Previous research explained the GPs' low readiness to implement BI for excessive drinkers by pointing out the large number of implementation barriers. By means of vignette study we tried to show that GPs' low readiness should be explained rather by inherent characteristics of BI and less by associated barriers such as insufficient financial reimbursement (11). In our vignette study we tested for the combined effect of the factors intervention type (BI vs. PI) and financial reimbursement (standard vs. above-standard) on the implementation readiness of GPs. We found that GPs' reported readiness to implement BI was lower than their readiness to implement a pharmacological intervention. However, only a tendency that did not reach statistical significance could be found for the second factor (financial reimbursement). GPs trended toward a slightly higher implementation readiness under the condition of above-standard financial reimbursement.
To reduce alcohol consumption several approaches have been created (e.g. Motivational Interviewing (19)). These approaches have been evaluated positively by efficacy studies and are thus recommended for use (1). It is probably also true that GPs are generally ready to motivate excessive drinkers to reduce their alcohol consumption and they probably know very well about the health costs of excessive drinking (25). But they are nevertheless not ready to treat these patients with SBI. Although our results were only recorded in Germany, they seem to be valid internationally too. Comparative studies are available, for instance, for Switzerland and France (26), as well as Sweden and the Netherlands (27). Both studies provide corresponding results. Although Swedish patients received more SBI than Dutch patients, only 6.0% received advice on how to reduce their alcohol consumption in comparison to 4.7% in the Netherlands. Moreover, the vast majority of 91% of French GPs did not use any test to screen for hazardous drinking in comparison to 77% of Swiss GPs. Even more daunting in this context may be the systematic review about strategies to improve the implementation of SBI (28) including results from 13 countries such as USA, Australia, GBR, and Spain showing that none of the tested strategies showed significant improvement regarding patient outcomes. We believe hence that the substantial number of barriers reported in previous studies indicates some 'deeper' conflict being rooted in the inherent characteristics of BI and we'd like to ask, is the time ripe now to quit forcing the implementation of BI into routine care?
It was not the intention of our study to show the superiority of pharmacological interventions to treat excessive drinkers and our results should not be understood as recommendation to prefer this type of intervention to BI. Pharmacological interventions to reduce the alcohol consumption are heavily disputed and a recommendation to use, for instance, anti-craving drugs would be inadequate considering the complexity of the issue. Nonetheless, GPs seem to perceive some advantages in using a pharmacological intervention and the question may be raised: What are those advantages and how can their understanding and a better understanding of GPs be used to improve BI? This question can hardly be answered easily. Some speculations may however be provided.
Firstly, GPs may prefer habitual behaviors (29). Habitual behaviors are usually well-practiced, well-structured, and are mostly associated with lower risk of failure. Accordingly, GPs may be used to ending their treatment with prescribing a drug, leading to a preference for such a behavior. Time pressure and workload, quite common in general practice, may even facilitate the use of habitual behaviors. Secondly, GPs may prefer to maintain social control over the course of consultation. A drug can be prescribed in a predictable structured act without the risk of inducing a less controllable dispute about idiosyncratic barriers and costs of reducing alcohol consumption and the vague issue of a healthier lifestyle. Finally, GPs may prefer the pharmacological intervention based on their belief that such an intervention may be more effective than a mere verbal intervention like BI. They may, for instance, perceive that BI's effectiveness depends primarily on a patient's self-control in contrast to a pharmacological intervention influencing the organism more directly. Risky consumers are not disobedient in general and may know about the unhealthiness of their behavior. But they may have failed several times to achieve a lower consumption level leading to resignation and a low self-efficacy regarding their self-control.
There are some limitations to the present study. First, the GP's readiness to implement BI was compared with their readiness to implement a pharmacological intervention newly available in Germany and not very common. This intervention was used in the control condition because of being also designed to reduce the alcohol consumption of excessive drinkers. The comparison of the two interventions may be perceived as problematic for some reasons. Both interventions, BI and the pharmacological intervention, may include overlapping elements such as information and advice. They may furthermore differ in several associated aspects beyond the pure mean of intervention such as duration, effectiveness, or perceived eligibility regarding the target group. GPs may hence tend to prefer one of the two also for personal reasons based on their - probably heterogeneous - past experiences with these interventions. Both interventions should not be understood thus as equivalent or comparable solutions for the problem of excessive drinking. Because of their multiple differences it's impossible - based on our results - to attribute the GP's lower implementation readiness to specific aspects of BI. But this was not the intention of our investigation. It was the intention to compare the effect of the intervention type - BI versus another intervention with other characteristics for the same target group - with the effect of the financial reimbursement to show the greater relevance of the factor intervention type. And this was done. A second limitation was the recruitment of GPs based on voluntary participation should be noted implicating selection bias and the risk of results with limited representativeness. Voluntary participation was implemented by the Ethics Committee of the University of Halle and is a widespread practice in this field of research. Thirdly, the following aspects may limit the generalizability of our results. As a predictor of implementation readiness, we used the self-reported intention of GPs. Self-reports, especially regarding desired or recommended behaviors such as BI, are possibly biased towards social desirability and would hence indicate a higher readiness than will be shown in routine care. However, using intention as a predictor of future behavior is an accepted practice in psychological research (30). Its predictive validity can be increased by using items specific to the context in question. This condition was met in our study. Moreover, GPs had to assess their implementation readiness based on written information about both interventions in the artificial context of case vignettes. Written information may provide a reduced and abstracted picture of the real conditions possibly inducing a weaker effect than real world conditions would do. Case vignettes are however a valid strategy to measure clinical competence (31). Finally, it should be mentioned that other factors may have influenced our findings, such as GPs’ concerns about the risk of stigmatization or a lack of skills. Future research should seek to clarify to what extent such factors, which have thus far been studied less, actually impair the GPs' readiness to implement SBI.
What are the implications for practical care? In Germany, primary care dependency prevention via SBI does not work effectively because all of the efforts to increase the GPs' use of SBI haven't been effective so far. Past research has emphasized that barriers may prevent GPs from implementing BI into routine care. It showed also that the elimination of barriers such as insufficient financial reimbursement did not lead to widespread implementation as may be expected (16) suggesting doubts regarding the belief that even a substantial increase of financial reimbursement would produce a primary care system with systematic screening of all patients and BI if screened positive. If inherent characteristics of BI are the problem its concept should be modified. Our results suggest the idea that inherent characteristics of BI may have a more negative effect on the implementation readiness of GPs than barriers such as an insufficient financial reimbursement. We show that GPs were more ready to treat an excessive drinker with a newly and even critically disputed pharmacological intervention than to use BI. Further research should investigate the crucial characteristics of BI to initiate a modification process finally leading to more effective primary care dependency prevention.