a GP who had formulated a general plan/reported a low level of plan enactment (1), and patients who had a consultation with a GP within the control condition (0).
Data at the patient level included GPs’ provision of smoking cessation care, as reported by patients, as dependent variables, including being asked about smoking, being advised to quit, and being provided with quit smoking follow-up. Patient-reported smoking cessation care was included as a dichotomous variable (1 = yes, 0 = no). The model was adjusted for differences between characteristics of the patients who visited the GPs at baseline and those who consulted the GPs post-intervention (gender, cultural background, and smoking status).
Univariate analysis was used to examine the main effects of GP plan specificity and GP-reported plan enactment on their provision of smoking cessation care (as reported by patients). In addition, interaction analysis was used to examine whether or not the effects of GP plan specificity on the delivery of care depended on the extent of GP plan enactment. Finally, subgroup analyses were performed to examine whether the effects of GP plan specificity and plan enactment on delivered smoking cessation care, differed between GPs with different baseline intentions to routinely implement smoking cessation care. In all models, we included time (baseline (0)/post-intervention (1)) by group (control group (0)/low plan specificity or low plan enactment (1)/high plan specificity or high plan enactment (2)) interaction effects since we included different cohorts of patients at baseline and post-intervention.
This study was approved by the Medical Ethical Board of the Leiden University Medical Centre (P10.125).
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
MV and MC contributed to the conception and design of the study, acquisition of the data, statistical analyses and interpretation of the data, drafting of the manuscript, and gave final approval for the submission of the manuscript. JP contributed to the interpretation of the data, drafting of the manuscript, and gave final approval for the submission of the manuscript. NC, MS, AK, and WA contributed to the conception and design of the study, drafting of the manuscript, and gave final approval for submission of the manuscript. All authors read and approved the final manuscript.