Using a cross-sectional study design, we surveyed employees at randomly selected child care centers located in the county and licensed by the state of Ohio. The anonymous questionnaire covered personal and work characteristics, pertinent medical history, and receipt of or intention to receive the pH1N1 and seasonal influenza vaccines. Demographic questions from the Behavioral Risk Factor Surveillance System Survey Questionnaire [18] and influenza vaccine questions from the National 2009 H1N1 Flu Survey Questionnaire [19] were used.

We used the Theory of Planned Behavior [20, 21], a widely applied theory in predicting social and health behavior, in developing some questions for our survey. The theory states that a person’s attitude (positive or negative feelings towards a behavior), perception of subjective norms (the perception that there is social pressure to perform or not perform the behavior), and perceived behavioral control (the perception of choice and availability of resources necessary to perform or not perform the behavior) influence the person’s intention to perform the behavior [20].

We created questions drawing from the three domains of the Theory of Planned Behavior to assess knowledge about and attitudes towards each vaccine [22]. Knowledge and attitudes questions were examined by extent of agreement with statements about each vaccine, using a four-point Likert scale (i.e., disagree, somewhat disagree, somewhat agree, and agree). Additional attitudes questions were examined using a four-point scale with bipolar adjectives (e.g., very good, somewhat good, somewhat bad, very bad).

The study population consisted of all employees ≥18 years old working at randomly selected licensed child care centers in the county. As of January 2010, the county had 362 licensed child care centers. Of these centers, 135 served infants <18 months. Since caregivers of children <6 months were among the initial target groups by the ACIP to receive the pH1N1 vaccine [13], 38 centers were selected randomly from this list of 135 centers. We visited all participating child care centers from January 30–March 1, 2010. All part-time and full-time employees working at the facility on the date of our visit to the center were invited to participate.

As a public health response, per the guidelines of US Title 45 Code of Federal Regulations Part 46, this evaluation was determined to not require review by an institutional review board.

Information about center characteristics, including center type and number of employees, was collected from center directors. Additional information, including National Association for the Education of Young Children (NAEYC) accreditation, and child capacity was obtained from the Ohio Department of Jobs and Family Services website [23].

Survey responses using a Likert scale were categorized as “expressed agreement” if respondents marked “agree” or “somewhat agree” and as “expressed disagreement” if respondents marked “disagree” or “somewhat disagree.” Internal consistency for the attitudes, subjective norms, and perceived behavioral control variables was measured using Cronbach’s coefficient (α) after adjusting for directionality. We created composite scores for variables within the attitudes, subjective norms, and perceived behavioral control domains where α > 0.7 by calculating the mean of the individual scores for each respondent. All composite scores were based on a four-point scale. A higher score on the composite score represented more favorable attitudes about the vaccines.

Characteristics of child care center employees who reported receipt of each vaccine were compared to those who denied receipt of the respective vaccine. Responses to the knowledge and attitudes questions were also compared among each group. We conducted bivariate analyses using the Student’s t test, χ² test, or Fisher’s exact test. We also used logistic regression for the bivariate analyses of the composite scores for the attitudes, subjective norms, and perceived behavioral control domains. Bivariate analyses were conducted using SPSS (SPSS Inc., Chicago, Illinois). All tests were two-tailed, and statistical significance was set at P < 0.05. We calculated odds ratio and 95% confidence intervals. We then used a stepwise backward elimination multiple logistic regression model to identify factors independently associated with receipt of each vaccine. The model was then analyzed using a Generalized Estimating Equations (GEE) model to account for the random effect of center with SAS 9.2 (SAS Institute, Cary, North Carolina).

Thirty-two (84%) of 38 invited licensed child care centers agreed to participate in the evaluation. The six declining centers cited lack of interest and/or lack of time as reasons for not participating. Participating and declining centers had similar characteristics including center type and number of employees. Most participating centers were independent or religiously-affiliated centers (75%) versus chain or corporate- or university-affiliated centers (25%). Four (12%) centers were NAEYC accredited. Fifteen (47%) centers had capacity for ≥100 children, and 16 (50%) centers had ≥15 employees.

We visited all 32 participating child care centers, and 384 (95%) of 403 employees ≥18 years old working on the day of the visits completed a survey. The median age of respondents was 30 years (range, 18–81 years), and 105 (27%) were ≤24 years old. Demographic and work characteristics of survey respondents are shown in Table 1.

Of 384 respondents, 45 (12%) reported having asthma or another chronic lung disease, 16 (4%) reported having heart disease, and 14 (4%) reported having diabetes. In total, 315 (82%) respondents denied having an underlying medical condition that would put them at high risk for influenza-related complications. These conditions also included kidney disease, liver disease, current diagnosis of cancer, any immunosuppressive disease, and taking immunosuppressive therapy.

Forty-five (12%) respondents reported having received the pH1N1 vaccine between October 2009 and survey administration. Rates of pH1N1 vaccination among respondents ranged from 0 to 83% by child care center. Thirteen (41%) of the 32 centers had 0% pH1N1 vaccination rates among their staff.

The most common places where respondents received the pH1N1 vaccine were a doctor’s office (35%), work-place (24%), clinic or health center (9%), health department (9%), and school (9%). pH1N1 vaccination rates were 14% for respondents caring for young infants (0–5 months), 13% for those caring for older infants (6–12 months old), 16% for those caring for toddlers (13 months–3 years), and 11% for those caring for children ≥4 years old.

Eighty-five (22%) respondents reported having received the 2009–2010 seasonal influenza vaccine between August 2009 and survey administration. Rates of seasonal influenza vaccination among respondents ranged from 0 to 57% by child care center. Four (12%) of the 32 centers had seasonal influenza vaccination rates of 0%.

The most common places where respondents received the seasonal influenza vaccine were a doctor’s office (45%), pharmacy or drug store (27%), clinic or health center (13%), hospital (5%), and workplace (5%). Seasonal influenza vaccination rates were 14% for respondents caring for young infants, 18% for those caring for older infants, 23% for those caring for toddlers, and 21% for those caring for children ≥4 years old.

Twenty-five (7%) respondents reported receiving both the pH1N1 and the seasonal influenza vaccines, 20 (5%) respondents reported receiving the pH1N1 but not the seasonal influenza vaccine, and 60 (16%) respondents reported receiving the seasonal influenza but not the pH1N1 vaccine. A total of 278 (72%) respondents reported receiving neither influenza vaccine.

The most common main reason cited by respondents for receiving the pH1N1 or seasonal influenza vaccines was “to protect myself/my family” (54% for pH1N1 and 65% for seasonal influenza). The next most common reasons were “my doctor recommended that I receive the vaccine” (16% for pH1N1 vaccine and 20% for seasonal influenza) and “to protect the children I care for” (14% for pH1N1 vaccine and 6% for seasonal influenza).

Of the 339 respondents who had not received the pH1N1 influenza vaccine, 65 (19%) reported they would definitely (n = 15) or probably (n = 50) get one. In contrast, 274 (81%) reported they would definitely not (n = 127) or probably not (n = 147) get one. Of the 298 respondents who had not received the seasonal influenza vaccine, 56 (19%) reported they would definitely (n = 13) or probably (n = 43) get one. In contrast, 242 (81%) reported they would definitely not (n = 127) or probably not (n = 115) get one. One respondent did not answer this question. Main reasons cited by respondents for not intending to receive either vaccine are shown in Table 2.

Most respondents had positive attitudes towards both vaccines, as most believed the vaccines to be “beneficial” (69% for pH1N1 and 82% for seasonal influenza) versus “harmful,” “good” (71% for pH1N1 and 80% for seasonal influenza) versus “bad,” and “wise” (68% for pH1N1 and 79% for seasonal influenza) versus “unwise.” These three measures of attitudes towards each vaccine had a high Cronbach’s coefficient; α = 0.879 for pH1N1 and α = 0.865 for seasonal influenza. Thus, for subsequent analyses, we created one positive attitudes composite score for each vaccine by calculating the mean of the scores for the three items.

Respondents’ beliefs about the pH1N1 and seasonal influenza vaccines are shown in Table 3. Respondents’ agreement with subjective norm and perceived behavioral control statements about both vaccines is also shown in Table 3. The seven subjective norms items (Table 3) regarding each vaccine had a high Cronbach’s coefficient; α = 0.835 for pH1N1 and a = 0.839 for seasonal influenza. We created one subjective norms composite score for each vaccine by calculating the mean of the scores for the seven items.

The latter three perceived behavioral control items (Table 3) regarding each vaccine had a high Cronbach’s coefficient; α = 0.733 for pH1N1 and α = 0.784 for seasonal influenza. We created one perceived behavioral control composite score for each vaccine by calculating the mean of the scores for the three items.

We found no statistically significant associations between age and mean number of years worked in child care or at the current center with receipt of either vaccine. Other demographic and work characteristics and their association with receipt of either vaccine are shown in Table 4. Agreement with belief statements about each vaccine and their association with receipt of the corresponding vaccine are also shown in Table 4.

Respondents with higher positive attitudes composite scores for each vaccine, or those who had more positive attitudes towards the vaccines, were more likely to have received that vaccine (OR = 9.52, 95% CI = 4.99, 18.16 for pH1N1 and OR = 5.87, 95% CI = 3.57, 9.65 for seasonal influenza). Respondents with higher subjective norms composite scores for each vaccine, or those who felt external pressure from others to receive the vaccine, were more likely to have received that vaccine (OR = 4.93, 95% CI = 3.10, 7.86 for pH1N1 and OR = 4.05, 95% CI = 2.80, 5.86 for seasonal influenza). Respondents with higher perceived behavioral control composite scores for each vaccine, or those who felt personal control over whether or not to get the vaccine, were more likely to have received that vaccine (OR = 1.52, 95% CI = 1.02, 2.27 for pH1N1 and OR = 1.42, 95% CI = 1.06, 1.90 for seasonal influenza).

Variables with P < 0.05 associated with receipt of the pH1N1 vaccine and those associated with receipt of the seasonal influenza vaccine were selected and entered into a stepwise backward elimination multiple logistic regression model and then a GEE model. Factors independently associated with receipt of the pH1N1 vaccine included having some college or technical school or higher (OR = 3.87, 95% CI = 1.69, 8.84), caring for toddlers (OR = 3.25, 95% CI = 1.15, 9.19), having positive attitudes towards the vaccine (OR = 5.23, CI = 3.32, 8.22), feeling external pressure to get it (OR = 3.10, 95% CI = 2.01, 4.77), and feeling personal control over whether or not to get it (OR = 1.68, 95% CI = 1.03, 2.74).

Factors independently associated with receipt of the seasonal influenza vaccine included believing in the efficacy of that vaccine (OR = 3.16, 95% CI = 1.44, 6.92), having positive attitudes towards it (OR = 2.83, 95% CI = 1.46, 5.48), feeling external pressure to get it (OR = 3.34, 95% CI = 2.04, 4.84), and feeling personal control over whether or not to get it (OR = 2.07, 95% CI = 1.06, 4.04). Factors independently associated with not receiving the seasonal influenza vaccine were caring for young infants (OR = 0.36, 95% CI = 0.16, 0.83) and having children 6 months–17 years old in the household (OR = 0.43, CI = 0.25, 0.74).