Factorial Design for Improving Influenza Vaccination Among Employees of a Large Health System
Objective. As healthcare personnel (HCP) influenza vaccination becomes a quality indicator for healthcare facilities, effective interventions are needed. This study was designed to test a factorial design to improve HCP vaccination rates.
Design. A before‐after trial with education, publicity, and free and easily accessible influenza vaccines used a factorial design to determine the effect of mobile vaccination carts and incentives on vaccination rates of HCP, who were divided into groups on the basis of their level of patient contact (ie, business and/or administrative role, indirect patient contact, and direct patient contact).
Setting. Eleven acute care facilities in a large health system.
Participants. More than 26,000 nonphysician employees.
Results. Influenza vaccination rates increased significantly in most facilities and increased system‐wide from 32.4% to 39.6% (
). In the baseline year, business unit employee vaccination rates were significantly higher than among HCP with patient contact; rates did not differ significantly across groups in the intervention year. In logistic regression that accounted for demographic characteristics, intervention year, and other factors, the use of incentives and/or mobile carts that provided access to vaccine at the work unit significantly increased the likelihood of vaccination among HCP with direct and indirect patient contact, compared with control sites.
Conclusions. Interventions to improve vaccination rates are differentially effective among HCP with varying levels of patient contact. Mobile carts appear to remove access barriers, whereas incentives may motivate HCP to be vaccinated. Education and publicity may be sufficient for workers in business or administrative positions. Interventions tailored by worker type are likely to be most successful for improving HCP vaccination rates.
Received August 8, 2008; accepted February 15, 2009; electronically published June 2, 2009.
Annually, in the United States, approximately 200,000 influenza‐related hospitalizations occur,1 including frequent outbreaks among persons who have been hospitalized for other reasons.2‐7 Patient‐to‐patient and healthcare personnel (HCP)–related transmission are often to blame, given that infected individuals can spread virus before symptoms have appeared and the fact that HCP, especially house staff, continue to work despite having febrile respiratory illness.8,9 Despite a long‐standing recommendation from the Advisory Committee on Immunization Practices that HCP who have contact with high‐risk patients should receive influenza vaccine yearly, their vaccination rate in the United States remains low (41.7%).10 Previous research to determine the reasons for low HCP influenza vaccination rates found that barriers include lack of time, fear of adverse effects, lack of awareness about the vaccine, and disbelief in the vaccine’s efficacy or in the need to be vaccinated.11‐14 Motivators include the desire to protect oneself and one’s patients from influenza, to reduce sick leave, and because vaccination was recommended by the employee health department.11,12 A number of interventions aimed at eliminating the barriers and reinforcing motivators have been tested, resulting in HCP vaccination rates ranging from 24% to 78%.11‐16
The Task Force on Community Preventive Services has recommended combinations of interventions to increase demand for vaccinations and interventions to increase access in healthcare settings as part of a multiple‐component program.17 Most research has focused on HCP who provide direct care to patients. In any hospital, there are employees who do not provide direct patient care but who could easily transmit influenza to patients. Yet, little is known about which interventions are most applicable for hospital employees with varying degrees of patient contact.
For several years, the University of Pittsburgh Medical Center (UPMC), the largest health system in western Pennsylvania, has attempted to increase influenza vaccination among all of its 43,000 employees, regardless of their level of patient contact. After the 2005–2006 influenza season, a sample of HCP from 6 UPMC facilities were surveyed, and the findings revealed differences between vaccinated and unvaccinated HCP in beliefs about the risk of influenza disease, transmission, vaccine safety and efficacy, social support for vaccination, and attitudes (eg, HCP have a duty to be vaccinated).18 The most frequently cited reason to be vaccinated was to protect oneself, and the most important incentive was convenience. The most important reason for failure to be vaccinated was that it was not viewed as a priority. These findings were used to plan the 2006–2007 influenza season interventions.
The purpose of this study is to evaluate the effectiveness of multiple‐component interventions to promote uptake of influenza vaccination in this health system and to determine whether their effectiveness differs across groups of HCP with varying levels of patient contact.
Methods
Interventions and Study Design
Eleven UPMC facilities participated in this before‐after trial. In all facilities, the UPMC (1) provided influenza vaccine for administration to all employees free of charge; (2) produced posters and other promotional materials that addressed myths about influenza, the vaccine, and motivations for choosing to be vaccinated or not; (3) encouraged peer vaccination on clinical units; and (4) sent chief executive officers of each facility a letter containing the influenza vaccination rate for that facility for the previous season, to encourage improvement. Several biweekly issues of the system‐wide staff newsletter promoted the vaccine campaign and announced the times and places where employees could be vaccinated; one article featured a minority physician vaccine “champion” to address previously reported racial disparities in rates.
Two additional interventions—incentives and use of mobile vaccination carts—were implemented using a factorial (or quadrant) design (Table 1), such that 4 facilities had incentives only, 2 facilities had mobile vaccination carts only, 3 facilities had both incentives and carts, and 2 control sites had neither. Initially, there were 4 incentive sites at which vaccinees selected an envelope containing a thank you note, 10% of which contained a notice of winning a $10 grocery store gift card. Some sites established their own incentive program independent of the investigators. Specifically, at 2 sites, a lottery was held among vaccinees, with winners each receiving a paid day off, and 1 site offered a party to the unit with the highest influenza vaccination rate. These facilities were assigned to the appropriate incentive quadrants for analysis.
Mobile carts were staffed by contracted emergency medical technicians to provide convenient access to influenza vaccine. The carts visited each patient care unit and each non–patient care, nonbusiness unit at least once during all shifts, including weekends, during October and November.
This quality improvement project was approved by the Quality Assurance Committee of UPMC.
Data Collection
Influenza vaccination data from the 2005–2006 and 2006–2007 seasons were recorded on paper logs in each facility and were then manually entered into Occupational Health Manager software (PureSafety). Vaccination data were combined with information from the human resources database, which includes data on job title, location (facility and unit or department), race, and sex. Only employees who were employed during the influenza season (October 1 through December 31) were included in the database, such that employees would have been available for vaccination and exposure to the interventions.
Employees were stratified into 3 groups on the basis of the level of patient contact. HCP with “direct patient contact” were those whose jobs involved hands‐on or face‐to‐face contact with patients (eg, nurses, nursing aides, phlebotomists, therapists, transporters, and patient registration clerks). With the exception of staff physicians or hospitalists, physicians were excluded from the database, because it was assumed that many would likely have been vaccinated in their private medical offices and not in the hospital. HCP with “indirect patient contact” were those whose jobs entailed being in and around patient care areas but who did not necessarily work with patients (eg, environmental services employees, maintenance workers, and health unit coordinators). Business and/or administrative employees were those whose jobs involved no patient contact or time in patient care areas (eg, secretarial staff, laboratory technicians, medical records clerks, and upper management). Employees with a valid contraindication to influenza vaccine (eg, severe egg allergy) were ineligible for vaccine receipt and were not included in the study.
Statistical Analyses
Descriptive statistics for demographic characteristics and job categories for HCP in all participating hospitals were calculated. We used χ2 tests to compare (1) the differences in demographic characteristics, job categories, and vaccination rates between baseline and intervention years, (2) the influenza vaccination rates in baseline and intervention years in each participating site and overall, (3) the effects of intervention strategies overall and across job categories within each year, and (4) the differences in vaccination rates within each strategy between years. Finally, we performed separate logistic regression analyses for each job category that controlled for the age, race, and sex of HCP; the type of facility; the intervention year; and the type of intervention. To account for the sample cluster and varying size of each of the 11 participating facilities, SAS statistical software, version 9.1.3 (SAS Institute), was used to analyze the data, specifying that the primary study units were facilities and that HCP were nested within facilities. The procedure GENMOD, with influenza vaccination status as the binary outcome variable, was used to account for repeated measurement of the variables of interest for participating HCP in each hospital over 2 years. Statistical significance for all analyses was set at 
.
Results
Total sample sizes were 26,435 employees during 2005–2006 and 27,233 employees during 2006–2007. More than 75% of HCP were employed during both study years. As shown in Table 2, most employees were white, female, and less than 50 years of age, and approximately one‐half were involved in direct patient care. Table 3 shows the overall influenza vaccination rates for each season, by facility. Influenza vaccination rates increased significantly in 8 of 11 facilities and, overall, by 7.2 percentage points (from 32.4% to 39.6%; 
). In the baseline year, vaccination rates among groups of HCP in all hospitals combined differed significantly. Average vaccination rates for all hospitals were 31.1% for HCP with direct patient contact, 31.5% for HCP with indirect patient contact, and 35.3% for business and/or administrative workers (
). In the intervention year, vaccination rates among groups of HCP in all hospitals combined were 39.6% for HCP with direct patient contact, 38.4% for HCP with indirect patient contact, and 40.4% for business and/or administrative workers and no longer differed statistically (
).
The impact of the interventions differed among the 3 types of HCP (Table 4). Without considering any other demographic factors, vaccination rates of HCP with both direct and indirect patient contact significantly improved in facilities that used incentives and/or mobile carts but did not increase in the group that was exposed to educational programs and publicity alone. Vaccination rates of employees in business and/or administrative positions improved in facilities with either incentives plus publicity and education or publicity and education only; in fact, vaccination rates increased more in the latter sites.
Table 5 presents the results of separate logistic regression analyses for each job category that controlled for age, race, and sex of HCP; type of facility; intervention year; and type of intervention. For HCP with direct patient contact, older age, white race, community hospital setting, intervention year, and interventions that included incentives and/or mobile vaccination carts were associated with a significantly increased likelihood of influenza vaccination. For HCP with indirect patient contact, the factors associated with a higher likelihood of influenza vaccination were the same, except that, for this group, female sex was also associated with a significantly increased likelihood of influenza vaccination. For employees in business and/or administrative units, the factors older age, female sex, white race, community hospital setting, and intervention year were associated with increased likelihood of influenza vaccination, whereas incentives were not.
The likelihood of vaccination differed significantly by race; therefore, logistic regressions were performed separately for white and for nonwhite HCP (Table 6). For white HCP, outcomes were similar to the overall findings, except that, among workers with indirect patient contact, there was no difference in likelihood of vaccination between women and men. For nonwhite HCP, the sex difference reemerged in workers with indirect patient contact, and odds ratios were larger among nonwhite HCP in direct and indirect patient care job categories. Thus, we tested interactions between race and intervention variables. For indirect patient care workers, there were no significant interactions. However, for nonwhite HCP with direct patient care duties, the coefficients for carts with or without incentives—but not for incentives alone—were significantly greater than were those for white HCP with direct patient care duties.
Discussion
Most published studies from healthcare institutions that have reported successful influenza vaccination programs among HCP have primarily focused on nurses, physicians, and others who provide direct patient care.13,16 However, influenza can be spread to patients from employees with many other roles involving both direct and indirect patient care duties. Furthermore, studies that examine only overall vaccination rates may incorrectly estimate the effect of the intervention among individuals in different job categories.
The Task Force on Community Preventive Services has recommended multiple‐component programs that combine interventions to increase demand for vaccinations, interventions to increase access in health care settings, and interventions that are system based.17 With use of a variety of methods, including incentives and mobile carts, we found significant effects of the intervention on vaccination rates among all sites combined, with a higher likelihood of vaccination during the intervention year, while controlling for other factors (Table 5).
For both white and nonwhite HCP with direct or indirect patient contact, interventions that included mobile carts (with or without incentives) were superior to education and publicity alone for increasing vaccination rates, with larger effects for nonwhite HCP providing direct patient care. Given the racial disparity in influenza vaccination uptake, this combination of interventions may be a valuable strategy for increasing vaccination rates among nonwhite direct‐care HCP. As reported in our previous year’s survey,18 it appears that convenience is a key factor for vaccinating busy personnel with dedicated duties on a particular unit. Convenience may be especially important for large facilities in which leaving the unit to undergo vaccination is time‐consuming or for facilities in which the employee health office is only available for vaccination during the day shift. Although the itinerant nature of indirect patient care workers’ duties may allow them to more easily access influenza vaccination clinics or the employee health office, mobile vaccination carts may offer added convenience.
Business and/or administrative personnel typically have access to computers from which they can receive electronic publicity messages. Furthermore, they have no patient care responsibilities, thus affording them more flexibility to leave their workstations to receive influenza vaccine. These factors may account for their comparatively high baseline vaccination rates. However, influenza vaccination of these workers may have a limited effect on patient safety, because they may be located far from patient care areas. For this group, because intervention year was significant in multivariate analyses but incentives were not, health system–wide publicity efforts to raise rates were most likely responsible for the observed increase. On the basis of these findings, we suggest that interventions should be tailored to the specific job category to make efficient use of available resources.
Although influenza vaccination rates increased significantly in this large multiple‐facility health system, rates remained below ideal levels. Clearly, more work is needed to raise rates in this health system and elsewhere. National efforts include a call to action for vaccination of HCP,19 and HCP influenza vaccination has been dubbed the “next battleground” for patient safety.20‐22
What should be the next step? Some have proposed mandatory vaccination of HCP and/or declination statements.23 Clearly, mandatory vaccination is an option, with 1 health system reporting 98% vaccination following implementation of a policy that required HCP to receive influenza vaccination or to wear a mask during influenza season.19 Although several national organizations, including the Advisory Committee on Immunization Practices, the Hospital Infection Practices Advisory Committee, and the Joint Commission, have recommended that healthcare facilities institute HCP influenza vaccination programs, only 1 endorses mandatory vaccination.24
Signed declinations, although recommended, are also not commonly used,25,26 but they are strongly supported by infection control professionals. Where they are in use, signed declination forms are associated with significantly higher influenza vaccination rates.25,27,28 To our knowledge, neither of these strategies has been implemented in a health system as large as UPMC. Without mandatory vaccination or declination, we were able to incrementally increase vaccination rates among more than 26,000 employees, offering substantial benefit in the community, considering their many contacts.
Our results, along with other findings,17,25 support the use of a combination of strategies that include education, system efforts to increase convenience, and incentives. This study benefited from our ability to plan interventions and tailor publicity messages to address the motivators and concerns about vaccination based on the findings from a survey that followed the previous influenza season.18 We recommend using periodic surveys to identify and address employees’ concerns about influenza vaccination and tailoring the annual influenza vaccination program to different groups of HCP.
Strengths and Limitations
The strengths of this study include our use of a factorial design, a common electronic registry of immunizations, and statistical analysis that accounted for the clustered design and inclusion of 11 diverse facilities. This is the first study, to our knowledge, to have tested interventions in a factorial design in a multiple‐facility health system, and it is also the first, to our knowledge, to have examined the effect of interventions on influenza rates among HCP with 3 distinct job categories related to levels of patient exposure in acute care facilities.
It is possible that we missed some vaccinations that were given outside of the facilities, because we used an employee record database; thus, we would have underestimated the actual vaccination rates. The number of facilities in each intervention was limited by the number of facilities in the health system.
Conclusions
Because the effect of interventions to increase influenza vaccination rates among HCP differed according level of patient contact and race, multifaceted interventions in healthcare facilities should be tailored to HCP groups. The convenience of mobile vaccination carts for HCP with direct and indirect patient contact may remove time‐ and access‐related barriers, and incentives for HCP may encourage vaccine uptake.
Acknowledgments
Financial support. Centers for Disease Control and Prevention (IP000064‐02).
Potential conflicts of interest. M.P.N., R.K.Z., D.E.F., and C.J.L. have received partial salary support from Merck. M.P.N., R.K.Z., and C.J.L. have received partial salary support from MedImmune for unrelated research grants but have no conflicts of interest with regard to this project. All other authors report no conflicts of interest relevant to this article.
References
- 1. Thompson WW, Shay DK, Weintraub E, et al. Influenza‐associated hospitalizations in the United States. JAMA 2004; 292:1333–1340.
- 2. Cunney RJ, Bialachowski A, Thornley D, Smaill FM, Pennie RA. An outbreak of influenza A in a neonatal intensive care unit. Infect Control Hosp Epidemiol 2000; 21:449–454.
- 3. Weinstock DM, Eagan J, Malak SA, et al. Control of influenza A on a bone marrow transplant unit. Infect Control Hosp Epidemiol 2000; 21:730–732.
- 4. Salgado CD, Farr BM, Hall KK, Hayden FG. Influenza in the acute hospital setting. Lancet Infect Dis 2002; 2:145–155.
- 5. Pachucki CT, Pappas SA, Fuller GF, Krause SL, Lentino JR, Schaaff DM. Influenza A among hospital personnel and patients: implications for recognition, prevention, and control. Arch Intern Med 1989; 149:77–80.
- 6. Strausbaugh LJ, Bridges CB, Jernigan DB, Liedtke LA. Influenza: prevention and detection in acute care settings. Infect Med 2002; 19:310–317.
- 7. Bridges CB, Kuehnert MJ, Hall CB. Transmission of influenza: implications for control in health care settings. Clin Infect Dis 2003; 37:1094–1101.
- 8. Ohrt CK, McKinney WP. Achieving compliance with influenza immunization of medical house staff and students: a randomized controlled trial. JAMA 1992; 267:1377–1380.
- 9. Weingarten S, Riedinger M, Bolton LB, Miles P, Ault M. Barriers to influenza vaccine acceptance: a survey of physicians and nurses. Am J Infect Control 1989; 17:202–207.
- 10. Fiore AE, Shay DK, Haber P, et al. Prevention and control of influenza: recommendations of the Advisory Committee on Immunization Practices (ACIP) 2007. MMWR Morb Mortal Wkly Rep 2007; 56:1–54.
- 11. O’Reilly FW, Cran GW, Stevens AB. Factors affecting influenza vaccine uptake among health care workers. Occup Med (Lond) 2005; 55:474–479.
- 12. Canning HS, Phillips J, Allsup S. Health care worker beliefs about influenza vaccine and reasons for non‐vaccination—a cross‐sectional survey. J Clin Nurs 2005; 14:922–925.
- 13. Tapiainen T, Bar G, Schaad UB, Heininger U. Influenza vaccination among healthcare workers in a university children’s hospital. Infect Control Hosp Epidemiol 2005; 26:855–858.
- 14. Song JY, Park CW, Jeong HW, Cheong HJ, Kim WJ, Kim SR. Effect of a hospital campaign for influenza vaccination of healthcare workers. Infect Control Hosp Epidemiol 2006; 27:612–617.
- 15. Interventions to increase influenza vaccination of health‐care workers—California and Minnesota. MMWR Morb Mortal Wkly Rep 2005; 54:196–199.
- 16. de Juanes JR, Garcia de Codes A, Arrazola MP, Jaen F, Sanz MI, Gonzalez A. Influenza vaccination coverage among hospital personnel over three consecutive vaccination campaigns (2001–2002 to 2003–2004). Vaccine 2007; 25:201–204.
- 17. Centers for Disease Control and Prevention. Improving influenza, pneumococcal polysaccharide, and hepatitis B vaccination coverage among adults aged <65 years at high risk. MMWR Morb Mortal Wkly Rep 2005; 54:1–11.
- 18. Nowalk MP, Lin CJ, Zimmerman RK, et al. Self‐reported influenza vaccination rates among health care workers following intervention in a large health system. Am J Infect Control 2008; 36:574–581.
- 19. National Foundation for Infectious Diseases. Influenza immunization among health care workers. Bethesda MD, National Foundation for Infectious Diseases, 2008.
- 20. Pearson ML, Bridges CB, Harper SA. Influenza vaccination of health‐care personnel: recommendations of the Healthcare Infection Control Practices Advisory Committee (HICPAC) and the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep 2006; 55:1–16.
- 21. Joint Commission. Joint Commission establishes infection control standard to address influenza vaccines for staff. 2007. Available at: http://www.jointcommission.org/NewsRoom/NewsReleases/nr_06_13_06.htm. Accessed June 4, 2007.
- 22. Hoffmann CJ, Perl TM. The next battleground for patient safety: influenza immunization of healthcare workers. Infect Control Hosp Epidemiol 2005; 26:850–851.
- 23. Poland GA, Tosh P, Jacobson RM. Requiring influenza vaccination for health care workers: seven truths we must accept. Vaccine 2005; 23:2251–2255.
- 24. Lugo NR. Will carrots or sticks raise influenza immunization rates of health care personnel? Am J Infect Control 2007; 35:1–6.
- 25. Polgreen PM, Chen Y, Beekmann S, et al. Elements of influenza vaccination programs that predict higher vaccination rates: results of an emerging infections network survey. Clin Infect Dis 2008; 46:14–19.
- 26. Gazmararian JA, Coleman M, Prill M, et al. Influenza vaccination of health care workers: policies and practices of hospitals in a community setting. Am J Infect Control 2007; 35:441–447.
- 27. Borlaug G, Newman A, Pfister J, Davis JP. Factors that influenced rates of influenza vaccination among employees of Wisconsin acute care hospitals and nursing homes during the 2005–2006 influenza season. Infect Control Hosp Epidemiol 2007; 28:1398–1400.
- 28. Ribner BS, Hall C, Steinberg JP, et al. Use of a mandatory declination form in a program for influenza vaccination of healthcare workers. Infect Control Hosp Epidemiol 2008; 29:302–308.
-
The contents of this article are the responsibility of the authors and do not necessarily reflect the official views of the Centers for Disease Control and Prevention.