Between April 2006 and June 2007 a cross-sectional study was conducted at eight randomly selected non-academic acute general public hospitals in the geographic area of Caserta and Naples (Italy). All 550 HCWs working in the EDs were selected.
The medical director and the ED head of each hospital received a letter with the description of the survey and requesting consent for the HCWs to participate. The medical directors addressed a letter to all HCWs explaining the enrollment and the purpose of the survey, assuring that response was completely voluntary, that information provided would be used solely for fulfilling the research aims, and a self-administered anonymous questionnaire accompanied by an envelope to facilitate its return. Consent to participate was implied by the return of the completed questionnaire.
The questionnaire comprised five categories of questions: (1) demographic and occupational characteristics; (2) knowledge about the risks of acquiring and/or transmitting certain HAIs for/to a patient and standard precautions for prevention; (3) attitudes toward precautionary guidelines and perception of the risk of acquiring HAI; (4) practice of standard precautions; and (5) from which sources they received up-to-date information about HAIs [see Appendix A]. Correct answers to each item were based on a review of the available literature as well as policies and guidelines [6, 13].
The content of the questionnaire was validated with interviews and discussions with other experts in the field, and it was modified where necessary. Final questionnaire content, comprehensibility, clarity, and format were developed and validated on input of a volunteer sample of 30 HCWs in a small pilot-test.
The study protocol as well as the questionnaire was approved by Ethical Committee of the Second University of Naples.
Statistical analysis
Multivariate analysis was carried out using stepwise logistic and linear regression techniques to establish whether the predictor variables were independently associated with the following outcomes of interest: knowledge about the risk for a HCW of acquiring both Hepatitis C (HCV) and Human Immunodeficiency Virus (HIV) infections from a patient (Model 1); knowledge that using standard precautions (gloves, mask, protective eyewear) and hands hygiene after removing gloves are HAIs control measures (Model 2); perceived risk of acquiring a HAI from a patient (Model 3); using often or always gloves when at direct contact with a patient and hands hygiene measures after removing gloves (Model 4). For purposes of analysis, the outcome variables originally consisting of multiple categories were reduced to two levels. In Model 1, HCWs were classified, according to questions B1a and B1c, as those who knew the risk for a HCW of acquiring both HCV and HIV infections from a patient versus all others; in Model 2, they were grouped, according to questions B4 and B6, as those who knew that using standard precautions (gloves, mask, protective eyewear) and hands hygiene after removing gloves are HAIs control measures versus all others; and in Model 4, HCWs were grouped, according to questions D3 and D5, to whether they often or always used gloves when at direct contact with a patient and hands hygiene measures after removing gloves versus all others. The following independent variables were included in all models: gender (male = 0, female = 1), age (continuous, in years), working category (physician = 0, nurse = 1), number of years in practice (continuous), number of patients seen in a workday (continuous), number of working hours in a week (continuous), knowledge about the risk for a HCW of transmitting HCV and HIV infections to a patient (no = 0, yes = 1), knowledge that HCV and HIV infections can be serious (no = 0, yes = 1), knowledge that HCWs' hands are vehicle for transmission of nosocomial pathogens (no = 0, yes = 1), educational courses and scientific journals as sources of information about HAIs (no = 0, yes = 1), and need of additional information about HAIs (no = 0, yes = 1). The following variables were also included: knowledge that the use of standard precautions is a HAIs control measure (no = 0, yes = 1) in Model 1; knowledge about the risk for a HCW of acquiring HCV and HIV infections from a patient (no = 0, yes = 1), and knowledge that invasive procedures are a risk factor for HAIs (no = 0, yes = 1) in Models 2-4; marital status (single/separated/divorced/widowed = 0, married = 1), number of other persons in the household (0 = 0, 1 = 1, 2 = 2, 3 = 3, 4 = 4, >4 = 5), and knowledge that using standard precautions (gloves, mask, protective eyewear) and hands hygiene after removing gloves are HAIs control measures (no = 0, yes = 1) in Model 3; knowledge that hands hygiene after removing gloves is a HAIs control measure (no = 0, yes = 1), positive attitude towards the use of guidelines for HAIs control practices (no = 0, yes = 1), positive attitude toward hands hygiene measures to reduce the risk among patients (no = 0, yes = 1), positive attitude toward hands hygiene measures to reduce the risk among HCWs (no = 0, yes = 1), and perceived risk of acquiring a HAI (continuous) in Model 4.
The primary analysis was univariate and the variables significantly associated with the outcomes of interest at p-value of 0.25 or less were included in the final models. Then, one stepwise multivariate linear regression model and three multivariate logistic regression models were constructed and the significance level for the variables to entry in the model was set at 0.2 and for removal at 0.4. In the logistic models the association between predictors and outcomes was measured by odds ratios (ORs) and their 95% confidence intervals (CIs). All tests were two-tailed and a p-value of 0.05 or less was defined as statistically significant. The data were analyzed using the statistical software Stata [14].