Summary
Background
Current guidelines for the prevention of cardiovascular disease emphasize the importance of assessing global cardiovascular risk, but there is evidence that risk is often assessed inaccurately.
Aims
To compare general practitioner-reported global cardiovascular risk in French primary care patients with estimates based on established risk-scoring systems, and to identify factors accounting for any mismatch between the analyses.
Methods
Data on patients aged greater or equal to 50 years seen during two 3-day periods were provided by 619 general practitioners. Physicians rated each patient’s cardiovascular risk as low, moderate or high, according to their perception; in addition, risk was assessed using the Framingham and Systematic coronary risk evaluation (SCORE) risk-scoring systems.
Results
A total of 13,446 patients aged greater or equal to 50 years were included. Of 11,241 patients with no previous history of cardiovascular disease, 47% were considered by their physicians to be at low risk of cardiovascular disease and 14% to be at high risk. In that population, 72% of patients rated as high risk according to the Framingham system and 77% rated as high risk according to SCORE system were incorrectly assessed by their physicians; similar results were observed in patient cohorts based on whether or not patients had received treatment for dyslipidaemia. Weighted kappa analysis showed poor agreement between physician risk assessment and both the Framingham and SCORE risk-scoring systems.
Conclusion
This study underlines the mismatch between GP-estimated cardiovascular risk and the risk assessed using scoring systems, especially for high-risk patients.
Résumé
Rationnel
Les recommandations actuelles concernant la prévention des maladies cardiovasculaires mettent en avant l’importance d’évaluer le risque cardiovasculaire global. Ce risque est cependant souvent estimé de manière insuffisante.
Objectifs
Cette étude a été menée dans le but de comparer le risque évalué par les médecins généralistes chez les patients de plus de 50 ans, au risque recalculé par les échelles de risque, et de déterminer quels facteurs étaient déterminants pour la discordance entre les deux analyses.
Méthodes
Six cent dix-neuf médecins généralistes ont inclus et récupéré les informations concernant tous les patients de 50 ans et plus consultant pendant deux périodes de trois jours consécutifs. Ils devaient estimer le risque cardiovasculaire de chaque patient entre bas, modéré, sévère ; ce risque était par la suite recalculé selon les échelles de Framingham et de Systematic Coronary Risk Evaluation (SCORE).
Résultats
Treize mille quatre cent quarante-six patients de 50 ans ou plus ont été inclus dans l’étude. Parmi les 11,241 patients sans antécédent personnel de maladie cardiovasculaire, 47 % ont été considérés par le médecin comme étant à bas risque cardiovasculaire et 14 % à haut risque. Dans cette population, le risque cardiovasculaire avait été sous-évalué par les médecins chez 72 % des patients évalués à haut risque selon l’échelle de Framingham et 77 % de ceux évalués à haut risque selon SCORE et des résultats similaires étaient obtenus selon que les patient recevaient ou non un traitement hypolipémiant. Le test de Kappa pondéré a montré qu’il existait une faible concordance entre l’évaluation du risque cardiovasculaire par les médecins généralistes et les scores de Framingham et de SCORE.
Conclusion
Cette étude met en avant la discordance entre le risque estimé par les médecins généralistes et le risque évalué par les échelles de risque, particulièrement chez les patients à haut risque cardiovasculaire.
Background
Current guidelines for the prevention of cardiovascular disease emphasize the importance of evaluating a patient’s overall (global) cardiovascular risk – a process that requires consideration of both modifiable risk factors, such as hypertension and dyslipidaemia, and non-modifiable risk factors such as age and sex . This recommendation is based upon extensive epidemiological evidence, which shows that many patients have multiple cardiovascular risk factors, and that these risk factors have additive or synergistic effects on the overall level of risk . The impact of such clustering of risk factors is highlighted by data from the Framingham study, which showed that the risk of coronary heart disease in patients with three or more risk factors was approximately 2.4 times higher in men, and almost six times higher in women, compared with that in patients with no risk factors .
A number of validated risk-scoring systems are available for the estimation of global cardiovascular risk, including the Framingham risk-scoring system and the Systematic coronary risk evaluation (SCORE) system developed by the European Society of Cardiology . However, there is evidence that, despite widespread acceptance of the concept of global cardiovascular risk, risk assessment is not performed routinely in clinical practice. For example, in a survey of 754 general practitioners (GPs) in five European countries (France, Germany, Italy, Sweden and the UK), 81% of respondents reported using the current guidelines but only 18% thought that the guidelines were being widely implemented . Furthermore, there is evidence that risk is often estimated inaccurately even when guidelines are followed. For example, in the CONTROLRISK study in Spain, cardiovascular risk was correctly estimated in only 48% of patients treated in primary care and in 55% of those treated by specialists . Similar results were obtained in a study in hypertensive primary care patients in the UK, in which global cardiovascular risk (assessed using Framingham-based tables) was correctly estimated in only 21% of patients and was under-estimated in 63% .
To our knowledge, no studies have evaluated cardiovascular risk assessment in French patients. Hence, the present study was undertaken to compare assessments of cardiovascular risk by French primary care physicians with the actual level of risk, as calculated using the Framingham and SCORE systems, and to identify potential factors associated with inaccurate risk estimation. This paper presents results from patients without a previous history of cardiovascular disease (i.e. candidates for primary prevention of cardiovascular disease).
Methods
The study was a survey of 619 French GPs. The physician sample was chosen from the online Kantar Health France panel of 4134 primary care physicians and was representative of primary care physicians. A total of 3885 GPs were sent an e-mail invitation to participate in the study. Participating physicians were asked to complete an online questionnaire for all patients aged 50 years and older who were seen during one of two 3-day periods (16–18 December 2009 and 11–13 January 2010). In addition, physicians were asked to estimate the patient’s level of cardiovascular risk on a three-point scale (low, medium, or high). Risk scores for each patient were also assessed according to the Framingham and SCORE (for countries with low cardiovascular risk) systems using published algorithms . Risk was analysed according to the Framingham system (low, ≤ 10%; intermediate, 10% to ≤ 20%; high, > 20%) and the SCORE system (low, < 2%; intermediate, 2–4%; high ≥ 5%).
To allow analysis according to the total number of consultations, each doctor received a weighting based on the geographical region in which they practised, and the number of consultations was weighted according to the relative number of GP consultations in France and the age and sex distribution of the French population. All data relating to the total number of consultations were adjusted based on this weighting.
The data analysis was essentially descriptive. Due to potential gain in cardiovascular risk control, two cohorts of patients were analysed: those who were receiving treatment for dyslipidaemia and those who were not. Analysis was performed using QUANTUM ® software, version 5.8. Student’s t test was used at a level of 95–99% (i.e. 1–5% risk of error). Bases with less than 30 individuals are not statistically significant for this test.
Results
The mean age (standard deviation [SD]) of the participating physicians was 49.7 (7.8) years (50.0 [7.8] years after adjustment for weighting); 83% were men. Approximately 16% were practising in the Paris area and similar proportions were practising in all other regions of France. Overall, participating GPs believed that the two 3-day study periods were typical of their usual practice workload.
Study population
The 619 participating physicians provided data for a total of 13,483 patients (average 22 patients per GP), of whom 37 were excluded because of abnormal findings. Patients excluded from the analysis were those for whom blood pressure, total cholesterol, triglyceride, high-density lipoprotein (HDL) and low-density lipoprotein (LDL) values were inconsistent. Of the remaining 13,446 patients, 6380 (47%) were aged 50–64 years and 7066 (53%) were aged 65 years or older. A total of 11,241 patients (84%) had no previous history of cardiovascular disease and were included in the analysis. Of these, 3359 (30%) were receiving treatment for dyslipidaemia with statins or other lipid-lowering drugs.
Table 1 summarizes the cardiovascular risk factors present in the overall study population, the primary prevention population, and the cohorts that were treated or untreated for dyslipidaemia. For patients that had had a carotid ultrasonography (prescribed by a GP or specialist), results are presented depending on the reported atherosclerosis severity (none, mild, obstructive). In the overall study population, the average number of risk factors per patient, as defined by the Agence Francaise de Sécurité Sanitaire des Produits de Santé, was 1.7; 45% of patients had elevated LDL-cholesterol, 29% had uncontrolled hypertension and 18% had diabetes ( Table 1 ). In the primary prevention population, the mean number of risk factors per patient was 1.5 and the incidence of individual risk factors was similar to that in the overall study population ( Table 1 ).
| Overall study population ( n = 13,446) | Primary prevention population | |||
|---|---|---|---|---|
| Total ( n = 11,241) | Not treated for dyslipidaemia ( n = 7882) | Treated for dyslipidaemia ( n = 3359) | ||
| Sex, men/women (%) | 48/52 | 45/55 | 43/57 | 51 a /49 |
| Mean age (years) | 66.7 | 65.9 | 65.2 | 67.8 a |
| Total cholesterol > 2.4 g/L (%) | 17 | 17 | 18 a | 15 |
| LDL-cholesterol ≥ 1.3 g/L (%) | 45 | 48 | 52 a | 41 |
| HDL-cholesterol ≥ 0.6 g/L (%) | 36 | 38 | 40 a | 33 |
| TG > 2g/L (%) | 10 | 10 | 8 | 12 a |
| Family history of cardiovascular disease (%) | 16 | 8 | 6 | 14 a |
| Smoker (%) | 14 | 13 | 13 | 15 a |
| Uncontrolled high blood pressure (%) | 29 | 28 | 25 | 36 a |
| Diabetes (%) | 18 | 15 | 9 | 29 a |
| Mild atherosclerosis on US examination (%) | 18 | 13 | 7 | 27 a |
| Obstructive carotid plaque on US examination (%) | 4 | 2 | 1 | 5 |
| Chronic inflammatory disease (%) | 14 | 13 | 13 | 14 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
