Abstract
Background Prevention has become a legal obligation for French GPs, since a law was passed in March 2002.
Aim Measurement and analysis of preventive procedures performed by French GPs.
Design of study Observational survey.
Setting GP surgeries in Puy-de-Dôme, France.
Method Doctors completed a questionnaire about their socioprofessional characteristics, and a researcher completed another questionnaire about preventive procedures performed on the last 15 patients seen by each GP. Twenty preventive services were evaluated and, for each service, medical records, targets, and objectives were defined according to the national preventive care guidelines. The gap between guidelines and practice was explained by doctor characteristics. Statistical analyses were performed using χ2 and logistic regression.
Results Representative samples of 179 doctors and 2453 medical records were randomised. Four preventive services were performed in more than 75% of cases, and the gap was explained by the salaried job the doctors had. Ten preventive services were performed in 25% to 75% of cases and the gap was explained by the medical software used. The six remaining services were performed in less than 25% of cases and no explanatory variable was identified.
Conclusion Sixteen preventive procedures were insufficiently performed. The more a preventive service is performed the more the gap will be explained by GPs' socioprofessional characteristics. The gap for a preventive procedure performed in 25% to 75% of cases was mainly explained by management of the medical records. A nationwide policy to improve prevention performance in general practice seems to be essential.
INTRODUCTION
Apart from their own intrinsic efficacy,1–15 preventive services are more effective when carried out by primary healthcare doctors.16–19 A number of financial incentives have been developed worldwide to improve GPs' prevention performance, such as pay for performance in the UK,20 and the US.21,22 In France, since the March 2002 law that added prevention to the Public Health Code and the August 2004 law that introduced GPs as ‘gate keepers’, prevention has become one of GPs' main tasks.
The current survey was based on the framework of two previous studies conducted in the university hospital of Clermont-Ferrand in 2004 and 2006. Fourteen preventive procedures and associated improvements were measured after various incentives were applied. These studies showed that the hospital partially failed to fulfil its prevention mission.23
The current study aimed to measure prevention performed by GPs according to the national preventive care guidelines and to assess the distance between guidelines and practice.
METHOD
This was an observational survey conducted in the French department of Puy-de-Dôme from July to September 2007 and is based on two questionnaires: one to identify GPs' socioprofessional characteristics the GP questionnaire and one to measure the preventive services they performed: the medical records questionnaire.
How this fits in
The medical records questionnaire, based on medical records data, is available to measure both the prevention performed at the hospital and at the GPs' surgery. The main results and determinants found in this study suggest that a national preventive policy strengthening GPs in the primary healthcare doctor role is essential to improve the health of the French population.
Inclusion criteria for medical records were: patient aged 18 years or over, examination by GP either at home or in the surgery, and follow-up by the same GP for at least 5 years. Exclusion criteria were: patient living in a nursing home, irrelevance of primary and secondary preventive services (palliative care), and a medical history of cardiovascular diseases (myocardial infarction, stroke, or peripheral arterial disease), cancer (breast, colon, cervix, or prostate), or dementia. Tertiary prevention was not studied because related preventive services and frequency are different.
Medical record assessment
The survey was based on auto-assessment by GPs of their medical records; they were interviewed by the researcher filling in the questionnaire. To be registered by the researcher, a preventive procedure had to be performed and noted in the medical records. No personal data were collected. One researcher filled in all the medical records questionnaire.
The preventive care guidelines of the French National Authority for Health (Haute Autorité de Santé) were followed to determine medical records targets and objectives of the 20 preventives services evaluated:6,24–36 diphtheria-tetanus-poliomyelitis, pertussis, and influenza immunisations; rubella serology and toxoplasmosis serology; tobacco consumption detailed in three categories (still smoking in the group previously known as smoker, detailed evaluation of both tobacco consumption and addiction using test); blood pressure check; screening for dyslipidaemia and diabetes; measures of weight, height, and body mass index (BMI); screening for breast, cervical, colorectal, and prostate cancers; alcohol consumption and addiction evaluation; and research for a cognitive disorder and a risk of falling.
GP and medical records questionnaires
The GP questionnaire (Appendix 1) evaluated doctors' socioprofessional characteristics that are likely to have an impact on prevention procedures and investigated: sex of GPs; date of graduation; type of practice; location of facilities; salaried job; medical records; continuing medical education; student training; any other activity linked to medicine; and how the doctors used their spare time. These characteristics were the explanatory variables. This questionnaire was tested on the first 10 GPs.
The medical records questionnaire (Appendix 2) was developed from questionnaires used for the prevention studies carried out in the authors' university hospital, and was adapted to general practice. The questionnaire was tested during the interviews that the researcher had with the first three included GPs.
Characteristics of concordance levels
Preventive activity performance was classified into three groups: ‘high’, ‘middle’, and ‘low’, with preventive procedures performed in more than 75% of cases, in 25% to 75% of cases, or less than 25% of cases respectively. A search was carried out to see if there was concordance between these three performance groups and the determinants that had previously been identified.
Randomisation methods used to select GPs and medical records
The regional union of GPs selected all professionally active GPs in 2007 with a surgery in the study area for at least 5 years (some GPs may belong to a regional union without medical activity or work for the drug industry). They assigned each one a code and randomised them. Then, they produced a main list of 200 GPs and a secondary list of 150 GPs according to the CNIL rules (National Committee for Ethics in Informatics). All GPs of the main list were contacted by the researcher. If a GP refused, another was contacted from the secondary list.
Patients were randomised by the researcher at each GP's surgery. All patients were checked to see if they met inclusion/exclusion criteria, and if so they were included in the study, starting with the last patient who was examined on their last working day. Inclusion stopped when 10 to 15 patients were included.
Number of participants needed
Using the highest prevalence among diseases targeted by the preventive services evaluated, 2305 participants were needed for a 2% precision measure and an α error of 5%. To take account of GPs' workload, the number of questionnaires that needed to be completed per GP was 10 to 15, and the number of GPs needed was 153 to 230.
Statistical analyses
Two descriptive analyses were performed: one relating to GPs' socioprofessional characteristics, and one relating to the prevention performances. As the variables were qualitative, χ2 tests and logistic regression were successively performed to compare physicians' attitudes toward prevention and to identify meaningful explanatory variables based on the calculation of the adjusted odds ratio. A meaningful threshold of 10% was chosen for the bivariate analysis in order not to neglect any interesting variables.
A meaningful threshold of 5% was chosen for the multivariate analysis and the likelihood ratio test. A logistic regression model was valid when the percentage of concordance was at least 60%. Statistical analysis was performed using SAS software. Finally, to give a broader view of factors related to the prevention performance, an aggregated score was calculated on Excel software for each explanatory variable, by combining the step of entry into the model and the number of variables in each model.
RESULTS
Among the population of 565 GPs selected and randomised, 350 GPs were included in the two lists and 261 GPs were contacted; 179 agreed to participate, 72 refused, and 10 were excluded (seven had undergone surgery within the last 5 years, and three were not practising as GPs in 2007). The sampling rate was 68.6%. The sample was representative of the population (Table 1). Characteristics of GPs, as explanatory variables, are shown in Table 2.
Table 1 Comparison of GP population and study sample characteristics
Table 2 Characteristics of GPs as explanatory variables
Randomisation of the medical records was in accordance with the study protocol, except for two GPs who preferred to receive the questionnaires by post and one who selected medical records by alphabetical order. The response rates ranged from 99.2% to 100% according to preventive services; 2453 medical records were included.
GPs' prevention performance is presented in Table 3. The ‘high’ group (≥75% preventive procedures) contained four procedures: blood pressure check, screening for dyslipidaemia and diabetes, and measurement of weight. The ‘middle’ group (25–75%) included 10 procedures: pertussis, diphtheria-tetanus-poliomyelitis, and influenza immunisations; rubella serology; still smoking or not, and detailed evaluation of tobacco consumption in the smoking group; measurement of height and BMI; and screening for breast, colorectal, and prostate cancers. The ‘low’ group (<25%) included six services: toxoplasmosis serology; cervical cancer screening; alcohol consumption and addiction evaluation; and research for a cognitive disorder and a risk of falling.
Table 3 Preventive procedures performed by GPs expressed in percentages, numbers of targeted patients (n) and 95% confidence intervals (CI)
The main results of the bivariate analysis revealed that GPs who graduated recently performed less prevention than those who graduated at least 15 years ago. Female doctors performed more preventive activities, except for BMI assessment. GPs with a salaried activity performed fewer preventive procedures, especially those working in institutions for older people. Despite better results when GPs had computerised medical records, none of the medical software available allowed sufficient collection of all the preventive services evaluated.
Logistic regression analysis showed 12 meaningful models out of 20 possible (Appendix 3). Overall, when a preventive service was classified ‘high’, two main explanatory variables were identified by the logistic regression: ‘salaried job’ the GPs had, and ‘having a cultural leisure activity’. GPs who worked in institutions for older people performed less prevention than those who had a salaried activity at a hospital. When a procedure was in the ‘middle’ group, the main explanatory variable identified by the logistic regression for five models on the seven models performed, was the ‘software’ the GPs used. Moreover, the detailed results showed that all modalities of this explanatory variable were meaningful. When a service belonged to the ‘low’ group, no determinant was found, so the results will be treated in specific articles.
Finally, the aggregated score was calculated. For example, for the ‘pertussis immunisation’ model, the first variable entered was ‘software’, so that scored one point (3/3); the second was ‘sex’ (scored 0.66 points, 2/3); and the last was ‘practising any kind of sport’ (scored 0.33 points, 1/3). This score was calculated for all explanatory variables and was summed per variable (Appendix 4). The variable ‘software’, which entered into 66.7% of the models, had the highest aggregated score (7.12). The variables ‘salaried job’ and ‘having a cultural leisure activity’, entered into 50.0% of the models, had a score that was higher than 4 (respectively 4.41 and 4.08). Finally, the variables ‘sex’, ‘practising any kind of sport’, and ‘reading without a link to medicine’, entered into 50.0%, 66.7%, and 41.7% of the models respectively, obtained scores between 3 and 4 (3.78, 3.63, and 3.46 respectively). The remaining variables, entered in less than 60.0% of the models, had a score that was less than 3.
DISCUSSION
Summary of main findings
Cardiovascular risk factors were thoroughly evaluated by GPs, apart from the annual evaluation of tobacco consumption, which needs to be enhanced. Results obtained for height and BMI measures were insufficient because of the increased prevalence of overweight (29.5%) and obesity (10.7%),38 and a prevalence of undernourishment of around 7–10% in general practice.39 All immunisations were insufficiently performed and also need to be improved. Nonetheless, three explanations are possible concerning pertussis immunisation: the guideline was relatively new (2004), the targeted patients were not defined precisely enough, and this vaccine is only available in association with the diphteria-tetanus-poliomyelitis vaccination which is performed every 10 years.
Preventive services targeting fertile women showed a lack of communication between GPs and medical gynaecologists. The worst result was obtained for the evaluation of alcohol consumption: 11.22% of patients were asked about their alcohol consumption, whereas the prevalence of alcohol misuse seen in a general practice is 30%.40 Bivariate analysis identified sex, date of graduation, salaried job, and medical record as the main determinants of difference between practice and guidelines.
A concordance was found between the level of performance and identified determinants. The more a preventive service is performed, the more the departure from guidelines will be explained by the GPs' socioprofessional characteristics, whereas the departure for a preventive procedure belonging to the ‘middle’ group was mainly explained by management of the medical records. The aggregated scores strengthened the evidence that all available software was irrelevant, despite a better performance of prevention when medical records were computerised.
Strengths and limitations of the study
The entire study protocol was followed as expected; both the sampling rates and the response rate were high; the sample was representative of the given population; and the number of participants included was sufficient. The survey has three limitations. First, GPs were free to choose whether to participate, so those who accepted probably performed better preventive procedures. Second, translation of the preventive care guidelines into the study questionnaire objectives was more difficult for brief interventions in the case of addiction than translation of the recommended frequency for the other services. Consequently, the study only measured if consumption of alcohol or tobacco appeared in the medical record. Third, a procedure was only registered if it was noted in the medical records, so the results are probably underestimated, especially for immunisations because some GPs only note them on a vaccination card.
Comparison with existing literature
Two previous French studies confirm the poor results for diphteria-tetanus-poliomyelitis immunisation.41,42 Concerning tetanus vaccination, the study results are significantly lower than those of the two previous studies (P<0.001) and higher for diphteria immunisation (P<0.001). Concerning poliomyelitis vaccine, the study result is significantly higher than that found by Guthmann et al (P<0.001).41
Two earlier studies of influenza vaccination found results that are comparable to the present one.43,44 The first had a lower performance than in the present study (P<0.001), but identified sex as a determinant of performance, similar to the bivariate analysis of this study.43 The second survey did not reveal a significantly different percentage of vaccinations; unfortunately, the multivariate analysis carried out was not powerful enough to identify determinants.44 The weighing gave a better score in the present survey than in the two previous biennial studies conducted in 1995 and 1997 (P<0.001).45
Results for cancer screenings, classified into the second group remain under the goal of the national cancer plan 2003–2007.46 When results of the present study were compared to the annual data of the regional screening agency,47 the percentage of this study was found to be higher (P<0.001). This difference can be explained by the fact that women who had a mammography outside the regional screening programme are included in the present study but not in the agency one. For colorectal cancer, the study results were also higher than those of the regional screening agency (P<0.001). As for both breast and colorectal cancers, the 2007 report of the regional screening agency noticed a difference of performance between cities and rural areas. The bivariate analysis used in this study did not produce the same finding. The present study showed better results than a previous one based on 2–3 years of screening (P = 0.047).48
Implications for clinical practice and further research
French GPs have a heavy workload, time constraints, and fee-for-service remuneration that does not include incentives for prevention. A better organisation of primary health care based on the standardisation of medical software for preventive procedures and the coordination of medical professionals may reduce GPs' workload; more preventive services could be delivered by nurses, especially in the new ‘Health House’, an establishment grouping together different health workers like GPs, nurses, physiotherapists, which is mainly developed in rural areas. A different method of payment, such as pay for performance or per capita with preventive goals, could be introduced to induce more appropriate incentives. In the UK, pay for performance for GPs seems to have produced significant improvements for preventive services and chronic diseases.49,50
GPs systematically perform four preventive procedures. In this case, a second salaried job in institutions for older people and a monthly cultural leisure activity are linked to the distance between practice and guidelines. Other preventive services are insufficiently performed, especially immunisations, measurement of height, screenings targeting fertile women, and evaluation of alcohol consumption. When a service was performed in 25% to 75% of cases, the medical software GPs used was the main determinant identified. No software allows a global and systematic collection of prevention data. These three explanatory variables also had the highest aggregated score.
Consequently, to achieve better prevention performance, a national public health programme should be developed, simultaneously taking into account initial training and appropriate continuing medical education for all doctors, improving the tools for managing medical records with the addition of another kind of payment, evaluating preventive performance, and training nurses to perform some preventive services.
Acknowledgments
With thanks to the 179 GPs who participated to this study.
Appendix
| GP characteristics |
| 1. | GP number | | |
| 2. | How many years ago did you graduate? | | |
| ☐5–15 | ☐15–30 | ☐more than 30 |
| Characteristics of your medical activity |
| 3. | Do you practise: | | |
| ☐solo? | ☐in a group? | |
| 4. | Do you also have a salaried job? | | |
| ☐yes | ☐no | |
| 5. | If ‘yes’ what sort of job is it? | | |
| ☐in an hospital | ☐in an infant and mother health centre | ☐other |
| 6. | If ‘other’, please specify | | |
|
| 7. | Your medical files are recorded (2 answers are possible): | |
| ☐on a computer | ☐on paper | |
| 8. | If your medical records are computerised, would you tell us which software you are using |
|
| 9. | Do you participate in continuing medical education (CME)? | |
| ☐yes | ☐no | |
| 10. | If ‘yes’ what kind of CME do you perform (several answers are possible) |
| ☐reading medical journals | | |
| ☐medical journals with reading test | | |
| ☐peer review | | |
| ☐meetings organised by local formal CME groups | |
| ☐seminars organised by regional or national formal CME groups | |
| ☐other | | |
| 11. | If ‘other’, please specify | | |
|
| 12. | Do you accept medical students for their first or second training period in your surgery? |
| ☐yes | ☐no | |
| 13. | Do you participate to another kind of activity linked to medicine? | |
| ☐yes | ☐no | |
| 14. | If ‘yes’, please specify: | | |
| ☐trade unionism | | |
| ☐teaching at the medical university of Clermont-Ferrand | |
| ☐politics | | |
| ☐medical council | | |
| ☐organisation of formal CME | | |
| ☐member of an organisation or an association | |
| ☐other | | |
| 15. | If ‘other’, please specify | | |
|
| Characteristics of your spare time |
| 16. | Do you read any books, papers, journals with no link to medicine? | |
| ☐yes | ☐no | |
| 17. | If ‘yes’, would you specify how many times per week or per month: | |
| ☐more than once a week | ☐once a week | |
| ☐at least once a month | ☐less than once a month | |
| 18. | Do you go to the cinema, theatre, concert, opera…? | |
| ☐yes | ☐no | |
| 19. | If ‘yes’, please specify how many time per week or per month: | |
| ☐more than once a week | ☐once a week | |
| ☐at least once a month | ☐less than once a month | |
| 20. | Do you practise any kind of sporting activity? | |
| ☐yes | ☐no | |
| 21. | If ‘yes’, please specify how many time per week or per month: | |
| ☐more than once a week | ☐once a week | |
| ☐at least once a month | ☐less than once a month | |
Appendix 1 The GP questionnaire
Appendix
| | Number of GPs interviewed | | | |
|---|
|
|---|
| Please circle the correct answer: Y (yes); N (no); NA (not applicable) | Number of medical records questionnaire | | | | |
|---|
|
|---|
| Preventive services | Targets | Aims | Measures |
|---|
| Vaccinations and serology tests |
| Diphtheria | All patients | Last immunisation <10 years | Y | N | |
| Tetanus | All patients | Last immunisation <10 years | Y | N | |
| Pertussis | Patients aged ≤40 years | Last immunisation <10 years | Y | N | NA |
| Poliomyelitis | All patients | Last immunisation <10 years | Y | N | |
| Rubella | Fertile women | 1 vaccination checked or serology performed | Y | N | NA |
| Toxoplasmosis | Fertile nulliparous women | 1 serology test performed | Y | N | NA |
| Influenza | Patients aged ≥65 years | Last immunisation <1 year | Y | N | NA |
| Cardiovascular risk factors |
| Age | All patients | Men aged >50 years or women aged >60 years | Y | N | |
| Tobacco | All patients | Last evaluation of whether a smoker or not <1 year | Y | N | |
| Blood pressure | All patients | Last measurement <1 year | Y | N | |
| Dyslipidaemia screening test | Patients aged <80 years and with one cardiovascular risk factor, or patients aged ≥80 years old if a treatment was started before the age of 80 years | Last dyslipidaemia screening test <3 years | Y | N | NA |
| Screening for diabetes | Patient with at least one cardiovascular risk factor | Last sugar blood sample <3 years | Y | N | NA |
| Anthropometry |
| Measurement of weight | All patients | Last measure <1 year | Y | N | |
| Measurement of height | All patients | Last measure <1 year | Y | N | |
| Measurement of body mass index | All patients | Last measure <1 year | Y | N | |
| Cancers |
| Breast cancer | Women aged over 50 and up to 74 years | Last mammography <2 years | Y | N | NA |
| Cervix cancer | Women aged over 20 and up to 74 years with no medical history of cervical surgery | Last smear test <2 years | Y | N | NA |
| Colorectal cancer | Patients aged over 50 and up to 74 years if the date of the last colonoscopy for another reason than colorectal cancer was >5 years | Last faecal occult blood screening <2 years | Y | N | NA |
| Prostate cancer | Patients ≥75 years | Last question and clinical exam <1 year | Y | N | NA |
| Addiction |
| Alcohol consumption | All patients | Last measure of consumption <1 year | Y | N | |
| | Last classification of sort of addiction used, for example CAGE assessment for alcohol abuse <1 year | Y | N | |
| Tobacco consumption | Smokers | Last measure of consumption in number of packets a year or in number of cigarette per day <1 year | Y | N | |
| | Last classification of sort of addiction used, for example the Fagerström test for nicotine dependence for physical <1 year | Y | N | |
| Cognitive disorders |
| Mini-Mental State Examination (MMSE) | Patients ≥75 years | At least one MMSE performed | Y | N | NA |
| Patients ≥75 years | Last MMSE <1 year | Y | N | NA |
| Risk of falling |
| Asking ‘have you ever fallen?’ | Patients ≥75 years | Last evaluation <1 year | Y | N | NA |
| ‘Up and go’ test | Patients ≥75 years | Last evaluation <1 year | Y | N | NA |
| Walking and talking | Patients ≥75 years | Last evaluation <1 year | Y | N | NA |
| Standing on one foot | Patients ≥75 years | Last evaluation <1 year | Y | N | NA |
| Sternal press | Patients ≥75 years | Last evaluation <1 year | Y | N | NA |
Appendix 2 The medical records questionnaire
Appendix
| 25% < performance percentage ≤75% | Performance percentage >75% |
|---|
| | |
|---|
| BMI (P<0.001) | Pertussis immunisation (P<0.001) | DTP immunisation (P<0.001) | Height | Influenza immunisation (P<0.001) | Tobacco ‘smoking group’ | Breast cancer (P<0.001) | Weight measurement (P<0.001) | BP measurement(P = 0.001) |
|---|
| 1. | Software (OR 73.78) | Software (OR 16.71) | Software (OR 3.04) | Software (OR 8.04) | Software (OR 10.40) | Sex (OR 2.02) | Peer review (OR 1.72) | Salaried job (OR 4.49) | Salaried job (OR 4.09) |
|
| 2. | Read (OR 2.94) | Sex (OR 1.83) | Medical record (OR 2.60) | Location of facilities (OR 3.05) | Sport (OR 2.26) | Medical practice (OR 1.83) | Read (OR 3.57) | Cultural activity (OR 1.72) | Cultural activity (OR 2.24) |
|
| 3. | Salaried job (OR 9.51) | Sport (OR 1.65) | Sex (OR 1.53) | Salaried job (OR 2.70) | Read (OR 2.13) | Cultural activity (OR 3.69) | Software (OR 0.46) | Medical record (OR 2.09) | Software (OR 18.70) |
|
| 4. | Medical students training (OR 1.71) | | Sport (OR 1.44) | Read (OR 1.81) | Regional/national CME seminars (OR 1.46) | Another activity linked to medicine (OR 1.67) | Cultural activity (OR 2.48) | Local CME meetings (OR 1.78) | Date of graduation (OR 5.02) |
|
| 5. | Location of facilities (OR 3.18) | | Location of facilities (OR 1.37) | Date of graduation (OR 1.92) | Medical students training (OR 1.57) | | Date of graduation (OR 1.61) | Read (OR 2.40) | Sex (OR 3.11) |
|
| 6. | Local CME meetings (OR 1.81) | | Another activity linked to medicine (OR 1.26) | Another activity linked to medicine (OR 1.56) | | | | Sport (OR 1.55) | Sport (OR 2.17) |
|
| 7. | Reading test (OR 1.48) | | Reading test (OR 1.25) | Sport (OR 1.50) | | | | Reading of medical journals (OR 1.49) | |
|
| 8. | Cultural activity (OR 1.75) | | | Regional/national CME seminars (OR 1.35) | | | | Peer review (OR 1.37) | |
|
| 9. | Medical record (OR 21.52) |
| |
| 10. | Regional/national CME seminars (OR 1.54) |
| |
| 11. | Date of graduation (OR 2.45) |
| |
| 12. | Another activity linked to medicine (OR 1.69) |
| |
| 13. | Sport (OR 1.37) |
| |
| 14. | Sex (OR 1.29) |
Appendix 3 Main results of the logistic regression, result of the likelihood ratio test (P), and adjusted odds ratio (OR)
Appendix
| Explanatory variable | Percentage of model in which the explanatory variable is meaningful | Aggregated score |
|---|
| Software | 66.7 | 7.12 |
|
| Salaried job | 50 | 4.41 |
|
| Having a cultural leisure activity | 50 | 4.08 |
|
| Sex | 50 | 3.78 |
|
| Practising any kind of sport | 66.7 | 3.63 |
|
| Reading with no link to medicine | 41.7 | 3.46 |
|
| Location of facilities | 33.3 | 2.70 |
|
| Date of graduation | 58.3 | 2.56 |
|
| Peer review participation | 25 | 2.13 |
|
| Medical records | 25 | 2.01 |
|
| Another activity linked to medicine | 33.3 | 1.58 |
|
| Regional or national CME seminars | 33.3 | 1.43 |
|
| Local CME meetings | 16.7 | 1.23 |
|
| Practice | 16.7 | 1.08 |
|
| Medical students training | 16.7 | 1 |
|
| Medical journals with reading test | 16.7 | 0.68 |
|
| Reading medical journals | 8.3 | 0.25 |
|
| Other type of CME | 8.3 | 0.14 |
Appendix 4 Results of the aggregated scores for all the explanatory variables tested, classified from the higher score to the lower score
Notes
Funding body
Faculty of Medicine, University of Auvergne.
Ethics committee
The project was presented and accepted by the ethics committee of the general practice department of the Faculty of Medicine at Clermont-Ferrand.
Competing interests
The authors have stated that there are none.
- Received April 2, 2010.
- Revision received May 13, 2010.
- Accepted July 19, 2010.
- © British Journal of General Practice, 2011
