ABSTRACT. Context. Immunization has an essential impact on public health worldwide. Numerous studies have shown the efficacy of different vaccines to protect individuals from various diseases. However, some parents choose not to vaccinate their children for reasons such as, among others, doubts regarding their usefulness, concerns over safety or efficacy, etc. Physicians are known to exert a direct influence on immunization rates by answering questions and clarifying misconceptions. Yet, it is unknown how they immunize their own children.Objective. We sought to assess how physicians interested in vaccination issues immunized, or would immunize, their own children.Design, Setting, and Participants. An 11-question, Webbased survey with a total of 102 discrete answers was sent to 2070 Swiss physicians in October 2004. All physicians were subscribers to a nonprofit, Web-based expert network (InfoVac, www.infovac.ch) that distributes monthly newsletters and answers question within 2 days on immunization issues. The InfoVac network reaches >95% of pediatricians in Switzerland but <20% of general practitioners. All responses were anonymous, and no identifier could be used to trace the participants of the survey. Questions were divided into 2 parts: (1) physicians who were parents were asked which vaccines they gave to their own children and at what age, and (2) all physicians were asked which vaccines they would give to their own child and at what age if they had a newborn child in 2004. Vaccines available in Switzerland at the time of the survey were offered as possible replies, and recommended vaccines were considered as those noted in the Swiss federal immunization schedule issued yearly. One question compared their immunization practice between their own children and their patients. Sociodemographics, qualifying year, membership in different professional groups, and their type of practice were also requested.Statistics. Standard descriptive statistics were used for sociodemographic characteristics. Univariate statistical analyses were performed for each variable to determine its relationship to the dependent variable, being a pediatrician or nonpediatrician. Logistic-regression analysis was used to calculate the adjusted odds ratios (ORs) and 95% confidence intervals (CIs), controlling for any statistically significant demographic variables that might function as confounders (gender, parenthood, workplace, year of diploma, and type of practice). For all statistical tests, differences were considered significant at P < .05.Main Outcome Measure. We performed a comparison of past and projected immunization rates in the children of pediatricians and nonpediatricians.Results. One thousand seventeen valid questionnaires were received (response rate: 49.1%; pediatricians: 53.3%). Nine hundred fifteen physicians (90%) had >1 child. All physicians reported immunizing children in their practice. Pediatricians were more likely to be women and to work in private practice than nonpediatricians but less likely to belong to a sel...