The American Journal of Managed Care August 2010
Employees' Willingness to Pay to Prevent Influenza
The willingness of employees to pay to prevent influenza demonstrates a strong preference to protect themselves and their household members; however, modifiable barriers to vaccination persist.
Objectives: To quantify employees' preferences, as measured by willingness to pay, to prevent influenza in themselves and in their child and adult household members and to examine factors associated with willingness to pay.
Study Design: Prospective observational cohort study of a convenience sample of employees from 3 large US employers. Participants had at least 1 child (<17 years) living in their household for at least 4 days per week.
Methods: Each month from November 2007 to April 2008, employees completed Web-based surveys regarding acute respiratory illness in their household. In the final survey, employees were presented with descriptions of influenza and questions regarding their willingness to pay to prevent influenza. Factors associated with willingness to pay were examined using multivariate ordinary least squares regression analysis of the log of willingness to pay.
Results: Among 2006 employees, 31.3% were female, the mean age was 41.7 years, 85.3% were of white race/ethnicity, and the mean household size was 4.0. Employees' median (mean) willingness to pay to prevent influenza was $25 ($72) for themselves, $25 ($82) for their adult household members, and $50 ($142) (P <.01) for children. However, influenza vaccination rates were approximately equal for children (27.5%), employees (31.5%), and other adult household members (24.5%). This finding may be explained by barriers such as cost, dislike of vaccinations, and disagreement with national influenza vaccination recommendations, which were significantly associated with lower willingness to pay for prevention of influenza (P <.05).
Conclusion: Employees expressed a stronger preference to prevent influenza in their children than in themselves or other household members; however, modifiable barriers depress vaccination rates.
(Am J Manag Care. 2010;16(8):e205-e214)
These findings are the first willingness-to-pay estimates of employees’ preferences for influenza prevention among child and adult household members.
- Employees were willing to pay twice as much to prevent influenza among their children than among themselves or other adult household members; however, influenza vaccination rates among children were similar to those among employees and other adult household members. This suggests that extending employer-based influenza clinics to dependents may improve vaccination coverage.
- Notable barriers to vaccination include disagreement with national influenza vaccination recommendations and dislike of vaccinations.
- Targeted educational interventions regarding vaccination benefits may be necessary to improve vaccination coverage.
Yearly seasonal influenza vaccination is recommended by the Centers for Disease Control and Prevention (CDC) as the first and most important step in protecting against contracting and spreading seasonal influenza, as supported by studies.6-12 In the United States, target populations for vaccination include children aged 6 months through 18 years, individuals with underlying medical conditions, pregnant women, adults 50 years and older, and healthcare workers and household contacts of individuals at high risk for influenza complications.13 However, influenza vaccination coverage for the general population and for target groups remains suboptimal.13 Several studies14-16 have examined the role of various factors contributing to vaccination coverage, and findings suggest that facilitators to vaccination include previous vaccination and provider recommendation, while barriers to vaccination include cost, consideration of adverse effects, and fear of contracting influenza from the vaccine.
Although costs are a barrier to vaccination, there is little research regarding how much individuals are generally willing to pay for influenza prevention and particularly for prevention in household members. Existing estimates of willingness to pay for influenza prevention focus on healthcare workers’ or adults’ willingness to pay for influenza prevention in hypothetical children.17,18 The objectives of this study were to quantify employees’ preferences, as measured by willingness to pay, to prevent influenza in themselves and in their child and adult household members and to examine factors associated with willingness to pay.
Data and Study Design
Study data were collected as part of the Child and Household Influenza-Illness and Employee Function (CHIEF) study. The CHIEF study was a prospective observational cohort study conducted from November 2007 to April 2008. Participants were employees of 3 large US employers, including a national retail chain, a transportation company, and a durable goods manufacturing company. Further identifying information for employers is restricted because of study confidentiality agreements.
The CHIEF study enrollment goal was 2400 employees. In October 2007, approximately 36,000 employees were mailed a letter offering a small monetary remuneration to participate in the surveys (also offered in Spanish and in paper form) and were provided with a Web address (URL) to a Web-based screener survey. The durable goods manufacturing company also disseminated information about the study through internal efforts. Therefore, the actual number of individuals who were aware of the study is unknown. Eligible employees were required to be covered under their employer’s private health insurance plan and to have at least 1 child (<17 years) who was covered under the same health insurance plan and lived in the employee’s household for at least 4 days per week. This study was approved by the New England Institutional Review Board, and informed consent was obtained from all employees.
The study included a Web-based baseline survey and 6 Web-based monthly surveys. The baseline survey contained questions about household sociodemographics, health behaviors, comorbidities, and employee workplace characteristics. In each monthly survey and for each household member, employees reported the presence of acute respiratory illness and the specific symptoms comprised therein during the previous month. For all surveys, employees responded to all questions regarding themselves and each of their household members.
Outcome data used for this analysis came from the final survey. In the final survey, employees were presented with descriptions of influenza based on the actual duration and severity of influenza-like illness symptom constellations that were observed in the study cohort during the prior survey months.19 Separate descriptions of influenza were presented for the employee, his or her child household members, and adult household members. Employees’ preferences to prevent the described influenza for (1) themselves, (2) their child household members, and (3) their adult household members were measured using a previously fielded willingness-to-pay method that included dichotomous-choice double-bounded questions, followed by an open-ended question asking for the respondent’s maximum willingness to pay; this maximum value was used for all analyses (Table 1), and the valuation technique is discussed in more detail by Prosser et al.17 Personal and household member influenza vaccination status, reasons for vaccination among the vaccinated, and reasons for nonvaccination among the nonvaccinated were also captured in the final monthly study survey. Only employees who responded to every monthly survey and had nonmissing willingness-to-pay responses were included in this study.
Factors associated with willingness to pay were examined using multivariate ordinary least squares regression analysis with the log of willingness to pay as the dependent variable. Log transforming the willingness-to-pay variable (which was skewed) allowed for approximation of normality of the error term. For each of 3 willingness-topay measures, the following 3 models were estimated: (1) all eligible employees were combined to examine the association between willingness to pay and sociodemographics, prior experience with influenza-like illness (since November 2007), comorbidities, and current vaccination status; (2) the sample was restricted to individuals who were vaccinated (or in the case of children and adults, >1 child and adult, respectively, in the employee’s household were vaccinated) and incorporated additional variables regarding reported reasons for vaccination; and (3) the sample was restricted to individuals who were not vaccinated (or in the case of children and adults, no children and adults, respectively, in the employee’s household were vaccinated) and incorporated additional variables regarding reported reasons for nonvaccination. Models were specified on the basis of a priori assumptions. Model output is presented as cost ratios, including the untransformed β coefficients from which the cost ratios were derived. For binary variables, the cost ratio is interpreted as the ratio of willingness to pay by employees for whom the binary indicator is 1 to willingness to pay by employees for whom the binary indicator is 0. For continuous variables, the cost ratio is interpreted as the incremental relative increase in willingness to pay for a 1-unit (1-U) increase in the variable. Statistical analyses were conducted using STATA release 9 (StataCorp LP, College Station, TX).
During the time-limited enrollment period, 3686 employees completed the Web-based screening questionnaire. Of these employees, 2298 (62.3%) met the inclusion and exclusion criteria and completed the baseline survey. The percentage of employees who completed the survey each month ranged from 95.3% to 97.4% (data not shown). The final sample included 2006 employees (87.3% of 2298 initial responders) who had nonmissing willingness-to-pay responses in the final survey.
Table 2 summarizes the characteristics of employees and their household members. Employees were 31.3% female, their mean age was 41.7 years, 85.3% were of white race/ethnicity, 83.9% had attended at least some college, and the mean household size, including the employee, was 4.0 individuals. The demographics of employees included in the CHIEF study were comparable to US Census Bureau data on heads of households (also known as householders) 18 years or older.20 In 2007, heads of households were 29.9% female, they had a mean age of 49.3 years, and 81.6% were of white race/ethnicity. However, employees in the CHIEF study were more educated than the general US heads of households (83.9% vs 56% had attended at least some college). Overall, 31.5% of employees, 27.5% of child household members, and 24.5% adult household members were vaccinated during the 2007-2008 vaccination season.
Willingness-to-Pay Estimates and Associated Factors
Employees’ median (mean [SD]) willingness to pay for influenza prevention was $25 ($72 [$177]) for themselves (n = 2006) ($0 minimum and $400 maximum; 5th-95th percentiles, $0-$250), $25 ($82 [$280]) for their adult household members (n = 1835) ($0 minimum and $9999 maximum; 5th-95th percentiles, $0-$400), and $50 ($142 [$447]) for children (n = 2003) ($0 minimum and $9999 maximum; 5th-95th percentiles, $0-$500). The difference in willingness to pay for children versus for adults or themselves was significant (P <.01).
Table 3 gives willingness-to-pay estimates by the reported reasons for vaccination or nonvaccination. The most frequently reported reasons for vaccination were healthcare provider recommendations (for children) and a desire to protect family and household members (for employees and adult household members).
In the multivariate models in which all employees were combined to examine the factors associated with willingness to pay, the receipt of a vaccine in the current season was associated with significantly higher willingness to pay for children, adult household members, and employees (P <.01 for all models). The full results of these models are available from the corresponding author.
Table 4 summarizes results for the subset models in which willingness to pay was examined for the aforementioned vaccinated groups listed in the “Methods” section. Of the various reported reasons for vaccination (holding other factors constant), access to free vaccination was associated with significantly lower willingness to pay in all models (P <.01 for all). Neither belief in vaccine efficacy nor experience of an influenza-like illness was significantly associated with willingness to pay.
Table 5 summarizes results for the subset models in which willingness to pay was examined for the aforementioned nonvaccinated groups listed in the “Methods” section. Household income of $50,000 or more was associated in all models with significantly higher willingness to pay (P <.01). Black race/ethnicity was associated with a significantly higher willingness to pay for personal prevention (P <.05). Among various reported reasons for nonvaccination, disagreement with national influenza vaccination recommendations was associated in all models with significantly lower willingness to pay (P <.01). In the model for children, cost and dislike of vaccinations