We conducted an epidemiological study of the prevalence of influenza vaccination coverage in patients aged ≥ 65 and influenza vaccination of primary care physicians according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement (guidelines for reporting observational studies). Of a total of 1791 primary care centers, we randomly selected 253 in seven Spanish regions (Andalusia, Castile-Leon, Catalonia, Valencia, Madrid, Navarre and Basque Country) in the 2011–12 season. A questionnaire was administered anonymously to physicians working in these centers between March 1 and May 25, 2012, via the internet, to obtain information on the main independent variable (influenza vaccination of primary care physician) and other variables (age, gender, education, opinions and attitudes of primary care physicians to influenza vaccination).
The questionnaire was developed after reviewing the scientific literature on the subject, especially the questionnaire used in the study by Kraut et al.. The questions were adapted to the specific circumstances of the Spanish National Health System and two pilot tests were conducted among medical staff in the researchers' settings to validate understanding of the questionnaire and its length. The final questionnaire consisted of 23 questions (22 closed and one open). Questions were distributed in three sections: information on the history of chronic disease and vaccination, knowledge of influenza and the influenza vaccine, and sociodemographic information. The questions were structured to appear gradually, spread over a total of six screens. The first screen welcomed the participants and provided general information on the survey. The following four screens contained the survey questions, and the last screen contained a text thanking the participants. Using the tools provided by the web platform, 19 of the 23 questions were compulsory, i.e., they had to be answered in order to access the following question.
The target population was any physician providing direct patient care in primary healthcare centers. In these centers, influenza vaccination is administered without cost by nurses to all population groups for which it is indicated, including people aged ≥65 years, according to physician prescription. Participating centers were randomly selected from a list of the centers in each region. All physicians in each center who had an email address were initially included. The questionnaire was accessible for a month and an email reminder was sent every 10 days to physicians who had not accessed the questionnaire or had not completed the survey.
The following variables were collected: profession, age, and sex. We also collected the presence of contraindications to influenza vaccination in each HCW, influenza vaccination in the 2011–2012 season and information on physicians' knowledge of and opinions and attitudes to influenza and influenza vaccination. Variables related to knowledge of and attitudes to influenza vaccination were covered by a set of questions evaluated on a Likert scale with 5 categories: totally agree, agree quite a lot, neither agree nor disagree, disagree quite a lot, and totally disagree. Information on the dependent variable (vaccination coverage in patients aged ≥ 65 years treated by physicians who participated in the study) was obtained from regional primary care records and was included as study information associated with each individual physician survey.
The data analysis included physicians providing direct patient care in primary care centers who reported information about their vaccination in the 2011–2012 season and in whom we could recover vaccination uptake in their patients aged ≥ 65 years. Physicians in whom vaccination was contraindicated were excluded.
The researchers responsible for each region facilitated listings of the HCW from each center containing the e mail address patients ascribed to each physician and vaccination coverage in their patients aged ≥ 65 years. The Coordinator Center received all lists and formulated a new list which assigned a number corresponding to each participant. This numbering was used to anonymize listings. The new listing was loaded on the web platform. At the end survey period, a database of completed surveys was extracted and was cross-checked with the anonymized list to identify non-responders. To characterize non-answers, we obtained information from 49.1% of physicians who did not respond to the survey. Physicians who responded to the survey were compared to those who did not according to physicians' age and sex and vaccination coverage in their patients aged ≥ 65 years.
The answers to questions about knowledge and attitudes were dichotomized in two categories: positive (totally agree, agree quite a lot) and negative (neither agree nor disagree, disagree quite a lot, and totally disagree).
A bivariate comparison using the Chi-square test was made in vaccinated/unvaccinated physicians considering the different sociodemographic variables and the answers to questions about knowledge and attitudes. Vaccination coverage in patients aged ≥ 65 years was compared with the main independent study variables using the Student's t test.
The association between vaccination coverage in patients aged ≥ 65 years and the main independent variable (vaccination of their physicians) was determined using a multilevel regression model with input of variables with a significance of p <0.10. All statistical tests were two-tailed and the α error accepted was 0.05. The analysis was performed using SPSS version 18 (SPSS Inc., Chicago, IL).
All information collected was treated in strict observance of legislation on observational studies. The study protocol was approved by the Ethics and Clinical Research Committee of the Jordi Gol Institute for Research in Primary Care.