Health Search – Research Institute of the Italian Society of General Practice: the creation of a research database in general practice

OBJECTIVE: To describe a database used in general practice by analysing the characteristics of the physicians providing the data and the completeness of the data recording on the basis of the indicators of interest (smoking habits, weight, height, and the prevalence of hypertension). DESIGN: Descriptive study and multiple linear regression analysis of the relationships between structural variables and outcomes. SETTING: General practitioners (GPs) belonging to the Italian Society of General Medicine (SIMG), and enrolled in Health Search (HS). POPULATION: Six hundred and ninety-five voluntarily enrolled GPs and their patients. PRINCIPAL OUTCOMES: Descriptive statistics (mean and median values, standard deviation, frequencies) concerning the general characteristics of the GPs, the ways in which they connect to the network and their use of the clinical data management software, and epidemiological measures concerning the prevalence of hypertension among their patients, and the recording of weight, height and tobacco smoking habits. RESULTS: The geographic distribution of the GPs is homogeneous: 317 in Northern Italy (12.52 physicians per 10(6) patients), 134 in Central Italy (12.30 physicians per 10(6) patients) and 244 in Southern Italy and the Islands (11.89 physicians per 10(6) patients). Five hundred and ninety-five (85.6%) are males; their mean age is 46.7 years (SD +/- 3.8); the mean number of patients per physician is 1128; 69% have at least one post-graduate specialisation; 49% practise alone; and they work for a mean of 30 hours per week (SD +/- 13). Five hundred and fifty-three GPs were actually connected to the network as of January 2003. The data relating to the years 2000-2001 show an increase in the number of connected physicians (34 in 2000 and 261 in 2001) and in the number of connections (a median of 12 in 2000 and 17 in 2001). The GPs contact more than 80% of their patients every year. The frequency of the recording of data concerning smoking habits, weight and height increases in proportion with the frequency of connections. The prevalence of hypertension among the patients included in the survey is 5.4%. Multiple regression analysis showed that the variability in the prevalence of hypertension was not influenced by the frequency of PC use but by the different structure of the age classes of the patients. CONCLUSIONS: This study of the variables relating to GPs, their clinical practices, and their use of computerised records (examined by means of the chosen indicators) shows that the regular and complete recording of the principal data of health interest is feasible insofar as it is not biased by the characteristics of the GPs, and advantageous. The construction of the database therefore represents a first step towards the initiation of routine research into general practice in Italy.