Meeting of Investigators, Geneva, Switzerland
24-27 May, 2002
(Meeting to be held at Hotel Château de Coudrée, France)
This progress report is partly a description of what was planned or in it’s making at the Brussels meeting. Things have changed since then, however, and the project has taken on a new direction, which is described below.
In Spring 2000 we decided upon a slightly revised implementation and research plan in the Danish project. This was due mostly to the fact that we have strong evidence to believe that the Danish GPs themselves will regard the routine screening for risky drinking as inappropriate, and maybe even counterproductive.
This meant that screening would not be promoted as a routine procedure. It should be presented as an option, e.g. as screening-like procedures in broad preventive consultations. Instead, diagnostic indications of heavy, risky or harmful drinking would be highlighted, as would methods for overcoming the patients’ denials of heavy drinking or alcohol problems, methods for motivational interviewing, for counseling, for referrals to specialist treatment etc.
Thus, the content of our intervention, i.e. the knowledge and skills we wanted to implement in general practice, was the identification, the counseling and treatment and the proper referrals of patients with heavy drinking or problem drinking – in short Alcohol Intervention (AI).
However, this slight adjustment of the original content did not change the implementation and research design very much.
Preparation and planning of the intervention
During May and June 2000 we ran 2 focus groups with a total of 15 GPs participating, each group lasting for 2 hours.
The purpose of the focus group was to get information on GPs’ preconceptions of alcohol problems and prevention, their attitudes to and experiences with alcohol intervention, counseling, treatment etc. We also wanted to know what kind of barriers they perceived in the patients to take part in counseling, referrals etc. And are there any structural barriers that make an implementation of the EIBI or AI a difficult business? Moreover, we wanted to find out what knowledge and skills the GPs need to be able to deliver an optimal performance vis-à-vis heavy drinkers and alcohol abusers. And how would they prefer to have CME-activities delivered/implemented?
The main findings from the FGs were presented at the Brussels meeting. We concentrated on factors that the GPs perceive as barriers to a smooth and efficient implementation of prevention and handling of heavy and harmful drinking in general practice.
The implications for our demonstration project were discussed n Brussels.
Because we have had the feeling – and this was confirmed in our FGs – that the GPs are at unease about the reactions of their patients if they bring up alcohol as an issue during the consultation, we decided to investigate whether patients really dislike their doctors asking about their drinking habits and other life style issues.
First we tried to construct a small questionnaire (22 questions) as the data collection instrument. However, through two small pilot/validations we realised that this method was not feasible. Instead, we set a focus group with 8 patients from a general practice in Frederiksborg County addressing the same issues as the intended questionnaire. From the focus group we got the impression that patients do accept their GP to ask them questions about their alcohol consumption pattern when the problem presented might be alcohol-related. But screening-like procedures without relevance for the problem/disease would not be popular.
CME in small-groups
In Denmark, there is a network of small CME groups. The members of these groups themselves decide which topics they want to deal with – and how they will do that. The number of members is normally between 5 and 12. Some of the groups have a permanent character, while others are ad hoc-based and devoted to a special problem, so that when the problem or issue has been emptied, the group is dissolved.
In Frederiksborg County, which is our local intervention area for the Alcohol Project, more than 90% of the GPs are members of one or more groups.
Participants in the GP focus groups suggested that we used the small CME groups as arenas for disseminating knowledge and know-how and develop or train skills in handling risky drinking or alcohol abuse. This is one of the reasons that we prepared – and offered – an educational package to the small CME groups.
Another reason is that much implementation research within the medical field favours the use of local groups. New knowledge and new skills can be discussed with colleagues in the local medical culture, and this is important because a consensus here is paramount for the uptake in daily routine practice. Innovations, clinical guidelines etc. will get a more rapid footing in daily practice when the target group gets a sense of ownership through discussing and accepting the innovation, through translating or transposing a guideline to local needs and sentiments.
The educational package consisted of two meetings, each lasting two hours, and it was offered to the 21 small CME groups in Autumn 2001. The marketing of the CME-project was done through an article in the local GP journal, through letters of invitation to each of the small-groups, and indirectly through the medical audit registration in November 2000.
The doctors were told that participation would enhance their skills and competence in handling alcohol issues and motivating patients etc. Participation was free, i.e. the GPs did not have to draw resources from their CME account.
Before launching the package in the first CME group we made an extension of the offer. Each of the participating GPs could have a professional actor coming to the consultation room simulating a patient with a problem or a disease that might be alcohol-related. The simulated patient in all cases had a rather high alcohol consumption level (but this was not told to the GP unless s/he asked – or asked in a patient-centered way that got the patient to drop resistance and be sincere).
The simulated patient made reservation for a consultation the same way ordinary patients do. When coming to the consultation room the actor revealed that s/he was the actor. The consultation lasted about 15 minutes (which is normal for a consultation) and it was audio taped.
The tape was transcribed and used for feedback to the GP and for teaching and training at the second 2-hour meeting in the CME small-group.
Both sessions in the group were focused on topics like:
- – when should alcohol consumption be an issue?
- – the transtheoretical model of change
- – the spirit of motivational interviewing
- – raising the issue of alcohol consumption
- – avoiding resistance
- – other MI techniques
The first meeting was focused on having GPs think about their practice and attitudes towards handling risky drinkers and alcohol problems, and to realise what are their problems and needs in relation to this category of patients and this task. Thus, the meeting should function as an eye-opener (a starter that makes the GPs consider their own practice regarding their handling of heavy drinkers) and help the participants to define needs (knowledge, skills). Techniques of motivational interviewing and health behaviour change counseling were shown and discussed.
During the second meeting the GPs’ experiences with the simulated patient were discussed, and the transcriptions were used for this and for teaching, as mentioned above. Other specific issues and needs raised during the first meeting were also addressed.
However, in spite of what we regarded as a “special offer” and in spite of our marketing activities the participation rate has been rather low. We shall return to that below under the heading Problems and Miscellaneous.
The year 2001 was scheduled as our year of intervention where all CME groups were supposed to have had the 2 x 2 hour sessions. During that year we would monitor the project carefully, making it possible 1) to know how the implementation actually runs; 2) to allow for adjustments and removal of unforeseen barriers and problems; 3) to know which parts of the intervention are the most appropriate and effective ones.
We did not plan for a full monitoring. Instead, we would be less ambitious and have a panel of GPs and their partners (out-patient clinics, psychiatric wards etc.) whom we should visit regularly and ask relatively systematic questions about problems experienced, e.g. in asking patients about their alcohol consumption, in breaking the denials of problem drinkers, in the cooperation with referral institutions, about information and written materials needed.
After having participated in the small-group based CME and having tried out and practiced the new knowledge and methods in daily practice, the GPs might have questions and problems they want to discuss with each other and with those responsible for the intervention and the project. Therefore, we would try to set up meetings where such matters would be discussed and the GPs could be advised. We expected to have such meetings a couple of times during the year of intervention.
It was considered to supply the other elements of the intervention with an Alcohol Project-related homepage that is available only to the GPs in Frederiksborg County (this restriction to GPs in this county is made to avoid contamination to the control areas). The homepage is to bring new information of current interest and relevance for the handling of heavy drinking and alcohol problems; it should contain diagrams, forms, and other tools for downloading and use during consultation; there should be a discussion database; and a FAQ-site could be included.
Plans and Methods for the Evaluation
Of course, one wants to know whether this approach of disseminating knowledge and know-how has any impact on the performance of the target group, the GPs. Therefore, we had planned an outcome evaluation and a process evaluation.
We would use a quasi-experimental design with pre- and post-measurements, having the Frederiksborg County as the intervention group, and 5 other counties being the control group. In the intervention group all GPs were asked to fill in a medical audit registration form for all their adult patient consultations during a 2-week period in November 2000; and this procedure should be repeated after the intervention period (in the beginning of 2002). We expected that 50 of the 230 GPs in the county were willing to fill in the registration form twice. To have a control group of the same size we approached 275 GPs selected randomly from 5 randomly selected counties. (The reason for not selecting only one county for the control group was to avoid that the chosen county would turn out to be very active in this particular field during the intervention year, thus diminishing the ability to find a possible intervention effect).
The categories in the medical audit registration form were built partly on the focus group discussions, partly on the goals of the intervention itself (more activities of identification, assessment, motivational interviewing, counseling, referrals etc.).
Thus, outcome is here defined as the clinical performance of the GPs. Patient outcome measures such as morbidity, mortality, DUI, arrests for drunkenness, referrals to specialist treatment etc. were not considered appropriate because of small numbers as well as a certain time lag in such parameters, not to speak of possible confounders. Nevertheless, we would look into the available statistics at the end of the project to check whether our expectations were confirmed or not.
In order to interpret and qualify any positive or negative effect evaluation results, a process (or implementation) evaluation should be undertaken, the aim of which was to describe to what extent – or whether – the programme was implemented as planned, whether it ran smoothly, or whether there were any barriers that might explain sub-optimal outcomes.
The above-mentioned monitoring of the implementation process (for adjusting the intervention) should also work as a data collection for the process evaluation. The data and experience from the monitoring should be supplemented by individual qualitative interviews with GPs and other key persons and by focus groups with GPs as participants.
Problems & Miscellaneous
Having a low response rate at the pre-measurement
The medical-audit like 2-week registration form was sent to all GPs in the County of Frederiksborg (N=246) and a random sample of GPs in 5 other counties (N=275) – the former being the intervention group and the latter the control group. As an incentive the GPs were offered a gift if they would fill in the forms (a gift token of two compact discs, valid 34 EUROS).
However, the response rate was rather low. In Frederiksborg County we had a 34% response rate (84/246) and in the control group 22%. This problem is accentuated by the fact that only about half of those participating in the training sessions in the intervention group had beforehand made the medical audit registration.
Not getting GPs interested
Another problem that our project had run into was a minimal interest from the GPs to engage in the CME-activities offered in this project. Five groups asked us to come. The attendance rate, however, was not very high. About half of the members of these CME small-groups have participated in both meetings.
No quantitative evaluation
Therefore, with the prospect of having very few participants in the intervention group (half of whom have not filled in the medical audit registration in November 2000) we faced a serious problem of statistical power. At best, maybe, we must 25 GPs in the intervention group that have participated in the intervention and registered patient contacts both before and after the intervention year. Thus, remembering that this is a demonstration project that is to document the impact of a broad implementation of EIBI on the performance of GPs within a region, in casu: Frederiksborg County, we were in trouble. Repeating the medical audit in the beginning of 2002 made no sense.
An alternative could have been a qualitative, in-depth interview study with those who actually have participated in the CME-activities.
Reframing understanding of alcohol issues
As promised during the meeting in Bled January 2000 we have contributed to the development of a questionnaire by testing the one presented in Bled. An academic detailer in one of the counties of Denmark has at the end of his visits to the GPs asked each of them to fill in the questionnaire immediately. All 30 questionnaires have been sent to Newcastle for analysis. But according to Nick Heather, the Danish data did not have more sense or internal consistency than those from Spain (and UK?).
Reframing of the understanding of alcohol problems should take place in the small-group based CME activities where the GPs’ understanding and attitudes should be highlighted and – if needed – challenged and changed.
New project plan August 2001
The progress and problems described above made us change the project. The quantitative outcome evaluation has been abandoned, and the project is now turned into a smaller method-developing project where we try to get GPs interested in participating in workshops in their own practices and let them define themselves their needs for training within the area of health behavior change counseling and motivational interviewing.
As a result of the customisation we will not choose alcohol or hazardous and harmful drinking as the only topic. The GPs tell us that if the project aims at tobacco, exercise, overweight as well they would be more interested. We were convinced that alcohol problems do not “sell tickets”. Since we have offered to teach and train more generic skills usable within a much broader scope (but including alcohol issues). The project now has the following outline.
The teaching and training sessions are different from the ones already tried out (described above) by being less top-down and less directive. While hitherto we have taught skills that we as teachers regarded important and useful for the participating GPs, we now let themselves find out and decide what are their weaknesses that they want to improve/eradicate.
GPs are offered a multi-stepped workshop that allows them to find out about their wants and wishes and to get training tailored for this. Catch word-like one could say that the clinician is the expert in what and how to learn, and the trainer’s role is to facilitate and to provide useful ideas and skills. Training should start with what the GPs actually do in their everyday work setting.
And when the GPs are asked to decide what scenarios to focus on (i.e. clinical area – smoking, alcohol, nutrition, diabetes etc. – and communication skills), we believe that this approach will be far more interesting for them than the usual top down model where experts come and tell them what the ought to learn.
The course of the workshop starts with an introductory meeting performed by the project leader Thorkil Thorsen (TT) in the general practice clinic or health center. The participating GPs agree upon a topic for the consultations with simulated patients that will come a few days later. These consultations are audio taped, and the transcriptions are sent to the doctors for them to consider own weaknesses and strengths and issues for the training session 3 days later. The training sessions are run by TT and Sverre Barfod (SB). This procedure is repeated twice focusing on subsequent scenarios with the same simulated patient. After the third consultation and return of transcription there will be debriefing seminar where experience are summarised.
In overview the steps run like this:
(1 hour in own practice)
with simulated patient
GP gets transcription
Feedback-meeting: Discussing experiences;
teaching and training (2 hours in
with simulated patient
GP gets transcription
Feedback-meeting: Discussing experiences;
teaching and training (2 hours in own practice)
with simulated patient
GP gets transcription
training (2 hours in
Day 1 Day 4 Day 7 Day10 Day 13 Day 16 Day 19 Day 22 Day 25 Day 28
A full stepped workshop has a 6-week course and includes 4 meetings with trainers and GPs (a total of 7 hours) and 3 consultations (each lasting 15 minutes).
GPs are paid for their participating in meetings and get a standard fee for each simulated consultation.
We had planned to run parallel workshops but due to the workload especially for the secretary it is the best waiting for a course to be through before starting the next with other participants.
The introductory and debriefing meetings are audio tape-recorded for evaluatory purposes allowing comparison of GPs’ self-reported communication problems and appraisal of own skills etc. before and after. At the same time participants also appraise the workshop and the didactic methods used so that the concept can be improved continuously.
The consultation transcripts are compared in order to reveal possible improvements during the workshop period in using motivational interviewing health behaviour change counselling.
This evaluation will be supplemented with yet another simulated consultation a 2-4 months after the debriefing seminar. This consultation should then have the same standardized simulated patient and presented problem similar but not identical with that used at baseline. This can perhaps prevent confounding interpretation in the minds of both the simulated patient and the doctor and hopefully allow for a direct comparison of competence levels.
It is possible to evaluate the transcriptions in a quantitative way using principles developed by Rollnick, Miller, Botelho and others. For this purpose we are inspired of the paper of Miller: Motivational Interviewing Skill Code (MISC) (http://www.motivationalinterview.org/training/ MISC.PDF)
The danish Alcohol Group are discussing this now. An example of a scheme for analysis is attached.
The evaluation is available in danish (2003):
Thorsen T, Barfod, S, Beich, A. Fra brydning til dans. Rapport fra et forsknings- og udviklingsprojekt om
kommunikationstræning i almen praksis. 2003. Central Forskningsenhed for Almen Praksis, København.
Publication is going on and then a summary in english will be available.
One could wonder whether this reformulated and reorganised Danish project still fits into the WHO phase IV club.
On the one hand the workshops make up an intervention. And we shall evaluate the process and outcome of each workshop-series through interviews and observation (and use this for further development, refinement, and marketing of the workshops). We still aim at getting tools and skills for EIBI widely implemented in the local area.
On the other hand, we do not stick rigidly to the phase IV protocol. We will not be able to demonstrate whether or how it was possible to get EIBI integrated as a routine practice in GPs.
There will be a quantitative evaluation on the GP’s performance of the motivational skills taught.
Perhaps the costumization to the danish conditions has forced the project too far from the phase IV protocol? But the process going did not work and made a reframing and redirection of the project necessary.
However, the training in motivational interviewing skills has a high priority in the danish health authorities and has great interest in the CME organs of Family Medicine.
We, too, see it as the only realistic way to get danish doctors to talk more alcohol problems with their patients
Sverre Barfod, Annelise Zachariassen.
May 20th, 2002
AI Alcohol Intervention
CME Continuous Medical Education
DUI Driving Under Intoxication
EIBI Early Identification and Brief Intervention
FAQ Frequently Asked Questions
FG Focus Groups
GP General Practitioner
MI Motivational Interviewing
TT Thorkil Thorsen. Senior researcher. Project leader
SB Sverre Barfod, General Practitioner, Leader of investigators in Denmark.
 Abbreviations: See end of document
 We have here been inspired by Stephen Rollnick and his colleagues in Cardiff.