Study population
Data stem from two independent cross-sectional European surveys based on similar sampling methods and questionnaires. These surveys were coordinated by the same investigators at the Nivel Institute from Netherlands and carried out in Switzerland, respectively, in 1993 by the University of Saint-Gall and in 2012 by the University of Lausanne. The studies obtained the approval of the Swiss ethical review board.
The European study of tasks profiles of general practitioner (1993)
The 1993 data were collected through the Swiss participation in the European study “Task Profiles of General Practitioners” [11]. This study aimed to describe the range of services offered by general practitioners in European countries and their relationship to health care systems [12]. The project was funded by the European commission and involved 7233 responding physicians in thirty countries in Europe. The drawing of the sample and the organization of data collection through paper questionnaires was carried out at national level. The Swiss sample was drawn by a random sampling procedure from a national database of PCPs (the Swiss Medical Association, FMH, which covers around 95 % of the Swiss physicians), stratified by urban/rural area. A response rate of 50 % enabled to obtain a sample size of 200 Swiss PCPs. The representativeness of the sample was assessed comparing with national data on age and gender and was considered as good [13].
The quality and costs of primary care in Europe study (QUALICOPC, 2012)
The 2012 data were collected through the Swiss participation in the QUALICOPC study. This project aimed to analyze and compare how primary health care systems in 34 countries perform in terms of quality, costs and equity.
Surveys were held among PCPs in 31 European countries (EU 27 – except for France-, FYR Macedonia, Iceland, Norway, Switzerland, and Turkey) and 3 non-European countries (Australia, Canada, and New Zealand). In each country, a random nationally representative sample of around 220 physicians was drawn. Only one physician per practice or health centre was eligible to participate. In Switzerland, the participating physicians stemmed from a random sample of PCPs drawn from the two physicians’ associations in first care medicine (general medicine and general internal medicine) and stratified by canton, the SPAM network (response rate of 10 %) [14]. The representativeness in terms of gender, rural/urban implantation and age was cross-checked against national statistics and considered as satisfactory. Ethical approval was acquired in accordance with the legal requirements in each country. Details about the study protocol and questionnaire development have been published elsewhere [15, 16]. Data collection took place between January and June 2012 in Switzerland.
Data
In both surveys, PCPs were interviewed by self-administrated questionnaire sent by mail. The questionnaires were translated from the initial English master version by translators in the three national languages of Switzerland: German, French and Italian language.
For the present study, we only selected twenty-two questions that were strictly the same in both surveys (both formulations of questions and answers). These questions were related to two domains of health care indicators, according to Donabedian’s classification: structure indicators (“all the factors that affect the context in which care is delivered”) and process indicators (“all actions that make up healthcare”) [17]. Structure indicators were explored through socio-demographic features of PCPs in terms of sex, age, rural/urban practices location and language areas of Switzerland (German, French or Italian). The practices location was described by the PCP using five items: big city, suburbs, small town, semi-rural and rural area. The items were secondarily grouped in two categories: urban area (big city, suburbs, small town) and rural area (semi-rural and rural area). The age variable was dichotomized at the median value in the global sample. Questions related to output indicators were arranged in different areas:
General features
General practice characteristics were described through the following items: solo or group practice, unique activity as PCP or involvement in other paid activity (such as physicians in nursing homes, private companies, teaching activities…), salaried or self-employed activity and weekly workload as PCP (number of hours). The workload distinguished regular hours and after-hours (weekend days and nights).
The equipment and aims of use of computers and the regular meets with others health care professionals (kind of professionals and frequency) were explored.
Primary health care access
Care access was explored using different indicators: daily face-to-face patient contacts (number and duration), number of daily patient phone contacts with patients and weekly home visits. Finally, distance to other PCPs, outpatient clinics and hospital in terms of km was assessed (declared by PCPs).
Care continuity – collaboration
The frequency of meetings (more than once a month) with other health care professionals such as other PCPs, medical specialists, home care nurses and social workers was assessed.
Care activities
In this area, first, the role of PCP as first contact for different diseases was evaluated. Then, the PCPs involvement in the management of patients with different acute and chronic diagnoses was explored. For these two questions, the answer was initially assessed using a four-point scale “almost-always”, “usually”, “occasionally” and “seldom-never”. For the analysis, the two first items were secondarily grouped.
The practice of minor surgery and medical techniques was investigated. The answers, initially assessed using a four-point scale “almost-always”, “usually”, “occasionally” and “seldom-never” were presented as “seldom or never carried out” vs all other categories. By the same way, technical equipments of PCPs’ practices were evaluated.
Prevention
PCPs were asked about their involvement in domains of health prevention such as immunization of children, antenatal care and pediatric surveillance, blood pressure and cholesterol regular measures.
Analysis
For each indicator, we calculated frequencies, or means and medians, in 1993 and 2012 and compared them using Pearson’s Chi2 test for categorical variables or t-Student test for continue quantitative variables. Nevertheless, in order to take into account the difference of the sampling method between the two surveys (stratification by canton in 2012 and by urban/rural areas in1993), the evolution between the two periods was also tested using an adjusted model on urban/rural features (logistic or linear regression according to the dependant variable studied). Analyses were performed using STATA software.