ABSTRACTS

European General Practice Research Network

Abstracts from the EGPRN meeting in Bertinoro, Italy, 7–10 May 2009

Friday, 8 May 2009

Location: University Residential Centre Bertinoro

Address: Via Frangipane, 6, 47032 Bertinoro (FC), Italy

08.50–09.10 1st Keynote Speaker: Professor Jean-Karl Soler, MSc, Malta

Theme: Research using EPR: The potential of ICPC in primary care research

The theme of the EGPRN Bertinoro, Italy, meeting in May 2009 is “Data analysis of electronic patient record databases in general practice”. The theme is important because data from electronic patient records are a potentially valuable source of empirical information about day-to-day general practice, and which are collected routinely about millions of patients around the world every day. Such data should inform quality assurance, education, training, and research. The potential applications of such data and its analysis are truly endless, and include epidemiology of general practice (studying patients’ reasons for encounter, doctors’ diagnostic labels, and interventions to provide care), computerized decision-support systems, drug adverse-event surveillance, studying the process of diagnosis in general practice, studying co-morbidity, and many others. We have recently published (1) on the necessity of precision in measuring events in such databases in order to capture the maximum amount of information with reliability and validity. The International Classification of Primary Care fulfils all the requirements of a modern and precise tool for organizing data in primary care for clinical care, practice management, and analysis at the micro, meso, and macro levels for research, education, and quality assurance. Its purpose is to order the domain of family practice in the format of episodes of care. It provides a single terminology for the patient's RFE and the family physician's diagnosis, thus representing both sides of the same coin. It captures the changes (transitions) in the content of episodes of care over time. It follows strict taxonomic rules, and so its classes are mutually exclusive. It offers—if possible—one class for common (occurring >1 per 1000 patient years) reasons for encounter and diagnoses. Less common classes are included in “ragbags”. Its biaxial structure (chapters for body systems/problem areas and components identical throughout all chapters) results in three-digit mnemonic, alphanumeric codes. Its reliability and validity are supported by its coding rules and a growing comparative international database. In the coding process, localization takes precedence over aetiology. Symptom diagnoses take precedence over disease diagnoses that are uncertain (i.e., that do not fulfil the inclusion criteria). It does not cater for mind-body metaphors: “psychosomatic” and “somatoform” disorders are not included.

The presentation will look at data collected with ICPC in an episode-of-care model from three countries, and analyse utilization, frequencies of reasons for encounter, interventions and diagnoses (episodes), and research relationships between reasons for encounter and episode titles (research into diagnosis) and between episode titles themselves (co-morbidity) using pre- and post-test probabilities expressed as odds ratios.

Reference

1. Soler JK, Okkes I, Wood M, Lamberts H. The coming of age of ICPC: celebrating the 21st birthday of the International Classification of Primary Care. Fam Pract 2008;25:312–7.

Friday, 8 May 2009

Location: University Residential Centre Bertinoro

Address: Via Frangipane, 6, 47032 Bertinoro (FC), Italy

09.10–09.30 2nd Keynote Speaker: Professor Marco Cambielli, Italy

Theme: Problems related to data collection from electronic patient records in primary care: Reliability of data

Family practitioners and other staff working in primary care require comprehensive and accurate data on patients at the point of care, if they are to provide high-quality health services to their patients. Electronic patient records (EPR) are an effective method of achieving this objective, by dispensing with the need to use difficult-to-access, and often illegible, paper-based records. Electronic patient records underpin many information technology initiatives in primary care, such as screening for identifying patients at high risk of cardiovascular disease, call-recall systems for asthma and other long-term disease management programmes, computerized decision support systems for prescribing, electronic ordering of tests, and electronic referral systems to secondary care. These are all, however, dependent on comprehensive and accurate coded data.

There are known to be large variations in the accuracy and completeness of the clinical information stored in electronic patient records. In some systematic reviews, it was found that the recording of consultations was generally high (typically greater than 90%), but assigning a morbidity code during each consultation was more variable. Some authors identified major omissions in the diagnoses. The recording of diagnoses in primary care was less complete and, when a diagnosis of disease was recorded, it was generally less detailed than in the data held by various national registries.

Soler et al. (1) described the progress of the International Classification of Primary Care (ICPC): the wide use of the ICPC facilitates inter national comparisons of clinical practice and coding in primary care. Key areas for further work are the development and evaluation of data quality standards for use in electronic patient records, and the evaluation of methods for improving data quality. Electronic patient records offer enormous benefits, not only for patient care but also, when aggregated, for secondary analysis; and when linked with other health and social care datasets, for outcomes mea surement, quality improvement, public health surveillance, and research. These benefits cannot be fully realized without high-quality data. Systems in which “free text” natural language (reflecting clinicians thought processes) could be coded and used for additional functionality are still far in the future (2). Markers of quality should comprise internal reference standards based on objective and diagnostic EPR elements that have high positive predictive value (3).

References

1. Soler JK, Okkes I, Wood M, Lamberts H. The coming of age of ICPC: celebrating the 21st birthday of the International Classification of Primary Care. Fam Pract 2008;25:312–7.

2. Majeed A, et al. Accuracy and completeness of electronic patient records in primary care. doi:10.1093/fampra/cmn047.

3. Thiru K, et al. Systematic review of scope and quality of electronic patient record data in primary care. BMJ 2003;326:1070–4.

Saturday, 9 May 2009

Location: University Residential Centre Bertinoro

Address: Via Frangipane, 6, 47032 Bertinoro (FC), Italy

08.45–09.05 3rd Keynote Speaker: Professor

Giampiero Mazzaglia, Italy

Theme: Computerized patient databases ingeneral practice: Experience in Italy and Europe

Before the advent of automated databases, the identification and appropriate follow-up of large cohort of patients required a major effort to ensure complete and accurate collection of the required information. To date, the increasing number of general practitioners (GPs) who use computers for administrative purposes, to keep patient records, and to generate prescriptions allows researchers to con duct epidemiological studies in a highly efficient manner.

The Health Search-Thales database (HSD) was set up by the Italian College of General Practitioners (SIMG) in 1998 with the primary aim of carrying out epidemiological research in a community-based setting. The HSD contains data from the computer-based patient records from a selected group of GPs distributed across Italy who voluntarily agreed to collect data for the database and attend specified training courses. The database contains information on the age, gender, and identification of the patient, and GP registration information, which is linked to prescription information, clinical events and diagnoses, hospital admission, and causes of death.

The HSD has been the source for a number of peer-reviewed publications on the prevalence of disease conditions, drug safety, and prescription patterns in Italian primary care. At national level, the information from the HSD is used for developing prescribing indicators for appropriate use of drugs. Such indicators are published annually by the National Department of Health and National Medication Agency in its annual report. At international level, the HSD is involved in several projects, including the EU-ADR project funded by the European Commission within the 7th Framework Programme. This project federates the clinical records of 30 million Europeans in Denmark, Italy, the Netherlands, and the UK, with two main aims: 1) to discover new drug-safety signals, and 2) to substantiate signals, using causal reasoning based on Bradford-Hill criteria, semantic mining of the bio-medical literature, and computational analysis of biological and chemical information (drugs, targets, anti-targets, pathways, etc.). In conclusion, although proper utilization of such databases requires multi-disciplinary skills, this information is an invaluable source of data for epidemiological studies.

Saturday, 9 May 2009

Location: University Residential Centre Bertinoro

Address: Via Frangipane, 6, 47032 Bertinoro (FC), Italy

09.05–09.25 4th Keynote Speaker: Professor Mike Pringle, Nottingham, United Kingdom

Theme: Unlocking the potential of the GP electronic record: Experience in the UK

In this presentation, I will explore the UK experience of using GP electronic records for research and health services delivery. In particular, I will cover:

1. Data quality: supporting and improving it using

PRIMIS +

I am Service Director of PRIMIS+ , a state-funded initiative to cascade training to family practices in order to improve clinical data recording and use. PRIMIS + provides a voluntary analysis tool for practices to assess their data quality (CHART) and a facility to compare their data quality with other practices (CHART Online). The evidence on data quality will be presented.

2. Service delivery support: through QSurveillance and QRisk

These two projects are run in my department in Nottingham. QSurveillance can monitor epidemics (a flu pandemic, etc.) and the effects of disasters (the Buncefield fire, for example). QRisk is a new risk score for cardiovascular disease which includes deprivation.

3. Research: using large databases such as QResearch, GPRD, THIN, and UK Biobank

These databases are gathering more extensive and more useful data. I will give an update on progress in UK Biobank, of which I am on the Board, towards its recruitment of 500,000 people.

4. European collaboration: the results of eHID

This recently completed research project looked at the potential for using GP data sets for comparison of prevalences of common diagnoses across Europe. I will illustrate some of the findings.

Through these four examples, I will try to illustrate the potential and challenges in using routine GP data for secondary purposes.

Presentation 1: Theme paper

Friday, 8 May 2009, 09.30–10.00

Prevalence and management of chronic heart failure (CHF) in a network of electronic care record (ECR) databases of 114 GPs in the Veneto region

Maurizio Cancian¹, Alessandro Battaggia¹, Mario Celebrano¹, Francesco Del Zotti¹, Bruno Franco Novelletto¹, Mario Saugo²

¹SVEMG, Scuola Veneta Medicina Generale, SIMG Società Italiana di Medicina Generale, and ²Local Health Authority, Verona, Italy. Tel: + 39 045 596 772; E-mail: [email protected]

Background: CHF is connected with increasing prevalence and new tools for diagnosis and treatment, but few studies are conducted in the general practice (GP) setting.

Research question: What is the prevalence and management of CHF of 114 GPs in the Veneto region?

Method: A baseline audit involved 114 GPs, divided into 20 groups, of all local districts in the Veneto region, using the same ECR, assisted centrally with an extraction procedure for CHF ICD-9 codes and an SQL query for finding “suspect CHF”. Every GP matched these lists with criteria from the SIGN guideline 2007, chosen after “agree” scoring.

Results: 1905 CHF patients were extracted from 155,617 patients, with a prevalence of 1.2% (95% CI 1.17–1.28) and considerable variability among the 20 GP groups. Of the CHF patients, 1213 (64%) were extracted as ICD-9 code, 692 (34%) after the SQL query. Echocardiography (ECO) was done in 57% of cases and ejection fraction measured in 31% of patients with an ECO, with ECG in 76% of patients. At least one annual blood pressure measurement was taken in 89% of patients, and electrolytes + creatinine in 88%. Considering 611 patients with spironolactone and ACE/ARBs, 10% had no annual recording of creatinine + potassium, and 30% just 1 per year. There was no annual recording of weight in 43% of patients. ACE-inhibitors or ARBs were used in 77% of CHF patients, beta-blockers in 46% (first-choice beta-blockers: bisoprolol, carvedilol, nebivolol, in 36%). Even for follow-up and for drugs, there was considerable variability.

Conclusions: Data show the strength of integrating databases in GP for the evaluation of CHF, as well as some weaknesses. There is room for improvement in the diagnostic process (coding, diagnostic criteria), in follow-up, and in therapy. Our audit, based on ECR data collection and on local groups, and the observed variability lead us to discuss the internal and external levels of change needed.

Points for discussion: 1) Are there similar studies in other countries? 2) How is the gap between the diagnostic process and new diagnostic tests and drugs in CHF perceived in European GP?

Presentation 2: Theme paper

Friday, 8 May 2009, 10.00–10.30

Predictability of persistent frequent atten dance: An historic 3-year cohort study

Frans T. Smits, H. J. Brouwer, H. C. van Weert, A. Schene, G. ter Riet

Department of General Practice, Division of Clinical Methods and Public Health, PO Box 22700, 1100 DE Amsterdam, the Netherlands. Tel: [phone]; Fax: [phone]; E-mail: [email protected]

Background: Few patients who attend their general practitioner (GP) frequently continue to do so. While transient frequent attendance may be readily explicable, persistent frequent attendance often is not. Persistent frequent attenders increase GPs’ workload while reducing work satisfaction. It is neither reasonable nor efficient to target diagnostic assessment and intervention for transient frequent attendees.

Research question: Is it possible to construct a prediction rule for selecting persistent frequent attenders using readily available information from GPs’ electronic medical records?

Method: An historic 3-year cohort study using data on 28,860 adult patients from 2003 to 2005. Frequent attenders were patients whose attendance rate ranked in the (age- and sex-adjusted) top 10% during 1 year (1-year frequent attenders) or 3 years (persistent frequent attenders). Using bootstrapped multivariable logistic regression analysis, we determined which predictors contained information on persistent frequent attendance.

Results: Out of 3045 1-year frequent attenders, 470 (15.4%) became persistent frequent attenders. The prediction rule could update this prior probability to 3.3% (lowest value) or 43.3% (highest value). However, the 10th and 90th centile of the posterior probability distribution were 7.4% and 26.3%, respectively, indicating that the model performs modestly. The area under the receiver operating characteristic curve was 0.67 (95% CI 0.64–0.69).

Conclusions: Among 1-year frequent attenders, six out of seven are transient frequent attenders. With the present indicators, our rule performs modestly in selecting those at risk of becoming persistent frequent attenders.

Points for discussion: What, not yet registered, factors can help to predict persistence of frequent attendance? Are persistent frequent attenders a group of patients suitable for preventive medical care? Is “persistent frequent attendance” more than (1-year) frequent attendance a useful concept in general practice?

Presentation 3: Theme paper

Friday, 8 May 2009, 11.00–11.30

Time trends in the incidence of peptic ulcers and oesophagitis between 1994 and 2003

Stefaan Bartholomeeusen, J. Vandenbroucke, C. Truyers, F. Buntinx

Department of General Practice, KU Leuven, Kapucijnenvoer 33, blok J, bus 7001, 3000 Leuven, Belgium. Tel: [phone]; E-mail: [email protected]

Background: The incidence of peptic ulcers has shown considerable fluctuation in the past. Most epidemiological studies have been performed in hospitals on patients with complications. General practice is the appropriate location to investigate the incidence of peptic ulcers. A gastroscopy can be performed ambulatorily and drug treatment started.

Research question: To examine time trends in incidence of peptic ulcer and oesophagitis and prescription of acid secretion-inhibiting drugs in the period 1994–2003.

Method: Retrospective cohort study based on the Intego database, composed of data extracted from the electronic medical records of 55 GPs in 47 general practices, spread over the whole Flanders region. Diagnoses are registered by means of program-specific keywords that are converted into the ICPC-2 classification in the central database.

The practices are selected on quality of registration. The population of the database matches the Flemish population. Incidences of new patients with duodenal or gastric ulcer or oesophagitis, and the proportion of patients with a prescription of an H2-antagonist or proton pump inhibitor, were calculated for the period 1994–2003.

Results: Between 1994 and 2003, the age-standardized incidence of patients with duodenal ulcer decreased from 2.22% (95% CI 1.94–2.50) to 0.85% (95% CI 0.70–1.00), and the incidence of gastric ulcer from 1.75% (95% CI 1.50–2.00) to 1.01% (95% CI 0. 84–1.18). The incidence of oesophagitis increased from 7.20% (95% CI 6.70–7.70) to 8.73% (95% CI 8.23–9.23). The propor tion of patients with a prescription of a proton pump inhibitor increased from 8.68% to 50.02%, and those with a prescription of an H2-antagonist remained stable at around 20%.

Conclusions: In 10 years, the incidence of peptic ulcers decreased considerably, while the incidence of oesophagitis increased. Different trends in society are probably responsible for this, such as a shift in diagnosis and the increased use of proton pump inhibitors.

Points for discussion: Are there similar trends in other countries?

Presentation 4: Theme paper, ongoing study with preliminary results

Friday, 8 May 2009, 11.30–12.00 h.

The German CONTENT database: Advanced options for multimorbidity analyses

Gunter Laux, T Kuehlein, A. Gutscher, J. Szecsenyi

General Practice and Health Services Research, University Hospital Heidelberg, Voss-Strasse 2, 69115 Heidelberg, Germany. Tel: + 49 6221 566207; Fax: + 49 6221 561972; E-mail: [email protected]

Background: CONTENT (Continuous Morbidity Registration Epidemiologic Network) is an ambitious project in Germany that aims to establish a system for adequate EMR keeping and analysis in primary care based on episodes of care. An episode is defined as a health problem from its first presentation by a patient to a doctor until the completion of the last encounter for it. The CONTENT database currently contains more than 90,000 patients and more than 1,000,000 encounters between doctor and patient.

Research question: The presented study aims to describe patterns of healthcare utilization based on patients’ multimorbidity.

Methods: The analyses were based on a total of 39,699 patients in a yearly contact group for which data entry based on episodes of care was performed using the International Classification of Primary Care (ICPC). In order to model the relationship between explanatory variables (age, gender, multi-morbidity) and response variables of interest (prescriptions, referrals, encounters) that were applied to measure healthcare utilization, we used multiple linear regression.

Results: Multimorbidity measured by the number of patients’ chronic conditions had a manifestly stronger impact on the number of encounters (beta=0.51, p<0.0001) in comparison to age or gender. Moreover, we observed that the number of patients’ chronic conditions had a significant impact on the number of different prescriptions (beta= 0.226, p<0.0001) as well as on the number of referrals (beta=0.3, p<0.0001).

Conclusions: Documentation in primary care on the basis of episodes of care facilitates an insight to concurrently existing health problems and related medical procedures. Therefore, the resulting data provide a basis to obtain co- and multimorbidity patterns and corresponding healthcare utilization issues in order to understand the particularly complex needs caused by multimorbidity.

Points for discussion: What techniques could be useful to minimize information bias in EMR keeping (i.e., entered data should be complete and consistent)?

Presentation 5: Theme paper, ongoing study with preliminary results

Friday, 8 May 2009, 12.00–12.30

Combined predictive model to improve general practitioners’ performance

Pasquale Falasca, A. Berardo, F. Di Tommaso, A. Accorsi

Epidemiologia Ricerca & Sviluppo, Azienda Usl di Ravenna, Via De Gasperi 8, 48100 Ravenna, Italy. Tel: [phone]; Fax: [phone]; E-mail: [email protected]

Background: The King's Fund and the Department of Health (NHS) has recently provided a predictive statistical model to evaluate the risk of urgent hospitalization in order to enable appropriate preventive measures for chronic diseases and to help to improve general practitioners’ (GP) performance.

Research question: Does implementation of strategies for management and control of chronic diseases cause a significant reduction in the risk of hospitalization?

Method: A data set, which includes information collated from different databases (inpatient, out patient, emergency care [EC], GPs, social services) is used to identify patients at a high risk of hospitalization. A combined predictive model has been developed with a multivariate logistic regression (stepwise) and applied to the population aged >65 years in the province of Ravenna (Italy). The chronic diseases involved are: asthma, coronary heart disease, heart failure, chronic obstructive pulmonary disease, depression, diabetes, hypertension, and cancer. Patients are divided into three groups according to their risk level: high risk (immediate charge), moderate risk (ordinary management), and minimum risk (interventions for prevention and self-care).

Results: 82% of all subjects>65 years (73,055) have average risk per year of 5.9%, 8% (6,710) have a risk of 16.4%, and 4% (4,411) have an average risk rating of 37.7%. Sixty-three per cent of patients>65 years with at least one chronic disease had urgent hospitalization. This percentage rises to 88% for the high-risk group. High-risk patients are admitted to hospital almost 10 times more often than low-risk patients; they also have 5.4 times more access to EC and 2.4 times more access to their GP.

Conclusions: This model has been proved to be a useful tool to predict hospitalization. It can also help GPs to stratify patients according to their risk and enable them to allocate appropriate resources and attention to the higher-risk groups.

Points for discussion: Combined predictive model; chronic diseases; risk of hospitalization.

Presentation 6: Freestanding paper

Friday, 8 May 2009, 14.00–14.30

Patient participation in decision making in primary care settings: What general practitioners do and what patients perceive

Roger Ruiz-Moral, Lucía Peralta Mungía, Luis Pérula, Mariana Martínez, Jose Ramón, Loayssa Lara, M. Teresa Carrion

Family Medicine Teaching Unit, Andalusia Health Service, Blanco Soler 4, 14004 Cordoba, Spain. Tel: + 34 9570 12544; E-mail: [email protected]

Background: Patient involvement in decision making (PIDM) is widely recommended and promoted, but it seems that general practitioners (GPs) scarcely foster these behaviours in consultations.

Research questions: To what extent do GPs involve their patients in PIDM in consultations? What is the patient's perception about her/his involvement in such decisions?

Method: We video-recorded 648 patients attending 95 family doctors for a new health problem (average 6 patients/doctor). A single trained observer rated all the encounters with the CICAA-D (17-items scale with a range of responses across 3 points, previously validated). Intra-rater reliability was assessed in a 25-interview sample by means of the test-retest (Cohen's kappa and global intraclass correlation coefficient). Patient perception about their participation was registered by three questions just after the encounter.

Results: We analyzed 637 encounters: in 386 (60%), no PIDM was detected. In 250 (39%) interviews, some kind of PIDM was observed. In terms of the latter, the patient involvement in 160 (64%) encounters was labelled as “participated decision making”, and in the remaining 90 (36%) as “shared decision making”. There were significant differences (p=0.000) between the global CICAA-D scores of those interviews considered as non-participative (mean 6.9) and those defined as “participative” (mean 9.88) and “shared” (mean 12.30). Fifty-nine per cent of patients would like the doctor to ask him/her about his/her idea regarding the proposed treatmment/s.

Conclusions: In a high rate of encounters, GPs do not facilitate patient participation. In practice, we observed that PIDM can occur at different levels. A doctor's communication quality is higher depending on the degree of participation. These results support our previous concept of PIDM and the value of the CICAA-D as a tool for measuring PIDM. A patient's desire to participate seems to increase when the doctor gives the opportunity to participate, and some patients perceive being involved even when no specific behaviour from the doctor is detected.

Points for discussion: The concept of “involve patients in decision making” in primary care; tools for measuring this phenomenon; correlation be tween external evaluation and patient appraisals.

Presentation 7: Freestanding paper

Friday, 8 May 2009, 14.30–15.00

The European EUCCLID pilot study on care and complications in people with type 2 diabetes in primary care

Kees Gorter, Johan Wens, Xavier Cos Claramunt, Pinar Topsever, Kamlesh Khunti, Anne Karen enum, Nikolay Khalangot, Margalit Goldfracht, Imre Rurik, Christophe Berkhout, Christos Lionis Guy Rutten

Julius Center for Health Sciences and Primary Care, Department of General Practice, University Medical Center Utrecht, PO Box 85500, HP Str 6.131, 3508 GA Utrecht, the Netherlands. Tel: [phone]; Fax: + 31 88 756 8099; E-mail: [email protected]

Background: Studies on quality of primary diabetes care are scarce and prone to bias due to selection by indication.

Research question: This study aims to test the feasibility of the set-up and logistics of a proposed cross-sectional EUCCLID main study in randomly selected primary care type 2 diabetes (DM2) patients in 11 European countries.

Method: In total, 22 general practices (one rural and one urban per country) were asked to participate. From each practice, five patients were randomly selected from a list of all DM2 patients known to the participating general practitioners (GPs) for whom the GP is the main diabetes care provider. Medical history, anthropometric and biolab data, and EuroQOL VAS scores of all participants were sent to a central laboratory and entered into a central database.

Outcomes: Feasibility of logistics, central laboratory and data flow, prevalence of indicators of metabolic control.

Results: In all, 103 patients were included from 22 GPs in 11 countries. A total of 1184 DM2 patients participated from the rural practices (19 [Israel] to 524 [UK]) and 1079 DM2 patients from the urban practices (nine [Israel] to 338 [UK]). After random selection, biomaterials and questionnaires were obtained from 52 patients from urban and 51 patients from rural practices. Patient characteristics: 55% female, mean age 66 years (SD 10.6). Biomaterials: mean body-mass index (BMI) 29.2 kg/m² (SD 5.3), ranging from 26 (Norway) to 34.5 (Spain); HbA_1c 7.1% (SD 1.3), ranging from 6.6 (UK and Hungary) to 8.0 (Ukraine); systolic blood pressure 134 mm Hg (SD 16.3), ranging from 126 (Hungary) to 144 (Ukraine); and total cholesterol 4.9 mmol/l (SD 1.2), ranging from 4.1 (UK) to 6.2 (Ukraine). Questionnaires: EuroQOL VAS mean 70 (SD 21.5), ranging from 46 (Ukraine) to 81 (Norway).

Conclusions: A European study on quality of care in a random selection of DM2 patients is feasible. There are large differences between countries.

Points for discussion: 1) To assess quality of diabetes care, data on patients’ adherence and on GPs’ clinical inertia and barriers to activity are needed. Suggestions on how to incorporate these in an European study. 2) Suggestions for questionnaires to include in the main study. 3) Suggestions for funding.

Presentation 8: Freestanding paper

Friday, 8 May 2009, 15.00–15.30

Are abstracts presented at EGPRN meetings followed by publication?

Paul Van Royen, Hagen Sandholzer, Frances Griffiths, Christos Lionis, Jan-Joost Rethans, Ferran Gali, Sophia Eilat, Eva Hummers-Pradier

Department of Family Medicine/General Practice, University of Antwerp, Universiteitsplein 1, 2000 Antwerp, Belgium. Tel: + 32 3 820 2699; Fax: [phone]; E-mail: [email protected]

Background: Research presented to a scientific meeting is often inaccessible to clinicians, unless it is also published in a journal.

Research question: To assess the publication rate of studies presented at 10 European General Practice Research Network (EGPRN) meetings, from two time periods (1999–2002 and 2005–2006).

Method: All authors were contacted by postal questionnaire or by e-mail. Presentations were rated as published if a copy of the (pre) publication or a letter confirming acceptance by a journal was included, or when the publication was found by a Medline search.

Results: Information was obtained on 249 presentations. In total, 111 (45%) presentations were published. Papers were most frequently published in the British Journal of General Practice, the Scandinavian Journal of Primary Health Care, and Family Practice. However, only 63% of the evidence was published in English-language Medline-listed journals, increasingly also in non-general practice journals. Discussion at EGPRN meeting was rated as helpful by 60% of the presenters who published their results. The major reason for non-publication was that the projects presented were not complete. This reflects the EGPRN philosophy to act as a discussion forum for research ideas or ongoing research at an early stage. The most frequent cause for non-publication was failing to submit the paper at the time of follow-up (105 responses).

Conclusions: At EGPRN meetings, research ideas or ongoing research are often presented and discussed. In light of this, the ratio of published work to presented work compares well with the results of international meetings of other specialities. The discussion is valued as very helpful for publication. Presenters, presumably those who are not native English speakers, need help in submitting and publishing their work in international Medline-listed journals.

Points for discussion: Do you see other reasons for non-publication of papers presented at EGPRN? Do we need a publication strategy for our EGPRN network in order to increase publication rates? How could the results of this survey help us in prioritizing future activities?

Presentation 9: Freestanding paper, ongoing study, no results yet

Friday, 8 May 2009, 14.00–14.30

Relative effectiveness of GINA-3 and -4 asthma medications in 6–18-year-old children: A network meta-analysis

L.B. v.d. Mark, Edo Lyklema

Department of General Practice, AMC – UvA, Meibergdreef 15, 1105 AZ Amsterdam, the Netherlands. Tel: + 31 6172 42670; E-mail: [email protected]

Background: Several treatment options are avail able for treatment of asthma in children. Current guidelines are based on some randomized controlled trials performed in children, but mainly on scientific evidence found in adults or based on consensus. For asthma, drug treatments are compared to placebo, not against each other, although such evidence is what a physician needs in practice. In the absence of a collection of large, high-quality, randomized trials comparing all eligible treatments, indirect comparisons of multiple treatments are a way to rank order treatments as to their effectiveness. Using direct and indirect comparisons, a “network structure” can be composed.

Research question: How do the different drug treatments’ effectiveness compare for children 6–18 years old diagnosed with moderate persistent asthma in steps 3 and 4 of the GINA guideline?

Method: A literature search was performed to identify randomized controlled trials of all direct and indirect comparisons of treatment options according to the GINA guidelines steps 3 and 4. Two independent reviewers assessed all references and extracted data from the included references. We performed a comprehensive “network” metaanalysis of treatment options for asthma in children for which randomized trials have been published and will rank the relative effectiveness of all options. Using logistic regression, the relative effectiveness of each treatment can be calculated. Bayesian methods will be explored as well.

Results: Thus far, more than 7500 articles have been screened, and of these 19 have been included. At this moment, we are analysing the data obtained and expect to be able to present the results at the conference.

Conclusion: A clear overview of the effectiveness of the treatment options for children 6–18 years old diagnosed with moderate persistent asthma.

Presentation 10: Freestanding paper

Friday, 8 May 2009, 14.30–15.00

The effect of different operational definitions of asthma on clinical prediction rules of child hood asthma

Karina van Wonderen, Lonneke van der Mark, Patrick Bindels, Wim van Aalderen, Gerben ter Riet

General Practice, Academic Medical Center Amsterdam, Meibergdreef 15, 1105 AZ Amsterdam, the Netherlands. Tel: [phone]; E-mail: l.b.vandermark@ amc.uva.nl

Background: Clinical prediction rules, by provid ingillness probabilities, may be used to help physicians choose management options. It is unclear how much asthma prediction rules may be affected by variation in the (operational) definition of asthma.

Research question: What is the impact of different asthma definitions on the proportion of children for which a physician remains in doubt with respect to management (“area of clinical indecision”) after application of a prediction rule?

Method: We performed a Medline search (from 2000 to 2008) for cohort studies on childhood asthma. We extracted data on the (operational) definitions of asthma used therein. Next, we constructed logistic regression models for six different definitions, using wheezing, eczema, and specific IgE as predictors. To illustrate the potential impact of using different definitions on clinical decision-making, we defined an area of clinical indecision (posterior probability between 25% and 60%) and calculated the number of children who remained inside this area after application of the prediction rule. Proportions, differences between proportions, and their 95% confidence intervals were calculated using the bootstrap.

Results: The search yielded 1238 references, of which 142 were included. These 142 references yielded 53 different definitions of childhood asthma. Fifty-seven per cent of definitions were based on a doctor's diagnosis (without time constraint) with or without other symptoms or medication use. In 11% of the papers, bronchial hyperresponsiveness or spirometry was a component of the definition. In 11% of the references, asthma was defined using asthma symptoms alone. The percentage of children whose posterior asthma probability was in the area of clinical indecision varied from 11.5% to 76.4%, depending on the operational definition.

Conclusions: The performance of a prediction rule may be highly dependent on the operational definition of asthma used to build it. Physicians should pay close attention to how asthma was defined when using a prediction rule.

Points for discussion: Crude prevalence rates may strike high; we used clinical data from an ongoing prospective cohort study for children with high risk for developing asthma.

Presentation 11: Freestanding paper, study proposal/idea

Friday, 8 May 2009, 15.00–15.30

The SpiFP (Spirometry in Family Practice) study: Study design

S. Cadier, P. Barraine, C. Gut-Gobert, C. Leroyer, Jean Yves le Reste

Dept. de Médecine Générale Brest, Université de Bretagne Occidentale, Rue Camille Desmoulins, 29200 Brest, France. Tel: + 33 6 74 35 27 89; Fax: [phone]; E-mail: [email protected]

Background: Smoking cessation is the only intervention that can modify the natural history of chronic obstructive pulmonary disease (COPD), a leading cause of mortality according to the World Health Organization. Smoking cessation, using the “minimal intervention” strategy, is applicable and relevant in a primary care setting. Whether spirometry testing, as a biomedical risk assessment tool, helps to increase smoking cessation rate remains controversial.

Research question: Is spirometry, associated with the “minimal intervention” strategy, more effective on smoking cessation rate than the “minimal intervention” strategy alone?

Method: The setting will be in family practice, with recruitment of 120 general practitioners (GPs) in charge of trainees in seven French universities. We will train all GPs for the “minimal intervention” strategy. GPs will then be randomized into two groups: “minimal intervention” strategy alone versus “minimal intervention” strategy and spirometry, with collective training in the case of spirometry use. Each GP will aim at including all smokers older than 18 years over a 1-year period. The end point will be at 12 months of the intervention. Each patient will be asked for smoking status, and a carbon monoxide test will be performed. Individual patients’ data will be computerized in one central database, using an eCRF.

Estimation of the study sample: Taking the hypothesis of at least a quarter more smoking cessation rate at 12 months in the spirometry group, the inclusion of 5000 patients in each group is needed (unilateral testing with an α error set at 5% and a β error set at 20%).

Points for discussion: All comments are welcome (no results and no conclusion, as this is a study design).

Presentation 12: Freestanding paper

Friday, 8 May 2009, 16.00–16.30

Musculoskeletal problems in overweight and obese children

Marjolein Krul, J. C. van der Wouden, F. G. Schellevis, L. W A. van Suijlekom-Smit, B. W Koes

Department of General Practice, Erasmus MC, PO Box 2040, 3000 CA Rotterdam, the Netherlands. Tel: [phone]; E-mail: [email protected]

Background: The obesity epidemic in children is spreading at alarming rates. Musculoskeletal problems can influence physical activity.

Research questions: Do overweight and obese children report more musculoskeletal problems in daily life than their normal-weight peers? Do overweight and obese children seek help for musculoskeletal problems more often than their normal-weight peers?

Methods: We performed a cross-sectional database and face-to-face interview study that included 2459 children aged 2–17 years from Dutch general practice. We collected self-reported height and weight (BMI), self-reported musculoskeletal problems in the 2 weeks prior to the interview, general practitioner (GP) consultations for musculoskeletal problems in 1 year, and age (two age groups were analysed, 2–11 years and 12–17 years, because of the proxy interview in the youngest age group). We calculated odds ratios (OR) and 95% confidence intervals (CI) for musculoskeletal problems in over weight and obese children, compared to normal-weight children.

Results: Overweight and obese children in both age groups (2–11 years and 12–17 years) reported significantly more musculoskeletal problems (OR [95% CI] 1.86 [1.18–2.93] and 1.69 [1.08–2.65]), than normal-weight children. The total group of children with overweight and obesity reported more lower-extremity problems, than the normal-weight children (OR 1.62, 95% CI 1.09–2.41), and they reported more ankle and foot problems than children with normal weight (OR 1.92, 95% CI 1.15–3.20). Overweight and obese children aged 12–17 years consulted the GP more often with lower-extremity problems than the normal-weight children (OR 1.92, 95% CI 1.05–3.51).

Conclusions: This study shows that overweight and obese children more frequently experience musculoskeletal problems, than normal-weight children. Children with overweight and obesity aged 12–17 years more frequently seek medical help for their lower-extremity problems than their normal-weight peers.

Points for discussion: 1) The vicious circle of being obese, having musculoskeletal problems, and physical inactivity: where will it most often start? What is the best place for intervention? 2) Do we need more research to gain more insight in this vicious circle?

Presentation 13: Freestanding paper

Friday, 8 May 2009, 16.30–17.00

General practice-based intervention for suspecting and detecting dementia in France: A cluster randomized controlled trial

Virginie Rondeau, Denis Pouchain, Jean-Francois Dartigues, et al., for the IMAGINE study group

General Practice, Paris Ile de France Ouest, 6, rue du dosteur Lebel, 94300 Vincennes, France. Tel: + 33 16259 14951; E-mail: [email protected]

Aims: To evaluate the impact of a general practice-based intervention on 3021 patients aged >75 years with spontaneous memory complaints, and to identify evocative signs of Alzheimer disease through a cluster randomized controlled trial.

Research question: What is the impact of an intervention on general practitioners’ (GP) capacity to identify Alzheimer disease in patients with memory complaints?

Method: 681 GPs, randomly selected and assigned to an intervention (n=352) or a control group (n = 329), and 214 specialists to confirm diagnosis. The intervention group participated in training sessions relating to dementia diagnosis and the use of brief neuropsychological tests. The control group dispensed care. The primary outcome was suspicion of dementia by GPs. The secondary outcome was accurate detection of dementia by the GPs.

Results: Most patients (96%) were already followed by the GPs for 10.8 years. Suspicion of dementia was twofold higher for GPs in the intervention group (adjusted OR= 1.99, p<0.0001). Probability of suspected dementia increased with patient age and decreased with educational level. However, the positive predictive value was not significantly different between the two groups (60.9% vs. 64.4%, p=0.41). GPs from the intervention group also had a higher probability (adjusted OR=2.24, p=0.01) of correctly detecting demented patients. The study showed that information on dementia and application of simple psychometric tests could improve the precision of a GP's diagnosis without changing the efficacy of detection of dementia.

Conclusions: Information on dementia and application of simple psychometric tests could improve the precision of diagnosis by GPs without changing the efficacy of the detection of dementia. Other methods should be tested in order to improve the detection rate, in particular intervention in patients with no cognitive complaints.

Points for discussion: How to improve GPs’ capacity to recognise Alzheimer disease in old patients with spontaneous memory complaint.

Presentation 14: Freestanding paper, ongoing study with preliminary results

Friday, 8 May 2009, 17.00–17.30

High prescription rate of ATD drugs in France: GPs’ motives

Alain Mercier, I. Aubin, L. Coblentz, M. Schuers, E. Lefebvre, E. Mauviard, P. Chazot, J. L. Hermil

Family Medicine, Rouen University, Faculty of Medi cine, Bd Gambetta, 76000 Rouen, France. Tel: [phone]; Fax: [phone]; E-mail: alain. [email protected]

Background: Antidepressant (ATD) drugs are very commonly prescribed in all industrialized countries. Consumption in France is higher than in other European countries. The reasons for this higher prescription rate are still unknown. In this study, we explore general practitioners’ (GP) attitudes on prescribing ATDs in France.

Research question: Understanding either contextual and/or diagnostic reasons to prescribe ATDs among French GPs.

Method: Qualitative data about reasons for ATD use were collected through GP focus groups. This sensitive topic was also addressed by in-depth interviews, to be sure not to miss any essential information, and to explore the points that remained unclear. The main points of the topic guide insisted first on social, family, or personal situations and patients’ requests. Secondly, it focused on possible “non-psychiatric” diagnoses. The analysis was made with a phenomenological approach and content analysis, using QSR Nvivo 8.0 software.

Results: This is an on going study. Preliminary results show that GPs’ prescriptions were mainly influenced by the social and professional conditions of their patients. They also used ATDs as a “diagnostic test” when dealing with unspecified symptoms, but only when they were associated with mental health problems. They commonly used ATDs for “non-psychiatric” diseases, such as migraine, colopathy, pain conditions, and sexual problems, but those conditions were always combined with mental health problems.

Conclusions: According to the GPs, some prescriptions seem to be justified in contrast to official guidelines.

Points for discussion: What do you know about recommendations and attitudes for ATD prescriptions in “non-psychiatric diagnoses” in other countries?

Presentation 15: One slide/five minutes, ongoing study, no results yet

Friday, 8 May 2009, 16.10–16.20

Management of acute chest pain in rural practice in Europe

EURIPA's Research Committee, Claudio Carosino

EURIPA, MEDICINA OGGI, Via Paganini 4, 43011 Busseto, Italy. Tel: + 39 05249 32812; Fax: + 39 05249 2626; E-mail: [email protected]

Background: General practitioners (GPs) often have to face the problem of acute chest pain. It is probably the most frequent symptom in emergency care, and certainly one of the most important, because it can disclose certain severe diseases such as acute coronary syndrome, pericarditis, or pulmonary embolism. In order to make a correct diagnosis and to provide appropriate therapy, the availability of certain technologies and the time needed for this are crucial. GPs working in rural and isolated areas probably have to overcome further problems and to find special ways for applying guidelines.

Research question: We would like to draw a picture of actual management of this problem in different rural settings across Europe, and to point out problems and barriers in following international guidelines.

Method: We choose the topic after a wide discussion on EURIPA's listserver and in a workshop during the Wonca Europe Conference in Paris 2007. We asked EURIPA's members to take part as volunteers in this project, and we gathered about 40 practices in different countries for the collection of clinical data. The protocol of the study is the result of further discussion on a special listserver for researchers and consists of a prospective longitudinal observational study. We developed a form in order to record each case of “new-onset (not due to a trauma) chest pain (for which we are usually asked for an unscheduled consultation, a home visit, or an after-hours call) when we have to deal with a differential diagnosis concerning severe heart or lung disease”. The form aims to collect details about the epidemiology of the problem (frequency, symptoms), the process of care (timing of intervention, decision procedures, diagnosis instrumental sup ports), and outcomes.

Points for discussion: Is the problem relevant? Is the methodology correct and suitable?

Presentation 16: One slide/five minutes, study proposal/idea

Friday, 8 May 2009, 16.20–16.30

Impact of virtual consultancy on the number of GP referrals to dermatology

Tiago Villanueva, Maria Raquel Vieira, Israel Macedo, Jorge Cardoso

Department of General Practice, USF AlphaMouro, Av. 5 de Outubro, 180, 1. esq., 1050–063 Lisbon, Portugal. Tel: [phone]; Fax: [phone]; E-mail: [email protected]

Background: In Portugal, dermatologists often criticize general practitioners (GPs) for their lack of dermatology knowledge and skills, as well as for writing poor referral letters. In many cases, GPs are criticized for referring patients who do not meet referral criteria. Virtual dermatology consultancy may help GPs decrease their referral rates to dermatology by allowing them to send pictures taken at point of care and to discuss clinical cases with experienced dermatologists via e-mail. Dermatologists then provide feedback, namely diagnostic or therapeutic guidance, which, in many cases, may allow the GP to manage cases adequately and thus avoid the need to refer the patient. Virtual dermatology consultancy can also function as a means of continuous professional development for GPs, and foster the integration of primary and secondary healthcare.

Research question: Does virtual consultancy in dermatology decrease the number of referrals from general practice to dermatology?

Method: This is a prospective cohort study (the study population comprises GP patients with dermatological lesions). For 1 year, a GP trainee will alternate in periods of 2 months between 1) exchange e-mails with a dermatology specialist based at the local Central Lisbon Hospital concerning dermatology cases of the practice and 2) making own decisions on how to manage patients without any e-mail advice. The GP trainee will send e-mails to the dermatologist, including a photo of the dermatologic lesion, as well as brief clinical information from the patient's history and physical examination. The dermatologist will then e-mail the GP trainee back with diagnostic and therapeutic guidance, as well as with advice as to whether to refer the patient. The GP trainee will then manage the case, or refer it, according to the dermatologist's advice.

Points for discussion: Is there a need for more widespread use of virtual consultancy in general practice? Use of virtual consultancy as a CPD tool for GPs; impact of virtual consultancy on integration of primary and secondary healthcare.

Presentation 17: One slide/five minutes, study proposal/idea

Friday, 8 May 2009, 16.30–16.40

Type 2 diabetes mellitus control in Mediterranean countries: A collaborative survey in the primary care setting. Second step: research protocol

Ferdinando Petrazzuoli, Xavier Cos, Frank Dobbs, Christos Lionis, Luc Martinez, Jean Karl Soler, Shlomo Vinker, Ksenija Kranjcevic, Pinar Topsever

SNAMID Caserta, Via Orientale 3, 81010 Ruviano (CE), Italy. Tel: + 39 08238 60032/[phone]; Fax: + 39 08238 60032; E-mail: [email protected]

Background: Issues of effectiveness regarding management of type 2 diabetes (T2D) in primary care have a central place in the literature. In the UK, an incentive scheme to improve quality of care delivered by general practitioners has recently been implemented with great success. This study aims to measure quality performance in T2D management across different Mediterranean countries (Italy, Spain, Greece, France, Malta, Israel, Croatia, Turkey) using some of the quality indicators for T2D established in the UK Quality and Outcome Framework (QOF).

Research question: What is the level of control of T2D among Mediterranean countries in the primary care setting?

Method: A retrospective cohort analysis of the medical records of patients with T2D having consulted in primary healthcare settings of different Mediterranean countries between 01/2007 and 12/2007. Records will be selected at random. Exclusion criteria: newly diagnosed patients; steroids or major surgery in the last 3 months; dementia; terminal illness; or extreme frailty. The variables (outcome and explanatory variables) to be evaluated will be: 1) sociodemographic parameters; 2) metabolic and clinical data: HbA_1c (as the main outcome indicator), urinary microalbumin, low-density lipoprotein (LDL) cholesterol, blood pressure, diabetic complications, and co-morbidities. To assess quality of management, we will use the following targets: HbA_1c <7%, SBP <130 mm Hg and/or DBP <80 mm Hg, microalbuminuria <30 mg/day, and LDL <100 mg/dl. The analysis will be performed at a patient level. A sample size will be calculated to detect a 10% difference in performance, with 95% confidence and 80% power. Since great variance of values and distributions for many quality indicatorsis expected, we will provide non-parametric boot strap techniques to compare arithmetic mean between countries.

Points for discussion: 1) Relevance of the study; 2) validity of the methodology.

Presentation 18: One slide/five minutes, study proposal/idea

Friday, 8 May 2009, 16.40–16.50

The improvement of quality of care in family practices using an electronic data system in Poland

Roman Topor-Madry, R. Kacorzyk, T Sobalski, Z. J. Król, A. Windak, W. Lukas, M. Milewska

Institute of Public Health, Jagiellonian University, Grzegorzecka 20, 30531 Krakow, Poland. Tel: [phone]; Fax: + 48 4260 366; E-mail: [email protected]

Background: Dedicated computer systems are not often used in family practices in Poland, even though the possible benefits are impressive, as shown by experiences in other countries. The drEryk is a computer software developed for the purpose of family practices and designed by family doctors to support the process of diagnosis, treatment, and monitoring patients using an electronic data system. Currently, it is used by near 400 practices in Poland.

Research question: The aim of the project is to assess the benefits of electronic documentation in family practices in Poland in improvement of quality of care, patient monitoring, and management.

Method: Participants in the project will be family practices using drEryk software. The practices agreeing to participate will create a network and share their experiences of using the system and the data. The data collection would consist of standardized questionnaires, which will be completed by family doctors periodically, and non-individual data coming from the practices. The published results will be aggregated and blinded. Each family doctor will have access to his/her individual data, which will be compared to data from the rest of the network. The main data would be quality indicators and data referring to management of the practice.

Expected results: The outcomes of the project are clinical and non-clinical. The main clinical outcomes are morbidity (with special aspects of multimorbidity in older age), use of drugs, use of procedures, basic screening procedures (blood pressure, blood glucose, and diabetes), and use of prevention advice. Non-clinical aspects include time of visit, waiting time, access to information, reporting for national statistics and insurance systems, management, etc.

Conclusions: The collaboration of family practices in the network is a very useful way for improving quality of care. Support from dedicated computer software would facilitate running single practices and give hard data for more fruitful network collaboration.

Points for discussion: 1) Electronic documentation in Poland (benefits of using and not using); 2) quality control and support based on an electronic computer system; 3) network collaboration and comparison of quality outcomes.

Presentation 19: One slide/five minutes, ongoing study, no results yet

Friday, 8 May 2009, 16.50–17.00

Project FIRE: Swiss general practice electronic database in an early stage

Bhend Heinz, Marco Zoller

General Practice and Health Services Research, University of Zurich, Rämistrasse 100, 8091 Zurich, Switzerland. Tel: + 41 (44) 255 98 55; E-mail: marco. [email protected]

Background: In countries like the UK, Germany, Australia, Malta, etc., electronic databases facilitate research in general practice settings. Ideally, data from practices using a primary care classification are uploaded directly from electronic medical re cords (EMRs) to a central server. Less than 13% of general practices are fully equipped with EMRs. The executive board of the Swiss Society of General Practitioners in cooperation with the Zurich Institute of General Practice started a project last year. The aim of the project is to build a research network of 100 general practices until 2010 using EMRs, and coding problem lists and reasons for encounter using ICPC-2.

Method: The project integrates the following steps. A pilot group of 15 general practitioners (GPs) attended courses on using ICPC and started coding in daily routine practice. A wiki website was set up to adapt a German and French ICPC, a thesaurus of terms used in Swiss general practice. A roadmap was published to standardize the architecture, formats, and important procedures in GPs’ EMRs. A mini mal data set was defined for the first stage. All companies offering practice information systems in Switzerland received detailed information about specifications for XML files and were asked to offer an export tool for the data set. The central server was set up to receive data from practices.

Results: In the second semester of 2009, we intend to collect data from pilot practices, including vital data, laboratory test results, drug prescriptions, and lists of problems and diagnoses. Analysis and benchmarking tools are under development.

Points for discussion: 1) What steps in completing the data set in further project stages have proven to be most useful? 2) Anonymized data without unique patient identifiers allow only cross-sectional data analysis. What do we have to take into consideration in order to prepare for the future?

Presentation 20: One slide/five minutes, ongoing study with preliminary results

Friday, 8 May 20 09, 17.00–17.10

Is the activity of a GP in an Emmaus community comparable to the usual activity of a GP working in an urban district?

Christophe Berkhout, Stéphane Cogez, Gilles Hebbrecht, Didier Duhot

Department of General Medicine, University of Lille 2–Law and Healthcare, Faculté de Médecine Henri Warembourg, Pôle Formation, 59045 Lille Cedex, France. Tel: + 33 607 307351; E-mail: [email protected]

Background: Emmaus communities are places were precariously living people can recover if they participate by their work in the communities’ shared costs and enter a reintegration programme. A weekly general practice intervention is carried out at the Dunkirk community's site.

Research question: Is the activity of a GP working in an Emmaus community comparable to the usual activity of a GP working in an urban district?

Method: Retrospective epidemiological survey of 5 years of registered data of ongoing encounters extracted from the electronic records and compared with data from the OMG database. Outcomes were the “results of consultation”. Analysis was at first descriptive, and differences in the activity of Emmaus community-based GPs versus “standard urban-based GPs” were tested.

Results: The data from the Dunkirk Emmaus community totalled 263 patients, 2075 encounters, and 2695 results of consultation. Analysis is on going. Preliminary results should be ready for presentation in Bertinoro. Results might reveal whether a GP needs special training in dealing with precariously living people.

Points for discussion: 1) Do you know other experiences of electronic data recording for the specific care of precariously living people with standard GP recording tools? 2) Can precariously living people reach the same quality of primary healthcare as their non-precarious counterparts?

Presentation 21: Theme paper

Saturday, 9 May 2009, 09.30–10.00

Leonardo Project: A patient empowerment approach based on the chronic care model in the setting of general practice. The role of specific web-based software

Ernesto Mola, R. Rizzo. A. Aquilino, A. Musilli, M. Panfilo, R. Rollo

Sezione di Medicina di Famiglia e di Comunità, ASSIMEFAC, via degli Aranci 2, 73100 Lecce, Italy. Tel: + 39 0832 642014; Fax: [phone]; E-mail: [email protected]

Background: Project Leonardo, a feasibility study promoted by a public-private partnership between the Puglia Regional Health Authority (Italy) and Pfizer Italia, was a disease management programme aimed to show that a team-based approach including care managers (CMs)—general practitioners (GPs) and specialists as “partners” of the patient with chronic conditions, promoting an empowerment approach—can improve adherence to treatment and follow-up, clinical outcome, and patient and health professional satisfaction. The whole programme was supported by web-based software, called Informacare.

Research question: What was the role of Informacare in collecting data? Did the software give useful decisional support to CMs? What were the benefits of using web-based software?

Method: 30 nurses/CMs in 20 family medicine group including 83 GPs were appropriately trained to support and coach 1160 patients in self-management of their chronic conditions and doctors in drafting an individualized care plan based on guidelines and needs, expectations, and the social and economic situation of patients.

Results: Patients enrolled in the Leonardo Project improved adherence to healthy lifestyles and to medications as well as clinical outcomes and satisfaction. The web-based tool allowed the management team to have real-time data organized in weekly reports, facilitating the project monitoring, focusing on attained results and critical areas that needed additional training. Informacare supported CMs and doctors in defining individualized care plans for each patient. The web-based solution also guaranteed data safety.

Conclusions: Project Leonardo demonstrates the feasibility of a patient-centred, team-based model of care, including well-trained CMs in general practices to manage patients with chronic conditions. The web-based tool Informacare appeared essential in the collection of data, the monitoring of step-by-step programme activities, and as decisional support. General practice databases could benefit from similar software in managing chronic patients.

Points for discussion: 1) Usefulness of EMR in managing chronic patients; 2) usefulness of shared software on the web; 3) EMR as decisional support.

Presentation 22: Theme paper, ongoing study with preliminary results

Saturday, 9 May 2009, 10.00–10.30

Do population characteristics make a difference?

Catharina van den Dungen, N. Hoeymans, F. Schellevis, G. Westert

Centre for Public Health Forecasting, National Institute for Public Health and the Environment, PO Box 1, 3720 BA Bilthoven, the Netherlands. Tel: [phone]/+ 31 1 34668 338; Fax: [phone]; E-mail: karin. [email protected]

Background: In the Netherlands, morbidity rates derived from general practice registration networks (GPRNs) are used as a public health indicator. These networks continuously collect anonymous information on patients about problems presented in general practice. In the Netherlands, many GPRNs exist, but morbidity rates differ.

Research question: What is the influence of adjusting for differences in socio-economic status (SES), degree of urbanization, and non-western population among Dutch GPRNs on morbidity estimations of diabetes, asthma, coronary heart disease (CHD), and herpes zoster?

Method: We used the recorded information of several Dutch GPRNs on the incidence and pre valence of their populations in 2007. The distribution of gender, age, SES, the degree of urbanization, and non-western population is known in each GPRN. We calculated the incidence and prevalence rates of diabetes, asthma, CHD, and herpes zoster. We compare the morbidity rates adjusted for age and gender with the rates adjusted for SES, degree of urbanization, or percentage of non-western population.

Results: Preliminary results: After adjustment for SES, ethnicity, and degree of urbanization, the variation in morbidity rates among Dutch GPRNs declined. In May, we want to present the final results from the study.

Conclusions: The tentative conclusion on the basis of the preliminary analysis is that adjusting for SES, degree of urbanization, and ethnicity has a positive influence on reliability, because the variation between them reduced. In further research, we recommend discounting these population characteristics.

Points for discussion: 1) What different population characteristics are important to consider? 2) What factors, in addition to population characteristics, can influence morbidity rates or improve the use of morbidity estimations?

Presentation 23: Theme paper

Saturday, 9 May 2009, 10.30–11.00

CONTENT EPR as a tool for guideline implementation

Thomas Kuehlein, G. Laux, A. Gutscher, J. Szecsenyi, S. Joos

General Practice and Health Services Research, University Hospital Heidelberg, Vossstr. 2, Geb. 37, D-69115 Heidelberg, Germany. Tel: + 49 (0)6221 564818; Fax: [phone]; E-mail: thomas.kuehlein@ med.uni-heidelberg.de

Background: Data from the CONTENT database (Continuous Morbidity Registration Network) revealed that, in spite of recommendations for treatment of uncomplicated lower urinary tract infections (uLUTI), trimethoprim (TMP) was prescribed in only 12% of cases (quinolones 60%). Doctors justified their behaviour with perception of a high rate of therapy failure.

Research question: Internal evidence and experience of the GPs seem to outweigh external evidence. Does a proper trial and control of success rate with the CONTENT EPR effect change?

Method: Mixed three-step study design. In a baseline focus group, opinions concerning guideline recommendations were assessed. In a 3-month trial period, participants prescribed TMP (2X150 mg/day for 3 days) to their uLUTI patients, documented in episodes of care in the CONTENT EPR. Therapy failure was defined as necessity to prescribe a second antibiotic or referral to specialists. In a second focus group, GPs were confronted with the results, and the effect on their opinion was explored.

Results: Therapeutic decisions were mainly driven by former hospital training and what was perceived as common therapy in letters after hospital discharge and from specialists. GPs felt no need to change a successful and common regimen. TMP had a success rate of 94% (84 episodes of uLUTI). The proper trial strongly changed GP opinion towards the use of TMP. Personal and group experiences were still named as having the strongest influence on decision (nine GPs in two focus groups). Statistical data analysis was seen as an important means of confirmation. Documentation in the CONTENT EPR was judged as easily feasible.

Conclusions: Internal evidence, experience, and peer-group opinion seem to be the strongest determinants for clinical decisions. Ownership of data acquisition and analysis gives strong support in the process of change and is easily feasible with an adequate EPR.

Points for discussion: Data quality in research databases supplied by EPRs is dependent on the perceived usefulness and meaningfulness of data acquisition on the side of the participating GPs. How can we further enhance GPs’ interest in working with their own data and improve confidence in external evidence to facilitate and promote our research and guideline implementation?

Presentation 24: Poster, ongoing study with preliminary results

Saturday, 9 May 2009, 11.30–13.00

Use of computer and electronic records among primary care doctors in Turkey

Bengu Pala, Murat Unalacak, Fatih Yuksel, Ilhami Unluoglu

Family Medicine, Eskisehir Osmangazi Universitesi Tip Fakultesi Aile, Hekimligi AD, 26480 Eskisehir, Turkey. Tel: + 90 5069 224080; E-mail: [email protected]

Background: Computer and electronic medical records are thought to help primary care physicians in their daily work and in making clinical decisions. We aimed to gather the thoughts of primary care physicians and learn how they use computers for clinical and scientific purposes.

Research question: What are the thoughts of primary care physicians about the use of computers and electronic medical records (EMR) for clinical and scientific purposes, and how do they behave in their clinical practice?

Method: We administered a questionnaire, which included questions aimed at revealing doctors’ behaviours regarding the use of computers and EMRs, to 86 primary care doctors who were selected randomly among participants of a national primary care conference. All analyses were performed with SPSS for Windows, version 13.0.

Results: Most of the 86 respondents (80.2%) had computers in their offices, and 87.2% had computers in their clinics available for common use. Of those, 89.5% used an electronic medical record system at the office. With regard to using an EMR system, 67.4% of respondents defined it as increasing their daily burden, but 65.1% pointed out an increase in the quality of their medical care; only 2.3% thought that using an EMR system is unnecessary. Of all respondents, 18.6% thought that using an EMR system decreased their daily burden. Of the respondents, 55.8% rated the EMR system that they used to be poor with respect to privacy of patient information, and 50% rated their system as being poor at saving patient data. On the other hand, a considerable number of respondents (61.6%) thought that using an EMR system was worth the time and effort in using it.

Conclusions: Most of the physicians agree that inclusion of computers and EMRs in daily clinical practice is useful, providing that the EMR system can be improved.

Points for discussion: 1) Can it be helpful to standardize the electronic medical recording systems among all primary care givers in order that it can also be helpful to standardize the further in service training?

Presentation 25: Poster

Saturday, 9 May 2009, 11.30–13.00

SumEHR in Belgium: Which practices have them, and how are clinicians using them?

Peter Burggraeve, K. Thomeer, L. Pas

Flemish Scientific Institute of General Practitioners, Domus Medica, Sint-Hubertusstraat 58, 2600 Berchem (Antwerp), Belgium. Tel: + 32 478 44 88 32; Fax: [phone]; E-mail: [email protected]

Background: Continuity of primary care requires availability of patient information at all times. In Belgium, a new standard for communication between practices was introduced in 2007, the “Summarized Electronic Health Record”, containing core health information of patients. It aims at the availability of data during out-of-hours services and further facilitates the exchange of health information.

Research question: How to develop quality exchange and data collection in primary care.

Method: Data were collected from a small convenience sample of practices participating in development of standards for electronic medical recordkeeping. A dummy patient was created and sent to 20 practices, representing 10 Flemish formally accredited software packages. A facilitator visited the practice to check how the patient was filed in the EMR. The XML (SumEHR) export was collected for further analysis. Results were discussed at a meeting with participants.

Results: The XML syntax (export message) was not valid for eight out of 10 software packages. Semantics or content showed a diversity in SumEHR export. Main reasons were coding behaviour of the general practitioner (GP) and presence (or absence) of links between the different levels of the EPR. A better outcome was found among practices using episode registration.

Conclusions: Technical problems make it impossible for different practices to communicate or ex change patient data in a secured way. To improve outcome and utilization, content of SumEHR needed to be better defined and application harmonized between software packages. More standardized data entry should be promoted.

Points for discussion: 1) How can we enhance use of more standardized input of patient data in the EMR? 2) Is development of a European standard for summarized records feasible? 3) Can health research information system content be improved through data collection networks?

Presentation 26: Poster

Saturday, 9 May 2009, 11.30–13.00

Analysing family physicians’ financial data in the Estonian Health Insurance Fund's Database for Family Physicians’ Bonus Payment System: Results after 3 years’ experience in Estonia

Katrin Västra, J. Viilup, H. Tarien

Health Care, Estonian Health Insurance Fund, Lembitu 10, 74001 Tallinn, Estonia. Tel: + 372 6208 315; Fax: + 372 6208 449; E-mail: [email protected]

Background: To promote family physicians (FPs) to be actively involved with disease prevention, to prevent the spread of infectious diseases, and to ensure more effective monitoring of ill persons in case of chronic diseases, a bonus payment (BP) system was implemented in 2006.

Research question: Can we use financial data sent by FPs to analyse quality of care, and how was the additional bonus payment applied with respect to improving quality of care?

Method: We have used the Estonian Health Insurance Fund's (EHIF) financial database as well as the activities reported by family physicians. The number of voluntarily participating FPs in the voluntary quality bonus system was 63% in 2006, 57% in 2007, and 78% in 2008. The quantitative analysis includes the target groups’ coverage and interventions provided, where family physicians involved in the bonus pay scheme and others are compared.

Results: The results after 3 years of implementation are encouraging. All FPs sent their financial data via electronic channels (submitted bills) to EHIF throughout all 3 years. FPs who participated in the programme performed more activities as compared to FPs who did not participate in the quality bonus programme (such as family nurse individual consultations, infant vaccination, management of diabetes, and essential hypertension). The indirect positive impact of the quality bonus system was the improved results of all FPs (participants as well not non-participants) compared to the results of the first and second year of the system.

Conclusions: The FPs’ financial data collection is suitable for the bonus payment system and monitoring of the policy. Patients with chronic diseases are more involved in testing, and those primary care practitioners participating in the bonus pay system are more actively engaged in preventive care.

Presentation 27: Poster, ongoing study, no results yet

Saturday, 9 May 2009, 11.30–13.00

A new tool for diagnosis of chronic systemic diseases in primary care

Stephania Kokkali, Celine Droesch, Dominique Gras, Jean Sibilia

Faculte dé Médecine de Strasbourg, Université de Strasbourg, 6, rue St Pierre le Jeune, 67000 Strasbourg, France. Tel: + 33 06 33 55 50 25; Fax: [phone]; E-mail: [email protected]

Background: Due to the low prevalence of chronic systemic diseases (CSD) in primary care, there are only a few papers dealing with the topic. These suggest an underdiagnosis and delay in treatment for patients with rheumatic arthritis managed by general practitioners (GPs). We hy pothesized that GPs and trainees in family practice are not comfortable with the diagnosis of these rare but unavoidable diseases.

Research question: Would a new tool, helping to screen our patients for diagnosis of CSD, be helpful for the GP and perhaps other specialities?

Method: 1) Taking into account the prevalence of CSD, we are building a reminder including a listing of clinical symptoms with their positive and negative predictive values. This will lead the GP to make different hypotheses. This electronic tool will allow the GP to request the right screening test (X-rays, antibodies, hormones) when he/she is confronted with a suspicion of CSD. 2) This reminder will be validated by one or more specialists in CSD. 3) This validated tool will be distributed via the Internet to all physicians in the Alsace region willing to answer a questionnaire dealing with CSD before and after using the tool. 4) An adapted questionnaire like the one mentioned above will be given to all our postgraduate trainees in family practice.

Results: The questionnaire for the trainees will assess the need for medical education relevant to CSD at the pre- and postgraduate level. The questionnaire for physicians before and after 3 months of using the tool could give us an evaluation of their satisfaction in using the tool. The rate of downloading will give an idea of their needs.

Conclusions: The tool and the trainees’ expectations should be available for the meeting.

Points for discussion: How to build a relevant questionnaire to assess satisfaction with the tool.

Presentation 28: Poster

Saturday, 9 May 2009, 11.30–13.00

Quality of encoded data from GPs’ EPRs

Marc Verbeke, Diego Schrans, Sven Deroose

Department of Family Medicine & Primary Health Care, Ghent University, De Pintelaan 185, 9000 Gent, Belgium. Tel: + 32 933 23542; E-mail: marc.verbeke@ ugent.be

Background: The quality of data analysis from EPR in general practice depends on the consistency of the reported data. The best way to manage the data from the EPR is by using classifications adopted for primary care. Encoded data are more manage able than free text, and ICPC is a good tool for GPs’ reporting.

Research question: Starting up a registration network, we want to have a look at the quality of the encoding of the data by the GP.

Method: For this observational study, we selected 13 GPs working with the same software. The software vendor offered the only free training the GPs had. The system was selected because of the implementation of a Belgian thesaurus defining more than 48,000 clinical concepts that can be used as labels in the different parts of the journal. By using this thesaurus, all of the data are ICPC/ICD encoded in the background without a GP being a specialist in encoding. After 3 months, the GP retrieved all used codes in the EPR. The use of the codes was split up into all the different parts of the EPR journal (SOAP).

Results: This retrieval illustrates on what level GPs encode spontaneously: 92.7% at the diagnostic level and 58.2% for describing clinical interventions. The RFE was encoded by 50% of the GPs. RFE was poorly encoded in follow-up encounters. GPs avoided repeat reporting; automated reporting offers better results.

Conclusions: Our results demonstrate that encoding data by GPs using a good tool is feasible, but we suppose this will be better after training. Within our observation, we have no idea of the quality of the content of the encoding. We suppose that when GPs get added value by using encoded information, they can do it quiet well.

Points for discussion: 1) How to improve quality of encoded data; 2) the importance of record structure; 3) added value for GPs from encoded data.

Presentation 29: Poster

Saturday, 9 May 2009, 11.30–13.00

Osteoporosis in premenopausal and postmenopausal women receiving long-term L-thyroxine therapy

Filippo D’Addio, Ferdinando Petrazzuoli, Nicola Buono, Carmine Farinaro

Medicina Generale, SNAMID, via Napoli 22, 81024 Maddaloni, Italy. Tel: + 39 08234 36049; Fax: [phone]; E-mail: [email protected]

Background: The impact of long-term L-thyroxine therapy (LTT4) on bone loss in pre- and postmenopausal women is a significant issue be cause of the large amount of women receiving this drug treatment. Scientific published studies show the effect on bone loss and on growing osteoporo sis of LTT4 given at suppressive (high) doses, but studies performed on women with LTT4 given at replacement doses show a light effect on bone loss without an increase in clinical osteoporosis. The analysis of drug prescriptions in GP databases is very useful for assessing the link between LTT4 and clinical osteoporosis.

Research question: Is the prevalence of clinical osteoporosis in pre- and postmenopausal women treated with LTT4 higher than that in all pre- and postmenopausal women?

Method: From data of all drug prescriptions performed during the years 2006 and 2007 from a drug prescription database of 10 Italian GPs, a sample of drug prescriptions for women over 40 linked to osteoporosis diagnosis (code 733 of ICD-9 classification) was extracted. In these women, we searched for permanent prescription of L-thyroxine. We calculated the prevalence of osteoporosis in all the sample and in women of the sample who took L-thyroxine. The statistical difference was evaluated with the software STAT-CALC EPI-INFO.

Results: Of 3575 women over 40, 397 (11%) had drug prescriptions related to osteoporosis; 264 (7%) had LTT4, and 46 (17%) of these women on LTT4 had drug prescriptions related to osteoporosis. The relative risk in this population was 1.7, and it is statistically significant.

Conclusions: The data from the drug prescription database of 10 Italian GPs confirm a higher risk of clinical osteoporosis in pre- and postmenopausal women on normal LTT4.

Presentation 30: Poster

Saturday, 9 May 2009, 11.30–13.00

Dizziness increases the risk of fractures: A retrospective cohort study

Carsten Kruschinski, O. Sheehy, J. Le Lorier, E. Hummers-Pradier

Institute of General Practice, Hanover Medical School, Carl-Neuberg-Str. 1, 30625 Hannover, Germany. Tel: + 49 511 532 8159; Fax: [phone]; E-mail: [email protected]

Background: Dizziness is known to be associated with the risk of falls. However, there is little direct evidence for the increase of fractures caused by dizziness.

Research question: Do patients suffering from dizziness show an increased fracture rate?

Method: Retrospective cohort study using a population-based administrative database. A cohort of 2442 patients with at least one diagnosis of dizziness between 01/1998 and 06/2003 was compared with 16,125 age- and sex-matched control patients. The main outcome measure was any kind of first fracture after the index date of dizziness. In addition, we stratified by osteoporotic and non-osteoporotic fractures. Statistical analysis was performed using the Cox proportional hazard regression model, adjusting for possible confounders.

Results: Analysis revealed a significant effect of dizziness as an independent contributing factor to fractures (adjusted hazard ratio [HR] 1.26, 95% CI 1.03-1.55). Moreover, a fracture in the year prior to the index date was highly associated with the incidence of a subsequent fracture (adjusted HR 2.69, 95% CI 2.09–3.47) as was male gender (adjusted HR 1.30, 95% CI 1.18–1.40). Stratified analysis revealed that each of these factors had an independent effect on the non-osteoporotic fractures (n=569), the adjusted HR of dizziness being 1.31 (95% CI 1.05–1.64). Osteoporotic fractures, however, were strongly correlated only with a precedent fracture (adjusted HR 3.91, 95% CI 2.31–6.63), and they were more frequent in female patients (adjusted HR 1.39, 95% CI 0.97–1.98).

Conclusions: Besides the “typical” elderly female patients being at risk of osteoporotic fractures, male patients suffering from dizziness should be carefully evaluated, and prevention strategies should be considered in order to minimize their risk of suffering non-osteoporotic fractures.

Points for discussion: 1) Validity of symptom-coded reasons for encounter such as dizziness; 2) matching.

Presentation 31: Poster

Saturday, 9 May 2009, 11.30–13.00

Melanoma: How to adapt postgraduate and continuing education to recent guidelines?

Juliette Chambe, B. Cribier, C. Perrigouard, R. Attuil, D. Gras

Faculté de médecine de Strasbourg, université de Strasbourg, 22 grand rue, 67000 Strasbourg, France. Tel: + 33 610 15 13 76; E-mail: [email protected]

Background: Melanoma is a rare and severe cancer. Although incidence is stabilizing in other European countries, it is still growing in France. Early detec tion of this type of skin cancer is the only way to reduce mortality. A new guideline promotes a strategy that places general practitioners in the centre of healthcare coordination.

Research question: Using that guideline, training for interns and general practitioners should improve their ability to identify thin melanomas.

Method: The training lasted half a day for the interns and 1 day for the GPs. They analysed the recommendation with a dermatology expert. Short-term evaluation consisted of a pre-test and immediate post-test, with open questions and photo analysis. Mid-term evaluation was sent 3 months later by e-mail. This consisted of open questions, photo analysis, and qualitative evaluation of the training and their change in practice.

Results: 16 interns and 27 GPs were volunteers. Basic knowledge was good in both groups (67% and 63%, respectively). The short-term evaluation at tested to global significant progress (78% and 76%, P<0.05). The best results concerned risk factors and Breslow index. The notion of solar prevention for children was a failure. Three groups of answers emerged in the picture analysis. A significant improve was noticed only in one of them, but recognition of melanoma progressed up to 88%. The mid-term evaluation showed a partial loss of efficiency in analysing malignant lesions, but it did not affect their attitude for 92%. Most of the participants appreciated the training, but wished to complete it with a clinical practice in dermatology.

Conclusions: One-day training improves the skill of interns and general practitioners in detecting early melanoma at short and mid-term. Future training should insist on the use of a risk score and systematic examination.

Points for discussion: How to use risk score for melanoma in everyday practice.

Presentation 32: Poster

Saturday, 9 May 2009, 11.30–13.00

Functional abdominal pain in children in Dutch general practice

Marieke Gieteling, J. van der Wouden, F Schellevis, M. Berger

General Practice, Erasmus MC, Dr Moelenwaterplein 50, 3015 GE Rotterdam, the Netherlands. Tel: [phone]; E-mail: [email protected]

Background: Functional abdominal pain (FAP) is a common complaint in childhood that has a major impact on the child's daily life. Despite high prevalence's found in open population studies, difficulties in management, and poor prognosis reported in secondary care studies, little is known about childhood FAP in general practice.

Research questions: To determine workload and prevalence rate of childhood FAP in general practice, and to evaluate patients’ characteristics and GPs’ management.

Methods: Data of patients aged 2–17 years consulting for FAP were retrieved from the Second Dutch National Survey of general practice (electronic patient record database), carried out in 2001. During 1 year, 90 general practices electronically coded all contact diagnoses, prescriptions, and referrals. We calculated the percentage of consultations for childhood FAP and its prevalence rate per 1000 person-years. We compared the consultation rate for psychological and functional problems between children consulting for FAP and those consulting for other reasons, and we determined referrals and prescriptions for FAP.

Results: FAP accounted for 2.5% (95% CI 2.48– 2.62) of all childhood consultations, and it had a prevalence rate of 63.3 per 1000 person-years. Ninety per cent of patients consulting for FAP only consulted once or twice for this problem. Children who frequently consulted for FAP consulted twice as much for psychological and functional problems (224 [95% CI 189–264] vs. 117 [95% CI 114–121], and 252 [95% CI 215–295] vs. 116 [95% CI 113–119] consultations per 1000 person-years, respectively). GPs prescribed medication in 39.9% of all FAP consultations and referred in 3.6%. During consecutive consultations, GPs referred and prescribed more.

Conclusions: FAP is a significant health problem in Dutch general practice. GPs are highly capable of managing this condition. Children with FAP referred to secondary care are a selection of frequent consulters presenting more often with psychological and functional complaints.

Presentation 33: Poster

Saturday, 9 May 2009, 11.30–13.00

Medical students’ disease

Joseph Azuri, Nissim Ackshota, Shlomo Vinker

Family Medicine, Tel Aviv University, 17 Emeq Bracha St., 67456 Tel Aviv, Israel. Tel: [phone]; Fax: + 972 3 695 4841; E-mail: [email protected]

Background: Medical school exposes students to vast medical knowledge. Studying a specific illness might raise wrong interpretation of physical symptoms, so that symptoms that were otherwise considered normal are now regarded as a true sign of an illness.

Research question: To examine the appraisal of self-health state, existing fear of morbidity, and level of health-related anxiety among medical students throughout their learning period.

Method: Anonymous questionnaires were distributed to 1st-6th year medical students at the Tel Aviv University medical school. The questionnaire was based on validated existing questionnaires. The variables tested were the degree of anxiety concerning health, fear of morbidity, demographic variables, and future desired field of specialization. The questionnaires were analysed according to their perceptual-cognitive vs. emotional-distress components.

Results: An inclination trend can be observed in the perceptual-cognitive process throughout the years of medical study (a rise from 16.71% to 31.11% in the rate of students who replied: to a large extent – to a very large extent, p <0.001). However, we observed a rise in the emotional-distress process during the third year (a rise from 2.82% to 14.22% in the rate of repliers: to a large extent – to a very large extent, p <0.001) and then a descent during the fifth year (from 15.28% to 6.19%, p <0.001). The number of reported clinic visits, medication therapy, or desired field of future specialization were insignificant to both processes. Nonetheless, it seems that medical students express more concern of cancer than cardiovascular morbidity (13.28% to 3.53%).

Conclusions: “Medical student's disease” should be regarded as a phenomenon depending on the years of learning. By breaking it down to two components, one can better characterize it and predict its beginning. Defining it as a normal process, one can assist in guiding medical students during their studies, thus reducing the level of anxiety.

Points for discussion: 1) Is this phenomenon a worldwide characteristic of medical studies? 2) Can this phenomenon be characterized by other variables (personal medical history, future desired field of specialization, etc.)?

Presentation 34: Poster, ongoing study with preliminary results

Saturday, 9 May 2009, 11.30–13.00

Data quality improvement in a cohort of diabetes patients in general practice

Debby Keuken, R. Bruessing, P. Keijser, H. Brouwer

Department of General Practice, Academic Medical Center, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands. Tel: [phone]; E-mail: [email protected]

Background: In 2007, a diabetes cohort was initiated by the Department of General Practice/Family Medicine of the Academic Medical Center (AMC), University of Amsterdam, the Netherlands. The aims were to enhance quality improvement of diabetes care and to provide an infrastructure for diabetes-related research. Since 2007, registration of diabetes care indicators has been implemented as part of a disease management programme. Previous studies have shown difficulties in obtaining a complete set of diabetes indicators.

Research question: What is the quality of diabetes care indicator registration in an established research network of general practitioners (GPs), and how can this be improved?

Method: We used a subset of data from the GP registration network of the AMC. In this network, data are recorded in electronic medical records on a routine basis and used for research purposes. For the 2585 patients known by their GPs with diabetes type 2, practice-based process measures for the diabetes care indicator HbA_1c is calculated over the year 2007 (12-month period). These values are presented to the 37 practices in the network during feedback sessions. For the records of patients with missing values for HbA_1c, the reasons for the missing values were investigated by the GPs. Causes for missing values were discussed in an expert team, and main limitations for indicator recording were formulated.

Results: Of the 2585 patients diagnosed with diabetes type 2, 772 (29.9%) HbA_1c values were missing in 2007. Analysis of missing HbA_1c indicator reveals several causes, including the patient is under specialized care, the patient does not attend his/her appointments, and the patient is not invited within the 12-month period. Limitations for indicator recording are formulated.

Conclusions: Our results demonstrate that various causes exist for missing HbA_1c values. Formulated limitations will be used to improve the quality of indicator registration.

Points for discussion: 1) Are the causes for missing indicator data presented here familiar to the attendants of the EGPRN? 2) Target levels for indicators of the diabetes care process should never be 100%. 3) High-quality electronic medical records are a basic requirement.

Presentation 35: Poster

Saturday, 9 May 2009, 11.30–13.00

Does diabetic patients’ ethnic background influence quality of care received in general practice?

Anh Thi Tran, A. K. Jenum, J. G. Cooper, T Claudi, M. F. Hausken, W Ingskog, J. Straand

Section for General Practice/Family Medicine, Institute of General Practice and Community Health, University of Oslo, Postboks 1130 Blindern, NO-0317 Oslo, Norway. Tel: + 47 22850 550; Fax: [phone]; E-mail: [email protected]

Background: A multiethnic general practice is challenging for GPs due to ethnic variations in prevalence and clinical course for common illnesses such as diabetes type 2 (DM2).

Research question: Does patients’ ethnic background influence the quality of care received for DM2 in general practice?

Method: We used software designed for this study to identify patients in GPs’ electronic patient records (EPJ). For patients with a diabetes diagnosis, predefined data were captured, e.g., ethnicity, measurements of HbA_1c, blood pressure (BP), cholesterol, eye investigation, smoking habits. For each identified patient, captured data were manually checked on site by an experienced research nurse. Ethnicity was categorized according to country of birth.

Results: In 2005, about 58,000 patient records in 11 practices (49 GPs) were screened. Of these, 2064 patients had a diabetes diagnosis, of whom 1653 had DM2 and were cared for by their GP, and were included in this study. Mean age varied across ethnic groups (from 52.3 to 66.4 years); native Norwegians were oldest. In all ethnic groups, most had their HbA_1c (91.4 to 95.2%), BP (85.2 to 92.5%), and cholesterol (92 to 97%) controlled. Immigrants were treated more intensely with oral hypoglycaemic agents (OHAs), or combined OHAs and insulin, whereas 18% of all minority patients vs. 28% of Norwegians were non-users. Compared to Norwegian DM2 patients, immigrant patients in all treatment groups had significantly higher HbA_1c (7.4 vs. 7.1% for OHA only, 8.4 vs. 7.9% for OHA and insulin combined, and 8.6 vs. 7.7% for insulin). Of the ethnic groups, 56.1-72.6% achieved treatment target for hyperglycaemia (HbA_1c <7.5%). Minority groups had lower BPs and received less anti-hypertensive therapy and statins.

Conclusions: Even if immigrant DM2 patients generally were younger than corresponding Norwegians, their glycaemic control was worse despite receiving more intensive treatment with glucose-lowering therapies.

Presentation 36: Poster, ongoing study with preliminary results

Saturday, 9 May 2009, 11.30–13.00

Ve.Di.Cli.S. Project (Verona Diabetes Clinical Study): Continuous quality improvement (CQI) in primary care management of type 2 diabetes

C. Andreoli, F. Del Zotti, S. Girotto, Giulio Rigon, A. Vaona

General Practice – Verona, Via Regaste San Zeno 37, 37123 Verona, Italy. Tel: [phone]; E-mail: [email protected]

Background: The increasing prevalence of type 2 diabetes in the western world brings many problems and challenges. The high-quality management of this condition at primary care level is crucial for sustainability of the healthcare system.

Research question: The aim of the study is to measure the quality of diabetes care among a group of Italian GPs during a period of 3 years.

Method: NICE clinical guidelines for type 2 diabetes (2005) were identified as the best guidelines because of high methodological quality. A panel of diabetes care quality indicators was identified to check the performance of a group of 30 GPs in Verona during a period of 3 years. Each indicator was adapted to the organizational context in which the GPs operate, by a focus group session. The group of GPs underwent four learning sessions (a full day for each session) aimed to standardize and improve knowledge of evidence-based management of diabetes. The performance data were collected by electronic queries sent by e-mail. The participants were required to run the queries in their professional database and send back the data to the scientific committee. The discussion continues online on the blog of the project.

Results: Table I shows the baseline (09/2008) performance data compared with data coming from another national database (Health Search) and the mean optimal theoretical performance level ex pressed by the participants.

Conclusions: At baseline level, the performance data in the Ve.Di.Cli.S GP group is better than the national level, but it is far from what GPs think it should be. After basic training, an e-mail/blog network is an efficient system for updating and collecting performance data.

Points for discussion: Table I. Results of baseline performance data (Sept. 2008). Ve.Di.Cli.S Criterion Baseline Health Search Standard% of Type 2 diabetes 5.30% 5.4% BMI 46% 35.4% 80% HbA .

Presentation 37: Poster

Saturday, 9 May 2009, 11.30–13.00

Managing type 2 diabetes mellitus in primary care: Measuring or controlling?

Stella Argyriadou, Charilaos Lygidakis, Anastasia Lygera, Eleni Thomaidou, Haritini Melissopoulou, Theodoros Makris

Primary Care/General Practice, Health Centre of Chrisoupolis, Mariou 7, 64200 Chrisoupolis, Greece. Tel: + 30 69 74875 603; E-mail: [email protected]

Background: Although personal nutritional profile and self-management is the cornerstone of type 2 diabetes mellitus (DM2) care, only a few patients are able to achieve long-term dietary changes and good metabolic control.

Research question: To what extent can an intervention based on a Mediterranean dietary model and nutritional self-managing improve DM2 control?

Method: GPs examined 500 users of health services of the Chrisoupolis Health Centre (age >44 years), collecting demographics, medical history, chronic therapies, and lifestyle information. Blood pressure, body-mass index (BMI), and laboratory examinations were performed, including lipid count, fasting blood glucose (FBG), and, in case of DM2, HbA_1c. Diabetics were randomly divided into two groups: a conventional treatment group (CG), with personalized diet and therapy, and an intervention group (IG), which followed a 6-hour educational course about Mediterranean nutrition, ways of quantity-and quality-controlling food consumption, and meal self-planning methods. A blood glucose monitoring system was provided to IG participants, and monthly follow-up meetings were scheduled. After 3 months, FBG and HbA_1c were re-evaluated.

Results: From 156 diabetics, 11 were newly diagnosed. Diabetics had a significantly higher BMI (median 30.00, p =0.005), and only 12 had HbA_1c levels within the recommended range (<7%). There was a significant interaction effect between the type of therapeutic strategy and multiple measures of FBG (F=25.34, p<0.001) and HbA_1c (F=49.13, p<0.001). Patients of the IG presented significantly lower values of FBG (median 120, U= 1268.00, p<0.001) and HbA1 (median 8.4%, U= 1159.50, p< 0.001) in the follow-up. Sixteen of them versus eight of the CG resulted in having a controlled HbA1 (<7%, p=0.003).

Conclusions: Motivating diabetics to become actively involved in their dietary assessment seems to have positive effects on their disease management. Future studies could be conducted by GPs to investigate the long-term effects of this methodology.

Presentation 38: Poster

Saturday, 9 May 2009, 11.30–13.00

Symptomatic treatment or antibiotics for uncomplicated urinary tract infection?

Jutta Bleidorn, Ildikó Gágyor, Michael M. Kochen, Eva Hummers-Pradier

Medical School Hanover, Institute for General Practice, Carl-Neuberg-Str. 1, 30 625 Hannover, Germany. Tel: + 49 511 532 2744; Fax: [phone]; E-mail: [email protected]

Background: Uncomplicated urinary tract infections (UTI) are common in general practice and are usually treated with antibiotics. This study was conducted to assess whether symptomatic treatment with ibuprofen was equally effective as cipro?oxacin for uncomplicated UTI.

Research question: 1) To assess the equivalence of ibuprofen and cipro?oxacin for uncomplicated UTI with regard to symptom resolution; 2) to demonstrate the feasibility of a randomized, controlled, double-blind drug trial in German general practices.

Method: A double-blind, multicentre, randomized, controlled equivalence trial was conducted in 29 German general practices. Inclusion criteria were typical symptoms of uncomplicated UTI. Patients were randomized to receive either ibuprofen (3 × 400 mg/d) or cipro?oxacin (2 × 250 mg/d) for 3 days. Intensity of main symptoms—dysuria, fre quency, low abdominal pain—was recorded at in clusion and after 4, 7, and 28 days, scoring each symptom from 0 (none) to 4 (very strong). The main clinical outcomes were symptom resolution on day 4 and number of patients with secondary antibiotic treatment.

Results: A total of 79 women presenting with at least one of the main UTI symptoms—dysuria and frequency—were analysed. On day 4, 21/36 (58.3%) patients in the ibuprofen group were symptom free versus 17/33 (51.5%) in the cipro?oxacin group. Ibuprofen patients reported fewer symptoms in terms of total sum score (1; SD 1.42) than those in the cipro?oxacin-group (1.3; SD 1.9) (difference −0.33 [95% CI −1.13; +0.47]). However, during the first week, 12/36 (33%) patients in the ibuprofen group re-consulted with ongoing or worsening symptoms, resulting in secondary antibiotic prescriptions, versus 6/33 (18%) in the ciprofeoxacin group.

Conclusions: There is a strong tendency towards equivalence of ibuprofen and cipro?oxacin for treatment of uncomplicated UTI. If this can be confirmed by further trials, antibiotics could be reserved for patients with persistent or recurrent symptoms, possibly resulting in a decrease of antibiotic use.

Points for discussion: Change of treatment strategies for uncomplicated UTI.

Presentation 39: Poster, ongoing study with preliminary results

Saturday, 9 May 2009, 11.30–13.00

Influenza and pneumococcal vaccination and screening rates of the elderly in a private outpatient clinic

Pemra C. Ünalan, Demet Dinç, Pinar Topsever

Family Medicine, Marmara University, MESA Koruevleri C4ID2 Altunizade, 34662 Istanbul, Turkey. Tel: + 90 216 3275612; Fax: [phone]; E-mail: [email protected]

Background: The influenza and pneumococcal vaccination (PPV) policy for elderly people in Turkey has changed in the last 3 years. The National Health Insurance System pays for these vaccines if the patient is over 65 years of age and in case of other known indications. However, these vaccinations are still not introduced as a strategic health objective for the elderly according to a defined age limit by the Ministry of Health.

Research question: What are the vaccination and screening rates of the elderly (over 65 years of age) if there is not reminder in the records?

Methods: One reason for non-vaccination is the payment made by the patient. To overcome this factor, we conducted a retrospective descriptive study including all patient records of patients over 65 years of age between January and December 2008 assigned to all outpatient clinics of a private hospital. Electronic patient records did not contain a prompt for immunization information, so these computerized clinical records were screened with the search terms “influenza”, “pneumococcal”, or “vaccination”. Descriptive data relating to those terms are evaluated manually.

Results: Eight people from a total of 15,889 encounters were asked about influenza vaccination and were immunized by their physicians. Five of them were men, and the mean age was 77.5 years. Six of the 15,889 encounters were asked about PPV; three of them were men, and the mean age was 72 years.

Conclusions: These results are alarming and emphasize the need to improve vaccine uptake. Communicating individually and monitoring the progress of vaccination are known to be helpful practices, but since electronic patient records do not have a standard window for this inquiry, it is very difficult to determine exact annual vaccine coverage and to discuss the factors that affect this result.

Points for discussion: Can we find any other way (method) to address our research question using electronic records?

Presentation 40: Poster, ongoing study with preliminary results

Saturday, 9 May 2009, 11.30–13.00

The MISURA project: A multidisciplinary prevention approach to decrease CV risk

Enzo Pirrotta, Maurizio Briganti, Alfredo Cuffari, Stefania Graziosi, Marianna Maranghi

Snamid Rome, SNAMID, via Aosta 42, 00182 Roma, Italy. Tel: + 39 0670 304368/mobile [phone]; E-mail: [email protected]

Background: MISURA is an acronym that can be translated as “keep yourself fit by regular habits”. It is a plan for active health promotion addressed to asymptomatic people at presumed high cardiovascular (CV) risk according to the 2007 Fourth Joint Task Force of the European Society of Cardiology (ESC) on cardiovascular risk prevention in clinical practice. Obstacles for people trying to change their lifestyle are socioeconomic status, social isolation, stress, negative emotions, and complex or confusing advice.

Research question: Is a multidisciplinary prevention approach to decreasing CV risk an effective and feasible tool in the primary care setting?

Methods: Data extracted from GPs electronic medical records (EMRs) will be shared with the University Medical Department “La Sapienza” in Rome. The project started in October 2008 and is planned to last 2 years. Three hundred and twenty people in good health and asymptomatic (160 for the interventional group and 160 for the control group), aged 35 to 59 years, at presumed increased CV risk will be enrolled. Total CV risk will be estimated using the Istituto Superiore di Sanità Progetto Cuore Individual Risk Score. A traditional prescriptive methodology will be used in the control group, while in the interventional group patients and their families will receive a 2-hour monthly structured education programme, involving GPs and psychologists. Outcomes: The process and quality indicators according to the ESC IV JTF are being evaluated.

Results: Enrolment is still in progress. So far, we have found 7% new-onset diabetes and 22% metabolic syndrome. CV risk score ranges from 1.8% to 24.5%, and BMI from 24.2 up to 42.1 kg/m².

Conclusions: A possible solution to overcoming obstacles in cardiovascular prevention is a new participatory and active model of management of chronic diseases, which includes a multidisciplinary approach and an effective partnership between primary and specialist services.

Points for discussion: 1) Relevance of the study in primary care; 2) methodology.

Presentation 41: Poster, ongoing study with preliminary results

Saturday, 9 May 2009, 11.30–13.00

Cardiovascular risk factor prevalence and therapy adherence in 502 patients with previous cardiovascular and/or cerebrovascular event

Matteo Rispoli, Marcello Barbarisi, Giovanni Cracas, Giovanna Di Buono, Bartolomeo Salvati, Antonietta Santoro, Ester D Arienzo, Alessandro Renna, Angela Fortunato, Salvatore Telese

Azienda Sanitaria Locale SA2, ASLSA2, Piazza G. Nicotera 6, 84127 Salerno, Italy. Tel: [phone]; Fax: + 39 089798667; E-mail: [email protected]

Background: Many studies have shown a high prevalence of risk factors in patients with previous acute cardio-cerebrovascular event, and a high rate of patients with poor therapeutic adherence, poorly motivated to change their lifestyle. These data motivated us to perform this evaluation in our practice.

Research question: What are the risk factor prevalences and cardiovascular drug utilization in the population with previous acute cardio- and/or cerebrovascular event in our practice?

Method: From April to May 2008, patients with previous diagnosis of stroke, transient ischemic attack (TIA), and/or coronary artery disease were selected from the database of the electronic medical records of 15 general practitioners in Salerno, southern Italy. Data collected were: blood pressure, smoking habit, waist circumference, glycaemia, triglycerides and cholesterol level, and drug therapy.

Results: 502 patients were enrolled (263 males, mean age 66 ± 10 years). High level of cholesterol (according to NCEP ATPIII) was found in 54 female (F) and 85 male (M) patients; hypertension (according to ESC-ESH 2007) in 120 M and 180 F; smoking habit in 80 M and 72 F; obesity (ESC2007) in 70 M and 57 F; hypertriglyceridaemia in 60 M and 40 F; and diabetes mellitus (ADA) in 50 F and 70 M. Seventy per cent of the patients were on platelet inhibitors, 40% on ACE-inhibitors and diuretics, 30% on nitrates, 20% on angiotensin receptor blockers (ARBs), 21% on beta-blockers, 20% on calcium antagonists, 10% on alpha-blockers, and 66% on statins (47% simvastatin, 20% atorvastatin, 20% pravastatin, 13% others).

Conclusions: In our population, there is still a high level of modifiable risk factors that can be dealt with by lifestyle change or medical therapy. Secondary prevention is regularly performed only a few months after the event, but then adherence to therapy and lifestyle tends progressively to worsen.

Points for discussion: 1) Relevance of the study; 2) validity of the methodology.

Presentation 42: Poster

Saturday, 9 May 2009, 11.30–13.00

Accuracy of symptoms and signs for coronary heart disease in primary care: A diagnostic study with follow-up

Stefan Bösner, Annette Becker, Maren Abu Hani, Heidi Keller, Andreas C. Sönnichsen, Jörg Haasen-ritter, Erika Baum, Norbert Donner-Banzhoff

Department of General Practice/Family Medicine, Karl-von-Frisch Strasse 4, 35043 Marburg, Germany. Tel: [phone]; Fax: [phone]; E-mail: [email protected]

Background: Diagnosing the aetiology of chest pain is challenging. GPs have to identify serious cardiac disease reliably, but they also have to protect patients from unnecessary investigations and hospital admissions. Data on the diagnostic accuracy of signs and symptoms for acute coronary events in low-prevalence settings are still lacking.

Research question: What is the diagnostic accuracy of symptoms and signs in patients presenting with chest pains in general practice?

Method: A cross-sectional diagnostic study with delayed-type reference standard was conducted in 74 GP surgeries in Germany with 1249 patients consecutively presenting with chest pains. GPs recorded the symptoms and findings of each patient on a report form. Patients and GPs were contacted at 6-week and 6-month intervals. Data on chest complaints, investigations, hospitalization, and medication were reviewed by an independent reference panel. Coronary heart disease (CHD) and indication for urgent hospital admission were the reference conditions. Main outcome measures were sensitivity, specificity, likelihood ratio, predictive value, and odds ratio for different variables of patients with a reference diagnosis. A secondary analysis was conducted for acute presentations, i.e., symptom duration <48 hours.

Results: The following signs and symptoms showed associations with CHD: known vascular disease, pain worse on exercise, patient assumes cardiac origin of pain, age, known heart failure, and diabetes mellitus. Negative associations were found for: stinging pain, cough present, localized muscle tension, and pain reproducible on palpation. For “indication for urgent hospital admission”, the following criteria were effective: pain radiating to left arm, known clinical vascular disease, home visit requested, known heart failure, and pain reproducible on palpation. Pain localization and temporal pattern, sex, and risk factors other than diabetes did not contribute.

Conclusions: Although individual criteria were only moderately effective, in combination they can contribute as an effective filter function in primary care.

Points for discussion: 1) Is there an “ideal reference standard” for diagnostic studies? 2) Methodological aspects for choosing univariate para meters for multivariate analysis; 3) limited diagnostic value of pain localization and temporal pattern of pain.

Presentation 43: Poster

Saturday, 9 May 2009, 11.30–13.00

Gestational hypertension in general practice: Screening and orientation of a pathological pregnancy by the general practitioner

Catherine Laporte

General Practice, University of Clermont 1,15 Rue De Cournon, 63000 Clermont-Ferrand, France. Tel: [phone]; E-mail: [email protected]

Background: Gestational hypertension (GH) is the first complication in pregnancy and the second cause of maternal death. Although delivery is the only available treatment, a decrease in morbimortality is observed through early screening and monitoring.

Research question: The goal of this study was to explore and better understand the attitude of general practitioners (GP) when confronted with GH, and to propose optimizations in the role of the GP.

Method: A qualitative study was pursued through semi-directive interviews of 14 liberal GPs in the Gard area between May and July 2008. These GPs monitored more than five pregnancies per year. The interviews were recorded and literally transcribed. A thematic analysis was then performed.

Results: GH, regarded as potentially hazardous, was systematically screened by GPs: monitoring of arterial pressure, proteinuria, and other clinical signs. The decision to reroute the patient to specialized monitoring was the negotiated consequence of theoretical knowledge, professional experience, and medical and paramedical environment. GPs re gretted the weak initial and ongoing training, the lack of coordination between professionals (obste triciansgynaecologists and midwives), and the undervaluation of their role in perinatality.

Conclusions: Management of GH could be improved by: training of GPs through an adaptation to the difficulties of practice; coordination between health professionals through a more pertinent action of the regional perinatality network; and clarification of the GP's role in the monitoring of pregnancy through rethinking the organization of the health system.

Points for discussion: Perinatality, gestational hypertension, French health system.

Presentation 44: Poster, ongoing study with preliminary results

Saturday, 9 May 2009, 11.30–13.00

Patient's gender influences delivery of preventive services in primary care

Ewelina Gowin, Wanda Horst-Sikorska, Magdalena Ignaszak-Szczepaniak, Joanna Dytfeld

Family Medicine, University of Medical Sciences Poznan, 49 Przybyszewskiego Street, 60–355 Poznan, Po land. Tel: [phone]; E-mail: ewego@poczta. onet.pl

Background: Most healthcare systems have recommended several preventive services to be delivered in primary care. Given the limited time and resources of GPs, it is necessary to choose which preventive procedures to offer and to whom.

Research question: What is the effect of patient's gender on the directly observed delivery rates of preventive procedures to 40-plus persons in primary care?

Method: The observational study was performed in 113 Polish primary care surgeries. A sample of GPs who work in Wielkopolska were asked by phone to participate in the study, and those who agreed were visited by observers recruited from final-year medical students. The observers directly recorded preventive procedures performed during the consultation of four consecutive patients aged 40 + . The collected data were blinded for the researchers. The GPs were not informed about the nature of the observation. For statistical analysis, the t test and chi-square test were used.

Results: A total of 1073 preventive procedures were registered during 450 consultations in 267 women and 183 men by 113 doctors. Mean number of procedures per patient was 2.37 (SD 1.6). Eleven categories of preventive procedures were recorded. The most common procedures were: serum glucose in 68%, blood pressure in 58.4%, and total cholesterol in 55.6% of patients. Six out of 11 procedures were offered to less than 10% of patients: dietary, exercise, and anti-alcohol advice, BMI, waist measurement, and anti-smoking. Men were more likely to receive anti-smoking (p=0.02) and anti-alcohol advice (p =0.0004) and BMI measurement (p = 0.02), while more women had total cholesterol measured (p=0.01).

Conclusions: 1) This study shows that the delivery of preventive services in Polish primary care falls below recommended levels. 2) Secondary prevention based on clinical examination and laboratory tests is performed more often than lifestyle advice. 3) Men receive more preventive procedures, in particular anti-smoking, anti-alcohol advice, and BMI measurements.

Points for discussion: 1) Which preventive procedures are of greatest value for GPs, and which are most valued by patients? 2) What are the differences in routine consultations of female and male patients?

Presentation 45: Poster, ongoing study, no results yet

Saturday, 9 May 2009, 11.30–13.00

Supplied and demanded services in general practice

Margreet Schipper, Nettie Blankenstein

VUUniversity Amsterdam, Conradstraat 80G, 1018 NK Amsterdam, the Netherlands. Tel: [phone]; E-mail: [email protected]

Background: In the Netherlands, there is an ongoing shift of services from secondary care to primary care. General practices are increasingly supplying multiple and more diverse services, e.g., evening surgery hours, vasectomy, health checks, while some existing services disappear, e.g., obstetrics, friendly visits. Little is known about the actual demand within a patient population. Also, little is known about the willingness of GPs to adapt to supplying other/more services.

Research questions: 1) What services are demanded by patients? 2) Do patient characteristics or patient populations influence this demand? 3) Are GPs willing to adapt and change their supply according to their patients’ demands?

Method: Patient questionnaires: Four general practices (in Purmerend, Amsterdam, and Hoofddorp) send questionnaires to samples of 300 patients each. In this questionnaire, demographic data and chronic diseases are collected as well as patients’ current use of services and the services they would like to be available at their general practice. GP questionnaires: The GPs working in the four participating practices will fill out a questionnaire on which services they are offering and which services they would be willing/unwilling to offer if their patients asked for them. GP interviews: In individual interviews, GPs consider the results: Are they willing to supply new services that are demanded by their patients? Are they willing to stop existing services that are not demanded by patients?

Analyses: Questionnaire scores will be analysed in Microsoft Access and Excel. Differences in demands between patient populations will be computed at the item level. Interviews will be audio-taped, and GPs’ considerations will be extracted qualitatively.

Results: Will be available in May 2009. Demands will be presented quantitatively, specified for different patient populations and characteristics. GPs considerations will be presented qualitatively.

Conclusions: Will be available in May 2009.

Points for discussion: 1) Can our Dutch results be generalized to other European countries? 2) How should GPs handle dilemmas arising from patient feedback, e.g., patients demand evening surgeries whereas GPs hate working late, or patients prefer hospital care while the pr ...

Presentation 46: Poster

Saturday, 9 May 2009, 11.30–13.00

To what extent are core competencies addressed in European general practice research? Classification of EGPRN meeting abstracts, 2002–2007

Yenal Izci¹, Pinar Topsever²

¹Family Medicine, Kocaeli University, Turkey, and

²Sallandstraat 28, 7412 WD Deventer, the Netherlands. Tel: + 31 6476 90695; E-mail:[email protected]

Aims: To investigate the use of core competencies in European general practice research in order to contribute to the development of a European research agenda by the EGPRN.

Method: Abstracts from 11 EGPRN meetings between 2002 and 2007 were descriptively evaluated according to date and city/country of meeting, country of origin of the principle investigator, multi-centre/international study design, and the six core competencies (CC) of the WONCA 2005 European definition of GP/FM. Results are presented as descriptives and frequencies (w,%).

Results: A total of 526 abstracts were analysed. When grouped according to countries, Turkey and Germany with 68 each (13%), the Netherlands with 60 (11%), France with 44 (8%), and the UK with 40 (8%) abstracts constituted 53% of all analysed abstracts, while the remaining 47% of abstracts originated from 26 other countries. “Specific problem-solving skills” (n=202; 38%), “primary care management” (n=110; 21%), and “patient-centred care” (n=103; 20%) were the most frequently addressed CCs, followed by “comprehensive approach” (n=76; 14%), “holistic approach” (n= 16; 3%), and “community orientation” (n=75; 3%). Four abstracts (1%) could not be assigned to any CC. Eighty-six abstracts (16%) matched a second CC: “specific problem-solving skills” (n=30; 35%), “patient-centred care” (n=21; 24%), “primary care management” (n=15; 17%), “comprehensive approach” (n=9; 10%), “community orientation” (n=6; 7%), and “holistic approach” (n=5; 6%). Forty-four studies (8%) were international, and 127 (24%) were national multicentre studies.

Conclusions: Most studies submitted to EGPRN meetings addressed “specific problem-solving skills”, whereas there was almost no attention to the “holistic approach” and “community orientation” competencies. This finding suggests several core competency-related research gaps in European general practice research, which is likely to be mirrored in the EGPRN abstracts.

Presentation 47: Poster

Saturday, 9 May 2009, 11.30–13.00

Antidepressant prescription analysis in primary care: A Verona survey from 2001 to 2005

Chiara Roni, R. Joppi, M. Andretta, C. Anecchino, G. Danti, F. Del Zotti, M. Font, L. Mezzalira, M. Romero, G. Tognoni, on behalf of the ISD Working Group

Pharmaceutical, Verona Local Health Unit “ULSS 20”, via Salvo d’Acquisto 7, 37122 Verona, Italy. Tel: + 39 045 8076057; Fax: [phone]; E-mail: croni@ ulss20.verona.it

Background: New antidepressants in particular have been increasingly promoted for prescription in primary care (PC) for an even wider range of depressive symptoms and disorders, despite the substantial lack of conclusive evidence specifically produced in PC settings.

Research question: What kinds of subjects are treated with antidepressants in Verona's PC setting? In what way and for how long are they treated?

Method: A retrospective surveillance of prescription data on 9065 patients (3.9% of Verona's population), aged 21–60 years, was conducted. The included patients had at least one GP prescription of antidepressants in the period March 2001–February 2002. The selected cohort has been followed up for 3 years. On the basis of the prescribed defined daily dose (DDD) number, three categories were identified: 1) patients who received less than 90 DDDs within 90 days were considered “poorly exposed”; 2) patients who received more than 240 DDDs within 240 days were considered “highly exposed”; 3) all other patients were considered “at intermediate exposure”.

Results: Women represented 67.8% of the studied population. The prevalence of antidepressant use increased from 1.9% at 21–30 years old to 6.2% at 51–60 years old. Seventy-eight per cent of the cohort were treated with newer antidepressants; 24% of patients were treated for up to 3 months, 11% showed an erratic pattern, and 7% were considered under chronic treatment (“highly exposed”).

Conclusions: The results apparently suggest that the majority of patients under antidepressants are not treated in accordance with current guidelines for depression (heterogeneity of depressive disorders; short and repeated treatment cycles) and are poorly compliant. The amount of chronic depressive patients seems to be relevant.

Presentation 48: Poster

Saturday, 9 May 2009, 11.30–13.00

Happiness among the elderly in communities: Thai Happiness Indicators (THI-15)

Chommanard Sumngern, Zaida de Azeredo, Rarchaneeporn Subgranon, Narirat Sungvorawongphana, Eduarda Matos

Community Health, Institute of Biomedical Science of Abel Salazar, University of Porto, Largo Prof Abel Salazar 2, 4099–003 Porto, Portugal. Tel: [phone]; Fax: [phone]; E-mail: chommanard @icbas.up.pt

Background: Happiness is an individual emotional or affective state that is characterized by a feeling of enjoyment and satisfaction. In relation to the elderly, this can lead to successful aging, resulting in a good life.

Research question: 1) How does happiness among the elderly differ across regions? 2) What are the factors that affect happiness in Thai older adults?

Method: Thai Happiness Indicators (THI-15) were used in this study of 306 participants 60 years of age and older in rural, suburban, and urban regions. The elderly were randomly chosen among those in Senior Thai Clubs.

Results: The elderly perceived their happiness as “good” (12%), “fair”(25%), and “poor”(62%). Among the regions, there was a difference (p<0.001) in happiness perception: 64.5%, 61.2%, and 22.8% in the rural area, suburban area, and urban area, respectively, considered their happiness as “poor”. We also verified that there was a difference (p<0.01) in “poor” happiness between male (57.9%) and female (42.2%). Education is related to happiness: those with higher education perceived “good” happiness more than those with lower education (p=0.007). There was also a difference between happiness and occupation; the results showed that the elderly who did not work had a better score of happiness than the elderly who had to work, especially those who worked on farms, who had the highest percentage in terms of “poor” happiness.

Conclusions: Happiness could be one of the indicators of mental health among the elderly in the community that healthcare providers should be concerned about, using proper interventions and/or activities in order to alleviate mental health problems among the elderly in the community, especially those who indicated “poor happiness”.

Presentation 49: Theme paper

Saturday, 9 May 2009, 15.20–15.50

Garbage in, garbage out? Validity of claims-based diagnoses as morbidity indicators in German general practice

Antje Erler, M. Beyer, C. Muth, F. M. Gerlach, R. Brennecke

Goethe University Frankfurt Main, Institute of General Practice, Theodor-Stern-Kai 7, 60590 Frankfurt am Main, Germany. Tel: + 49 69 6301 4132; Fax: [phone]; E-mail: [email protected]

Background: Coded diagnoses from claims records are frequently used as morbidity indicators. In Germany, the mandatory coding system is ICD-10 instead of ICPC, and documentation and coding of diagnoses are mainly performed for administrative and remuneration purposes.

Research question: The aim of the study was to examine whether claims-based diagnoses adequately describe episodes of illness of patients treated over a 1-year period, or whether an over-reporting of diagnoses occurs.

Method: In a retrospective cross-sectional study of a random sample of 250 patients from 10 general practices, we identified all health problems treated in the year 2003 on the basis of the patients’ medical records and summarized them as episodes of illness. We used percentages, kappa value, sensitivity, and positive predictive value to describe the match, correctness, and completeness of the most frequently coded episodes of illness compared to claims-based diagnoses within the same time period.

Results: In spite of a high mean of 6.1 claims-based diagnoses per patient, episodes of health problems were under-reported in 30% of cases, mainly relating to non-severe diagnoses frequently encountered in general practice, chronic conditions not requiring medication, and diagnoses justifying a screening test. An over-reporting for diseases not treated within the study period was observed in 19% of cases, most often in the case of permanent chronic conditions. In 11% of cases, the ICD-10 codes of claims-based diagnoses and the diagnoses in the medical records did not match (“erroneous codes”). For six of the diagnoses most common in general practice (hypertension, diabetes, hyperlipoproteinaemia, cardiovascular disease, back pain, and acute respiratory tract infections), correctness at 71 to 93% was higher than completeness (56 to 86%).

Conclusions: The suitability of claims-based diagnoses to describe morbidity in general practice is limited. Morbidity will be more often under- than overestimated when claims-based diagnoses are used as indicators.

Points for discussion: Is it possible to compare German morbidity data from general practice with data from other countries which use an episode-based documentation system and ICPC for coding of health problems?

Presentation 50: Theme paper

Saturday, 9 May 2009, 15.50–16.20

Incidence and clinical manifestations of rheumatic fever: Electronic medical records-based survey

Shlomo Vinker, Efrat Zohar, Robert Hoffman, Tuvia Baevsky, Asher Elhayany

Family Medicine, Sackler School of Medicine, Tel Aviv University, 77041 Ashdod, Israel. Tel: [phone]; Fax: [phone]; E-mail: [email protected]

Background: Most data about the incidence of rheumatic fever (RF) are collected from hospital records. We presumed that there may be cases of RF that did not require hospitalization, especially in countries with high-quality community healthcare.

Research question: To explore the incidence and characteristics of RF using community electronic medical records.

Method: A retrospective descriptive study among the members (more than 450,000) of the Clalit Health Services Central district during 2000–2005. The electronic medical records of members up to 40 years of age who had a diagnosis of RF in hospital discharge letters or during community clinic visits were retrieved. Patients with first episode of RF according to the modified Jones criteria were included.

Results: 44 patients had a first episode of RF. All patients were younger than 29 years of age. The annual incidence among patients aged 0–30 years was 3.2/100,000. Twenty-five per cent of the patients were both diagnosed and treated in an ambulatory care setting. The highest incidence was among children aged 5–14 years (7.5/100,000), and in males it was 2.26 times higher than in females. The incidence was higher among patients from large families, non-Jewish ethnicity, and rural areas.

Conclusions: A quarter of new cases of RF were treated in an ambulatory setting. Using the electronic medical record, we were able to evaluate the incidence and characteristics of RF. Any future estimates of the disease incidence should take into account that the disease is becoming an ambulatory-treated disease.

Points for discussion: 1) The medical record as a tool for the evaluation of disease incidence; 2) the shift of treatment of rheumatic fever to the community.

Presentation 51: Theme paper

Saturday, 9 May 2009, 16.20–16.50

Out-of-hours primary care in Belgium: Registration of reasons for encounter, diagnoses, and referrals

Philippe Ryckebosch, J. Brouns, J. Goris, E. Vermeire, H. Philips, P. van Royen

Hutsartsenwachtpost Deurne-Borgerhout, Ivan Maquinaylei 12, 2100 Deurne, Belgium. Tel: [phone]; E-mail: [email protected]

Background: In Belgium, during the weekends and on holidays, patients can turn to a general practitioner (GP) on call or go straight to a hospital's emergency department (ED). Unfortunately, there are no data as to what kind of problems GPs encounter when they are on call, which problems they handle themselves, or which patients are referred to a hospital.

Research question: What kind of medical problems do GPs encounter when they are on call? Which problems do they handle themselves, and which patients are referred to a hospital or to their GPs’ scheduled hours?

Method: In 2003, the first general practitioner deputizing service was opened in Deurne-Borgerh out. A new software package, the HWP-Mailer, was developed. The reason for encounter (RFE) and the main diagnosis are registered via ICPC2-ICD10 (Wonca) and the Belgian version of Thesaurus within a new codification module. The referral to an ED, to the patient's own GP, or the decision to have the patient referred for hospital admission are stored in the database. As such, data are available on RFEs, diagnoses, frequencies, and reasons for which people are referred.

Results: The most common RFEs are: R05 Cough, A03 Fever, and R21 Throat symptom/complaint. The most common diagnoses are: R74 Upper respiratory infection acute, D73 Gastroenteritis presumed infection, R78 Acute bronchitis/bronchiolitis. Of all patients, 6.44% were referred to a hospital-m-4.19% to the ED (most common diagnosis D88 Appendicitis) and 2.25% for admission (K90 Stroke/ cerebrovascular accident)—and 3.7% (R78 Acute bronchitis/bronchiolitis) were referred back to their own GPs.

Conclusions: With the development of a new out-of-hours health record, figures are available about the reasons for encounter, diagnoses, and referrals. These data will be helpful in refining the process of information gathering in primary care in Belgium.

Points for discussion: 1) Do RFEs/diagnoses registered by the GP on call differ from those during normal scheduled hours? 2) Do you know of similar studies of out-of-hours services in other countries? Were the results similar or different? 3) Could these data help us in planning future organisation of out of hours primary care services?