Objective. The economic impact of routine vaccination of preschool children with inactivated influenza vaccine was investigated.
Design. A decision analysis was performed using data from the literature. Direct and indirect costs of each vaccination strategy were calculated and compared with a strategy of not vaccinating.
Setting. Two settings were evaluated: a setting in which vaccination was available during flexible hours and a setting in which vaccination was available only during usual work hours (8:00 am–5:00 pm).
Results. Vaccination resulted in a net cost savings in both settings. The net savings per vaccine recipient were $21.28 in the flexible setting and $1.20 in the restricted setting. Although the analysis was performed for the inactivated vaccine, sensitivity analysis showed that the nasal vaccine could also result in a net cost savings depending on the price of the cold-adapted vaccine when it is licensed.
Conclusion. Vaccinating preschool children is economically advantageous. Serious consideration should be given to recommending vaccination in this age group.
- NHIS =
- National Health Interview Survey
Influenza infects children more often than adults,1,2 but little attention has been paid to vaccinating children. This is partly because serious complications of infection are uncommon in children.3–7 However, influenza can cause discomfort and symptoms that require children to restrict their ordinary activities or avoid day care. Illness in children affects the entire family. Young children who are too ill to attend day care or preschool need an adult caretaker at home. This often means that an adult must take time off from work to care for the child. The resulting lost productivity has been shown to be substantial in school-aged children,8 but no study has examined the economic impact of influenza or vaccination in preschool children. Therefore, we performed a decision analysis to examine the cost or savings attributable to vaccination in this population.
This analysis was conducted for children between 6 months and 5 years of age from the general US population. A societal perspective was adopted. Two settings were examined: a flexible setting in which the vaccine was available outside work hours and a restricted setting in which the vaccine was available only during usual work hours (8:00 am–5:00 pm, Monday–Friday). In the restricted setting, it was assumed that wages could be lost by employed parents to obtain vaccination for the child. Unemployed parents, those who worked part time, and those with in-home day care were assumed not to miss work to obtain vaccination. In the flexible setting, it was assumed that parents did not miss work to get the child vaccinated. Vaccination in each setting was compared with the strategy of not vaccinating children. Results were expressed as a net cost (savings) of vaccination per recipient compared with not vaccinating.
Direct costs included the cost of disease in children and the adult contacts to whom they transmitted the infection. Indirect costs included the wages lost by a parent who stayed home when a child became too ill to attend day care and the wages lost by an adult contact who became ill and missed work. Direct costs and indirect costs related to otitis media were included because vaccination against influenza has been shown to reduce the incidence of otitis media.9–11
The analysis was hypothetical. Costs and probabilities were obtained from the literature. Although the analysis was based on the injectable, inactivated influenza vaccine, use of the live, nasal vaccine was investigated in the sensitivity analysis.
The costs of vaccination, including supplies, personnel, administrative expenses, and the vaccine, have been found to be as low as $4 per recipient in a health maintenance organization12and as high as $10 in an outpatient clinic.13 The $10 figure was used for the analysis.
The current recommendation for the inactivated influenza vaccination is that each child less than 9 years old receive 2 injections, 1 month apart, during the first year of vaccination.14 In subsequent years only 1 injection is necessary. Because this analysis included children from 6 months to 5 years of age, it was assumed that none of the children would have received the primary series. In a steady-state system, children would receive the 2-dose series in their first influenza season and need only 1 dose in the subsequent 4 seasons. It was therefore assumed that one-fifth of the population would need a second injection in any year, raising the cost of vaccination by 20%.
The direct costs of disease included a physician visit for infected children and for secondarily infected adults who presented for medical evaluation. The cost of a visit to a pediatrician's office was assumed to be $51,15 and the cost of an emergency department visit was assumed to be $124.42.16 It was assumed that there would be 0.099 excess outpatient visits per child for influenza each year in the absence of vaccination.6 It was assumed that this figure included all visits related to influenza, including visits for influenza-associated otitis media. Because higher visit rates have been reported in some settings,1,2,17 the use of outpatient facilities was investigated through sensitivity analysis. Although 7.6% of non–injury-related outpatient visits by children are to emergency rooms,18 data for upper respiratory infections show a lower rate of 4.95%.19 The latter rate was used in the analysis.
The recent study by Neuzil et al6 analyzed excess antibiotic use attributable to influenza. This was found to average 0.072 excess courses of antibiotics per preschool-aged children. It was assumed that a course of antibiotics would cost $9.91.19
For adult contacts, it was assumed that 27% of ill patients would visit an outpatient clinic.1,2 The cost of a physician visit for an adult was assumed to be $69.51.13
Hospitalizations related to influenza have been analyzed in 2 recent studies. The first found an average of 0.00176 excess hospitalizations related to influenza per child per year.6 In the second study, children between the ages of 6 months and 1 year were not specifically separated from those under 6 months of age, and this might account for the slightly higher rate of 0.00196 per child.7 Because influenza vaccination is not effective before the age of 6 months, the former number was used. The average inpatient stay for a child with respiratory infection ranges from 3.1 days for bronchiolitis to 3.7 days for pneumonia.20 The average stay for influenza-related hospitalizations was assumed to be in the middle of this range (3.4 days). The average cost of a hospital day for respiratory disease was $828 in a recent study,16yielding a cost estimate of $3064 per hospitalization. This figure is lower than the median charge ($3400–$4200) for inpatient treatment of respiratory illness in children20,21 and lower than the cost of admission for respiratory syncytial virus infection.22
Indirect costs were based on the wages lost by parents because of influenza in their children, influenza in parents as a result of contact with their child, and time spent obtaining vaccination for the child. For this calculation, a day's salary was assumed to be $93.40 for women and $123.40 for men.23 To be conservative, the woman's wage was assumed for all caretakers. In other words, it was assumed that employed mothers rather than employed fathers would stay home with an ill child. Men lost wages only if they themselves became ill through contact with the child. It was also assumed that 68% of households had two parents.24 Ninety-seven percent of men and 65% of women with preschool-aged children were assumed to be working, 24% of those women only part time.24,25 In-home day care was assumed to be available in 13% of households.26 If a parent did not work or if in-home day care was available, no indirect costs were incurred. Part-time workers were assumed to lose half the wages of full-time workers during illness and were assumed not to miss work to obtain a vaccination for a child. It was assumed that families with multiple preschool children would bring them all at once for vaccination, resulting in 1.2 children vaccinated per visit.
Bed days per preschool-aged child with influenza average 0.89 days (range 0.6 to 1) in the most recent National Health Interview Surveys (NHISs),2,17 and this figure was used to estimate the homebound days. It was assumed that a bedbound child would need an adult caretaker at home.
The risk of transmitting influenza from a young child to an adult contact was found in a cross-sectional study to be 28.6%.27 Susceptible contacts were assumed to be adult members of the child's household, or 1.68 contacts per child. Transmission to other children or persons outside the home was not included in this model. Indirect costs of transmission to adult contacts were incurred when the adult lost wages because of missed work. Based on a recent clinical trial, vaccination was assumed to reduce lost work days by 0.52 days per recipient.13 This is consistent with NHIS data showing that women with influenza missed 0.593 to 0.776 days of work and men missed 0.563 to 0.568 days when ill with influenza.1,2 Any physician visits were assumed to occur while adults were too ill to work.
It was assumed that children with otitis would be bedbound for an average of 0.5 days each.1,2,17 Indirect costs of otitis media through lost wages were calculated in a similar manner to the indirect costs of influenza. The annual incidence of otitis media was assumed to be 0.627 in children less than 5 years old based on NHIS data.1,2 An earlier study by Biles et al28found the incidence to be 0.551 in children less than 8 years old. The higher number was used in this analysis because Biles's data included older children as well, a group with a lower incidence of otitis media. Otitis media is known to have a seasonal variation, with 44% of the total yearly incidence occurring during influenza season, from December to March.28 Thus the incidence of otitis media during influenza season was calculated to be 0.276.
Incidence and Efficacy
Several studies have estimated the incidence of influenza infection in the young pediatric population. One study. which defined an infection as a symptomatic, culture-proven episode, found an infection rate of 31% in children less than 3 years old who were seronegative before the influenza season.4 A multiyear study found an average 37% incidence of symptomatic infection among children less than 6 years old.3 These numbers are consistent with NHIS data from 1994, with an influenza incidence of 36.7% in preschool children.1 However, the 1995 NHIS found a much higher incidence of 53.6% in children less than 5 years old.2 Unfortunately, NHIS data represent a calendar year rather than an influenza season and are based on self-report. For the purposes of this analysis, an incidence of 37% was assumed to represent an average influenza season. The effect of more severe and less severe influenza seasons was investigated by sensitivity analysis.
The efficacy of the vaccine was assumed to be 83%9 in preventing clinically apparent infection in children less than 5 years old. The vaccine was found to be 32% effective in reducing otitis media in 6- to 30-month-old children during the influenza season10; this correlated well with other studies.9,11 Based on the results of other studies20,28 it was assumed that vaccine side effects would be mild and free of cost.
Sensitivity analysis was performed by varying key assumptions using ranges from the literature (Table 1). Vaccine efficacy and cost, disease incidence, rate of secondary transmission, and the effect of vaccine side effects were examined. The analysis was also run assuming that there was no excess cost of hospitalization for influenza.7 The use of outpatient services for ill children and the number of days spent in bed were investigated.
As part of the sensitivity analysis, the cost (savings) associated with use of the nasal vaccine was analyzed. It was assumed that no booster doses would be needed and that the vaccine would have an efficacy of 89% in children.11 The analysis examined how much a single dose could cost to maintain a net savings per recipient.
Finally, the analysis was run for a subpopulation limited to parents who had to miss work to obtain vaccination for their children.
Vaccination resulted in a net savings in the restricted and flexible settings (Table 2). The savings were greater if the vaccine was available outside traditional working hours. Specifically, vaccination in the flexible setting resulted in a net cost savings of $21.28 per recipient. This savings fell to $1.20 per recipient if vaccination was available only in a setting with restricted hours. If indirect costs were not included, vaccination would still result in a modest savings per recipient of $0.34.
Sensitivity analysis revealed that for a broad range of assumptions, vaccination in a flexible setting retained its net cost savings. The net savings associated with vaccination in a restricted setting were less robust.
The efficacy of the vaccine was assumed to be 83% based on a clinical trial.9 The literature reports an efficacy as low as 54% during a severe influenza season caused by antigenically drifted influenza A virus.31 At this level, vaccinating in a setting with flexible hours continued to result in a net cost savings, but vaccination in a restricted setting resulted in a net cost of $9.70 per child (Table 3). In fact, if the opportunity to vaccinate a child were restricted to normal work hours, there would be no net savings if the efficacy were less than 78%.
The incidence of influenza varies from year to year. In a year in which the incidence was below 14%, there would be no net savings from vaccination regardless of setting (Table 3). If the incidence fell below 36%, savings would no longer be possible in a restricted setting.
The vaccine has been shown to be effective in preventing otitis media during the influenza season.9,10 Even if there were no effect on the incidence of otitis media, vaccination during the base year in a flexible setting would still yield a cost savings. Vaccination in a restricted setting would result in a net cost of $0.85 per recipient if the indirect costs of otitis media were excluded. The effect of excluding the cost of hospitalization was similar: Vaccination in a flexible setting would continue to result in a net cost savings, whereas vaccination in a restricted setting would result in a net cost of $3.27 per recipient.
The cost per dose of vaccine was investigated (Table 3). Vaccination in a flexible setting continued to result in a net cost savings as the cost of the vaccination was increased to a maximum of $28 or in a restricted setting if the cost were increased to a maximum of $11 per dose.
We assumed that the transmission rate of influenza from a young child to an adult contact was 28.6%.27 Even if no transmission occurred within the household, vaccinating children less than 5 years old would continue to show cost savings in a flexible setting. The model assumed that no costs were associated with potential adverse effects of vaccination. Theoretical concerns center around risks of local, systemic, and allergic reactions. The incidence of fever in vaccine recipients was not greater than the incidence of fever in placebo recipients,29,30 so there would probably be no increase in absenteeism from day care because of fevers caused by vaccination. However, if there were costs associated with side effects, they would have to exceed $18 per recipient to eliminate the cost savings from vaccination in a flexible setting and $1 per recipient in a restricted setting.
The live attenuated intranasal influenza vaccine has been shown to have an efficacy of 89% in a single-dose regimen.11 This vaccine may be better tolerated by children and their families than an injectable product, making vaccination more attractive. However, it is unlikely that this new vaccine would cost as little as the current inactivated vaccine. Therefore, we determined the maximum cost of the nasal vaccine that could still result in a net savings. In a flexible setting the vaccine could cost as much as $36 per dose and still maintain a net cost savings. In a restricted setting, the cost per dose could be as much as $16.
We analyzed the costs associated with influenza vaccination in a specific subpopulation that would be most likely to benefit from the vaccine: families in which both parents work and there is no in-home day care. Targeting this subpopulation for vaccination with the inactivated vaccine resulted in a net cost savings of $43 per child in a flexible setting. In a restricted setting, in which parents would have to miss work to get the child vaccinated, there was a net cost of $9 per recipient.
Influenza vaccination of preschool children resulted in a net cost savings per recipient if performed in the general population regardless of setting. The savings were highest when the vaccination was available in a setting with flexible hours. This flexibility eliminated the need for employed parents to miss work to immunize their child.
We used a societal perspective that included indirect costs of time lost from work. It is important to note that vaccination also saved money in terms of direct medical costs alone. This is important because the perspective of a third-party payer would include only the direct medical costs of disease. If reasonably priced, the live attenuated nasal vaccine would also result in a net cost savings.
The analysis was undertaken from a conservative point of view. Factors that would decrease the cost of vaccination or increase the cost of disease would make immunization more attractive. For example, it might be possible to vaccinate children during an office visit for a medical problem or well child care. Vaccinating healthy children could reduce the risk of disease spread to high-risk children and adults. Conversely, factors that would decrease the attractiveness of vaccination include a lower incidence of influenza, as would occur in a milder year. Unfortunately, it is not possible to predict the severity of an influenza season with accuracy.
For the analysis, we considered only the theoretical transmission of influenza from young children to adult household contacts. Certainly the virus could be transmitted to other children or adults as well, increasing the cost savings from vaccination.
Vaccination saved the most money among parents who had to miss work to care for an ill child. In addition to the financial benefits, some parents might also consider vaccination appropriate to reduce the risk of their children suffering symptoms of influenza. Others might prefer a scheduled and predictable loss of work for a vaccine appointment to an unscheduled stay at home with an ill child.
Our previous analysis found vaccination to result in a net savings for school-aged children that ranged from $4 to $35 per recipient depending on setting.8 Younger children have a higher incidence of complications, including otitis media and hospitalization,1,2,6 but are less likely to have parents who are employed full time.24 These factors balance each other to result in similar findings in both populations.
Influenza vaccination for preschool children resulted in a net savings in both indirect and direct costs. The cost savings of vaccination in children is less than in working adults and older adults, for whom savings of more than $45 per recipient have been found.12,13 Reasons include the lack of serious complications of influenza in children and the fact that it is the parents rather than the ill children who incur indirect costs. In contrast, childhood vaccination against varicella results in a savings in indirect costs but not in direct costs.32,33
Childhood influenza causes morbidity in the child and the potential for transmission to susceptible adults. Strong consideration should be given to recommending vaccination against influenza in children.
- Received December 17, 1999.
- Accepted March 31, 2000.
Reprint requests to (M.D.N.) Virginia Commonwealth University, Box 980102, Richmond, VA 23298. E-mail:
- Adams PF,
- Marano MA
- Benson V,
- Marano MA
- Monto AS,
- Koopman JS,
- Longini IM
- White T, Lavoie S, Nettleman MD. Potential cost savings attributable to influenza vaccination of school-aged children. Pediatrics. 1999;103(6). URL: http://www.pediatrics.org/cgi/content/full/103/6/e73
- Medical Letter
- American Medical Association. Socioeconomics of Medical Practice, 1997. ML Gonzales, ed. Chicago, IL: American Medical Association; 1997:69, 89, 107
- Adams PF, Hendershot GE, Marano MA. Current estimates from the National Health Interview Survey, 1996. National Center for Health Statistics. Vital Health Stat. 1999;10(200)
- Schappert SM
- Lawrence L,
- Hall MJ
- Meurer JR, Kuhn EM, George V, Yauck JS, Layde PM. Charges for childhood asthma by hospital characteristics. Pediatrics. 1998;102(6). URL: http://www.pediatrics.org/cgi/content/full/102/6/e70
- Bureau of Labor Statistics 1999. http://stats.bls.gov
- Johnson O, ed. 1997 Information Please Almanac. Boston, MA: Houghton Mifflin; 1997:66,69,70
- Brunner B, ed. 1999 Time Almanac. Boston, MA: Information Please LLC; 1998:844
- National Household Education Survey 1995. http://nces.ed.gov/pubs/95824.html
- Frank AL,
- Taber LH,
- Glezen WP,
- et al.
- Edwards KM,
- Dupont WD,
- Westrich MK,
- et al.
- Copyright © 2000 American Academy of Pediatrics