Determinants of influenza vaccination uptake among Italian healthcare workers

Abstract

We analyzed seasonal influenza vaccination coverage among the Italian healthcare workers (HCW) in order to identify socio-demographic and clinical determinants of vaccination.

We used data from the survey “Health and health care use in Italy,” which comprised interviews of 5,336 HCWs For each respondent, information on socioeconomic, health conditions, self-perceived health and smoking status were obtained. After bivariate analysis, we used multilevel regression models to assess determinants of immunization. Overall 20.8% of HCWs (95%CI 19.7–21.9) reported being vaccinated against seasonal influenza.

After controlling for potential confounders, multilevel regression revealed that older workers have a higher likelihood of vaccine uptake (OR = 6.07; 95% CI 4.72–7.79). Conversely, higher education was associated with lower vaccine uptake (OR = 0.65; 95% IC 0.50–0.83). Those suffering from diabetes (OR = 2.07; 95% CI 1.19–1.69), COPD (OR = 1.95; 95% CI 1.31–2.89) and cardiovascular diseases (OR = 1.48 95% CI 1.11–1.96) were more likely to be vaccinated. Likewise, smokers, or former smokers receive more frequently the vaccination (OR = 1.40; 95% CI 1.15–1.70; OR = 1.54; 95% CI 1.24–1.91, respectively) compared with never-smokers as well as those HCWs reporting fair or poor perceived health status (ORs of 1.68, 95% CI 1.30–2.18).

Vaccine coverage among HCWs in Italy remains low, especially among those with no comorbidities and being younger than 44 y old. This behavior not only raises questions regarding healthcare organization, infection control in healthcare settings and clinical costs, but also brings up ethical issues concerning physicians who seem not to be very concerned about the impact of the flu on themselves, as well as on their patients. Influenza vaccination campaigns will only be effective if HCWs understand their role in influenza transmission and prevention, and realize the importance of vaccination as a preventive measure

Introduction

Seasonal influenza outbreaks represent a major public health challenge. The virus causes 3–5 million cases of severe disease each year, and is responsible for up to 1 million deaths annually.^1,2 The most effective way to prevent infection by the influenza virus and its potentially severe complications, is immunization. Vaccination is safe and well-tolerated and can prevent up to 70–90% of influenza-specific illness in healthy adults.¹ The World Health Organization (WHO), US Centers for Disease Control and Prevention (CDC) and the immunization guidelines in most of developed countries, recommend influenza vaccination of Healthcare Workers (HCWs). These professionals may be at increased risk of contracting influenza because they are exposed both in the community and in the workplace.^3,4 It is quite difficult to quantify this risk, but research suggests it may be greater for HCWs than for the general population.⁵ Annual influenza attack rates range from 5% to 10% in adults, while rates of 11–59% have been reported in HCWs caring for patients with influenza.^6,7 Influenza outbreaks can result in staff absences, disrupt health services and increase healthcare costs.⁵ Moreover, influenza represents a threat to patients, especially for those already at high risk conditions such as the immunocompromised individuals or those undergoing chemotherapy. Studies show that HCWs unknowingly risk acting as vectors for the virus, thus serving as an important reservoir for patients under their care.^5,7Despite this, influenza vaccination coverage among HCWs, remain low worldwide.⁸

In Italy, several studies have been conducted to try to assess the coverage of influenza-vaccination among HCWs, but they are all limited to small settings,^9,10 while nationwide data are lacking. The aim of this study was to assess the prevalence of influenza vaccination coverage, and the risk factors for non-vaccination among a representative sample of Italian population working in healthcare facilities.

Results

Overall 20.8% (95% CI 19.7–21.9) of the professionals in the sample reported being vaccinated against seasonal influenza in the last 12 mo (Table 1). The highest immunization coverage were registered for males (25.8%, 95% CI 23.8–27.9), and older workers, with proportions reaching 57.6% in HCWs over 65 (95% CI 53.9–61.3). High vaccination coverage were also found for those suffering from chronic diseases, such as COPD (53.4%, 95% CI 43.3–59.3), diabetes (52.0%, 95% CI 44.6–59.3) and cardiovascular diseases (44.6%, 95% CI 39.0–50.3). Another important variable associated with vaccination was perceived poor health conditions (49.7% of those with perceived poor health conditions were vaccinated, 95% CI 42.5–56.9, and 51.4% of those who were in very poor health, 95% CI 35.1–67.5). On the other hand, the lowest immunization coverage were registered among females (18.3%, 95% CI 17.1–19.6), the population aged less than 44 y old (11.2%, 95% CI 10.0–12.5), and those reporting a good self-perceived health status (12.5% of adults - 95% CI 10.5–14.9). Vaccination levels among current smokers was not different from the general population of HCWs (20.5%, 95% IC 19.0–22.0), while former smokers showed a higher coverage level (27.3%, 95% IC 24.9–29.8).

Table 1. Distribution of socioeconomic characteristics of professionals.

		Vaccine coverage during the previous 12 mo (n = 5,336)
		Not-vaccinated	Vaccinated	Pa
		N°	%	CI 95%	N°	%	CI 95%
	Total sample	4,226	79.2	78.1 -80.3	1,110	20.8	19.7–21.9
Sex	Male	1,306	74.2	72.1–76.2	455	25.8	23.8–27.9	*
	Female	2,920	81.7	80.4–82.9	655	18.3	17.1–19.6
Age Group	< 44	2,174	88.8	87.5–90.0	274	11.2	10.0–12.5	*
(years)	45–64	1,765	79.9	78.1–81.5	445	20.1	18.5–21.8
	65+	287	42.3	38.6–46.1	391	57.7	53.9 -61.3
Marital status	Single	1,036	85.3	83.2–87.2	179	14.7	12.8–16.8	*
	Married	2,608	79.2	77.8–80.6	683	20.8	19.4–22.2
	Separated/divorced	399	80.8	77.1–84.1	95	19.2	15.9–22.9
	Widowed	183	54.5	49.1–59.7	153	45.5	40.3–50.9
Education	post-graduate degree	192	69.3	63.7–74.5	85	30.7	25.5–36.3	*
	more than 4 y university degree	726	76.5	73.7–79.1	223	23.5	20.9–26.3
	3 y university degree	340	82.9	79.0–86.3	70	17.1	13.7–20.9
	High	1,228	85.6	83.7–87.3	207	14.4	12.7–16.3
	Medium	1,351	81.3	79.4- - 83.1	311	18.7	16.9–20.6
	Low	389	64.4	60.5–68.1	215	35.6	31.8–39.5
Area of Residence	North-West	1,024	79.1	76.8–81.2	271	20.9	18.8–23.2
	North-East	1,024	79.7	77.4–81.8	261	20.3	18.2–22.6
	Center	766	78.0	75.3–80.5	216	22.0	19.5–24.7
	South	984	79.8	77.5–82.0	249	20.2	18.0–22.5
	The Islands	428	79.1	75.5–82.4	113	20.9	17.6–24.5
Cardiovascular disease	Absence	4,062	80.6	79.5–81.7	978	19.4	18.3–20.5	*
	Presence	164	55.4	49.7–61.0	132	44.6	39.0–50.3
COPD	Absence	4,158	80.1	79.0–81.2	1,032	19.9	18.8–21.0	*
	Presence	68	46.6	38.6–54.7	78	53.4	45.3–61.4
Diabetes	Absence	4141	80.3	79.2–81.3	1,018	19.7	18.7–20.8
	Presence	85	48.0	40.7–55.4	92	52.0	44.6–59.3
Smoking habits	Never smoker	1,080	85.1	83.1–87.0	189	14.9	13.0–16.9	*
	Former smoker	933	72.7	70.2–75.0	351	27.3	24.9–29.8
	Smoker	2,213	79.5	78.0–81.0	570	20.5	19.0–22.0
Self-perceived health status	Very good	753	87.5	85.1–89.5	108	12.5	10.5–14.9	*
	Good	2,177	83.6	82.2–85.0	426	16.4	15.0–17.8
	Fair	1,187	71.8	69.6–73.9	467	28.2	26.1–30.4
	Poor	92	50.3	43.0–57.5	91	49.7	42.5–56.9
	Very poor	17	48.6	32.4–64.9	18	51.4	35.1–67.5

I: Confidence Interval 95% point estimates; ^ap < 0.05, Chi-square test

In the multilevel regression analysis (Table 2), older workers had a higher likelihood of being vaccinated (OR = 6.07; 95% CI 4.72–7.79). Conversely, the lower educational attainments were associated with a lower rate of vaccination among HCWs (e.g., low education: OR = 0.57; 95%CI 0.39–0.81). Those suffering from chronic disease, such as diabetes (OR = 2.07; 95% CI 1.19–1.69), COPD (OR = 1.95; 95% CI 1.31–2.89) and cardiovascular diseases (OR = 1.48 95% CI 1.11–1.96) are more likely to be vaccinated. Likewise, smokers and former smokers, are more frequently vaccinated against influenza (OR = 1.40; 95% CI 1.15–1.70; OR = 1.54; 95% CI 1.24–1.91, respectively) compared with never-smokers. HCWs reporting fair or poor perceived health status, compared with those with poorer health conditions, have higher odds of vaccination with ORs equal to 1.68 (95% CI 1.30–2.18) and 2.58 (95% CI 1.70–3.92), respectively.

Table 2. Variables associated to HCW’s vaccine coverage during the previous 12 mo results of multivariate modeling.

		OR	95%CI	P
Total sample (n = 5,336)
Sex	Male	1
	Female	0.75	0.64–0.89	*
Age Group	< 44	1
(years)	45–64	1.55	1.29–1.86
	65+	6.07	4.72–7.79
Marital status	Single	1
	Married	1.07	0.87–1.30	*
	Separated/divorced	1.11	0.82–1.50	*
	Widowed	1.61	1.17–2.23	*
Education	post-graduate degree	1
	more than 4 y university degree	0.83	0.60–1.14
	3 y university degree	0.74	0.50–1.10
	High	0.54	0.39–0.75	*
	Medium	0.51	0.37–0.71	*
	Low	0.57	0.39–0.81	*
Area of Residence	North-West	1
	North-East	0.92	0.64–1.32
	Center	1.00	0.68–1.46
	South	1.00	0.70–1.41
	The Islands	1.18	0.75–1.86
Cardiovascular disease	Absence	1
	Presence	1.48	1.11–1.96	*
COPD	Absence	1
	Presence	1.95	1.31–2.89	*
Diabetes	Absence	1
	Presence	2.07	1.46–2.94	*
Smoking habits	Never smoker	1
	Former smoker	1.54	1.24–1.91	*
	Smoker	1.40	1.15–1.70	*
Self-perceived health status	Very good	1
	Good	1.24	0.97–1.55
	Fair	1.68	1.30–2.18	*
	Poor	2.58	1.70–3.92	*
	Very poor	1.90	0.87–4.12

^a p < 0.05

Comparing the estimates obtained from the random-effects models with those of logistic regression models, the inclusion of random effects (i.e., residing in a different region of the Country) improved the accuracy of the models by approximately 20.6% (95%CI 11.78–36.14).

Discussion

Italian HCWs appear not to be very compliant with recommendations of the WHO and the local Ministry of Health. In fact, the mean coverage stops at 20.8%, with rates reaching a low 11.2% in younger age classes. In an Italian multicenter study conducted by Placidi et al.¹⁰ for a five-year period, the vaccination coverage ranged between 0 and 29%. The median value was 16–17% in 2005, 13% in 2006, 11% in 2007 and again 16–17% in 2008 and 2009 (only against the H1N1 variety of the influenza virus), and changed on the basis of individual risk perception and information about efficacy and side effects of the vaccine. Moreover, the results are in line with a local analysis made in the Apulia region, in Southern Italy, where an audit of vaccination coverage among 302 HCWs revealed that 32.7% declared that they had received at least one dose of influenza vaccine in the 2007/2008 season (95% CI: 27.3–37.9).¹⁴

This situation did not change even after the spread of the 2009 H1N1 influenza virus pandemic, as detailed in the study by Placidi et al.¹⁰ The coverage of influenza vaccination among HCWs ranges widely. However, the most important factors associated with vaccination were higher education, the presence of chronic diseases, and, most importantly, older age.

Previous analyses of data from the “Health and Healthcare use” survey, showed that, among non HCWs, persons aged 65 and older, 62.6% (95% IC: 61.9–63.1) reported being vaccinated against influenza.¹⁵ Looking at older (65 y and over) healthcare workers, where coverage reaches only 57.7% of vaccination rates (95% IC: 53.9–61.3), it appears that the levels of influenza vaccination coverage are rather similar in the two populations and even lower in the HCWs subgroup. Likewise, when considering adults and older people population with chronic disease, 57.6% of HCWs affected by COPD reported being vaccinated against seasonal influenza, similar to the vaccination coverage among the general population with COPD.¹⁶ The estimates coming from multilevel regression predicting HCW vaccination are more accurate when considering the regional level as a random parameter. These results highlight the effects of the important federal reform that has taken place in the Italian national health care system. It will be of interest to compare these data with the results of the upcoming edition of the survey “Health and Healthcare use” that will take place at the end of 2012 and in the first six months of 2013. Furthermore, the general low acceptance of H1N1 vaccination was probably the result of a combination of public skepticism, distrust of the health authorities’ indications and misinformation raised by a growing anti-vaccination lobby, who actively contested both the need and the safety of the pandemic influenza vaccine across Europe.¹⁷ However, it is well known that vaccination of HCWs can reduce workplace absences, deliver economic benefits for healthcare systems and provide cost savings for healthcare organizations.^5,18 Similar evidence has been gathered in an Italian setting. In fact, in a study performed in an Italian-teaching hospital, HCW vaccination was cost saving, with the economic benefits outweighing the costs by a factor of 4.5. The researchers found that the prevalence of influenza-like illness was significantly higher in unvaccinated HCWs, resulting in the loss of 64% more working days than in the vaccinated group.¹⁸ Despite inconsistencies,¹⁹ there is evidence that healthcare workers could play an important role in transmitting infections to their patients,⁷ as witnessed by the recent outbreaks of infectious disease in healthcare settings.^20,21 HCWs could be pivotal in counseling patients and mothers regarding immunization, therefore they should be more sensitive and adherent themselves to the prevention of infection disease through active immunization.²²

A recent Cochrane review suggested that some HCWs remain unvaccinated because of the lack of perceive risk, doubt vaccine efficacy and concerns about side effects and recommended to carry out new high quality RCTs to test interventions to increase vaccination rates among HCWs . Unfortunately, based on our analyses, HCWs vaccination behavior reveals that influenza vaccine in this group could be perceived to be even less important than in the general Italian population. This behavior not only raises questions regarding healthcare organization, infection control in healthcare settings and clinical costs, but also brings up ethical issues concerning physicians which seem not to be very concerned about the impact of the influenza on themselves, as well as on their patients. We share Hofmann et al.²³ opinion that on the fact that influenza vaccination campaigns will only be effective in the long run if HCWs understand their role in influenza transmission and prevention, and will start taking responsibility of their behavior.

Limitations and strengths of the paper have to be considered when interpreting and generalizing its findings. On the one hand, among the limitations, we must consider that the ISTAT survey did not include a detailed description of individuals employment and position within the healthcare organizations, therefore it is impossible to differentiate those subjects working in administrative services and those in direct contact with patients. In addition, limitations of our study are related to the typical bias risks of cross-sectional studies, i.e., the possibility of a residual confounding, the limited validity and reliability of self-reported measures, and the impossibility of establishing causality links to explain the correlations found. On the other hand, also major strengths have to be considered. First of all this study draws on the data of a large national survey that is weighted to be representative of the population in general. Moreover, the use of this large data set decreases the likelihood that residual confounding may account for the observed correlations. Last but not least, the validity and reliability of both the dependent variable (self-reported vaccination²⁴) and of most of covariates (e.g., self-reported health²⁵) have been already established by the literature, thus confirming the overall consistency of study findings.

Methods

Data were drawn from the survey "Health and use of health care in Italy," a national cross-sectional survey that supplies information about perceived health status, disease symptoms, chronic disability conditions and other social determinants of health, conducted every five years by the Italian Institute of Statistics (ISTAT). The health survey is performed to monitor health care needs, and the use of healthcare services, mainly to guide healthcare policy.

Each participating subject completed a self-administered questionnaire, and had a face-to-face interview with ISTAT data collectors. The latest edition of this survey, conducted between December 2004 and September 2005, gathered data on 50,474 families and 128,040 individuals, representative of the Italian population in terms of age and gender.^11,12 The new edition of the survey, scheduled for 2010, has been postponed because the ISTAT was involved in the organization of the 2011 national census. The update is presently in progress (September–December 2012) and will continue in 2013 (March–June), as reported on the ISTAT website (http://www.istat.it/it/archivio/7740).

From the above sample, 5,336 persons, who declared to work in healthcare facilities, were considered. As a dependent variable, we used the reply (yes/no) to the question: "Did you get vaccinated against influenza in the last 12 months?." Moreover, the following socio-demographic characteristics were available for each participant: age, gender, marital status, area of residence and level of education. The geographic area of residence was classified by Italian region according to the ISTAT criteria (North-East, North-West, Center, South and Islands). Education was categorized as follows: Low (none or a primary school diploma), Medium (middle or junior high school diploma), High (high school diploma), Undergraduate (3-y university degree) and Graduate (bachelor or higher degree). The survey also investigated exposure to unhealthy lifestyles, such as smoking (current smoker, former smoker or never-smoker) and the presence of chronic disease such as diabetes, cardiovascular disease (hypertension, previous acute myocardial infarction, previous stroke or other cardiovascular disease) and chronic obstructive pulmonary disease (COPD). The SF-1 question assessing self-rated health was used as a proxy of general health status.¹³

Bivariate analyses were performed to study the distribution of variables in the sample using chi-square tests, as appropriate. Mixed effect models for binary data were developed to adjust for confounders, and to evaluate which factors were independently associated with influenza-vaccination (1, when vaccinated; 0 when not vaccinated), taking into account the role of different Regions of residence and the sampling design. Independent variables were grouped as socio-economic and clinical, and coded as follows: a) among the socio-demographic variables we have considered age class (less than 44 y old; between 45–64 y old; more than 65 y old), area of residence (North-East, North-West, Central, South and the Islands), educational level (medium-low and medium-high), b) clinical variables considered were, diabetes, cardiovascular disease and COPD . A single-item question was used to assess self-perceived health status, according to the following classification: Excellent; Very good; Good; Fair and Poor. The significance level for variables to enter the multiple logistic regression models was set at < 0.2 and for removing them from the model the level was set at > 0.4. Standard post-estimation tests (F-statistics, Pseudo R2 and Hosmer-Lemeshow tests) were used to assess the model's validity. The accuracy of the logistic model to predict the variables associated with HCWs' vaccination rate was assessed by receiver operating characteristic (ROC) analysis using the DeLong nonparametric method. We used sampling weights in all of the analyses, to reflect the multistage sampling design of the survey.¹¹

Analyses were performed with STATA, version 10.1 (Stata Corp., College Station, TX, 2007). The level of significance was set at 0.05.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.