Cost-effectiveness Analysis of a Rotavirus Immunization Program for the United States

(1)

Cost-effectiveness

Analysis

of

a Rotavirus

Immunization

Program

for

the

United

States

Jean C. Smith, MD, MPH*; Anne C. Haddix, PhD; Steven M. Teutsch, MD, MPH; and

Roger I. Glass, MD, PhD*

ABSTRACT. Objective. To estimate the economic

con-sequences in the United States of routine childhood

im-munization of children younger than 1 year of age with a

rotavirus (RV) vaccine.

Design. Cost-effectiveness analysis of a national RV

immunization program from the perspective of the

health care system and the perspective of society.

Esti-mates of disease incidence, medical expenditures,

pro-ductivity costs, vaccine efficacy, and vaccine coverage

rates were derived from published literature and

unpub-lished vaccine trial reports. The impact of changes in

estimates of vaccine efficacy and medical costs was

de-termined by sensitivity analysis.

Main Outcome Measures. Incremental cost

effective-ness, expressed as savings per case of RV diarrhea

pre-vented.

Reults. Given a vaccine efficacy rate of 50% and a

vaccine cost of $30 per dose, an RV immunization

pro-gram would prevent more than 1 million cases of RV

diarrhea, 58 000 hospitalizations, and 82 deaths per year.

A vaccine program would cost $243 million per year but

would yield net savings of $79 million from the

perspec-tive of the health care system and $466 million from the

perspective of society. The incremental cost effectiveness

was a savings of $459 per case prevented from the societal

perspective and $78 per case prevented from the health

care system perspective. Sensitivity analyses

substanti-ated net savings over a wide range of variables, and cost

effectiveness increased with greater vaccine efficacy or

decreased vaccine cost.

Conclusions. Economic and disease reduction benefits

would be realized from the use of an RV vaccine that is

partially protective against severe RV diarrhea. These

findings suggest that immunization with an RV vaccine

would be cost effective and cost saving. Pediatrics 1995;

96:609-615; rotavirus, childhood immunization, vaccines,

cost effectiveness, decision analysis.

ABBREVIATIONS. RV, rotavirus; DTP,

diphtheria-tetanus-pertussis.

Rotavirus (RV) is the most common cause of

Se-vere diarrhea in children worldwide. Virtually all

From the *Viral Gastroenteritis Section, Respiratory and Enteric Viruses Branch, Division of Viral and Rickettsial Diseases, National Center for

Infectious Diseases, tPrevention Effectiveness Activity, Epidemiology

Pro-gram Office, and §Division of Training, Epidemiology Program Office,

Centers for Disease Control and Prevention, Atlanta, GA.

Received for publication Feb 16, 1995; accepted Apr 20, 1995.

Reprint requests to (J.C.S.) Viral Gastroenteritis Section (Mail Stop G-04),

Centers for Disease Control and Prevention, 1600 Clifton Road, NE, Atlanta, GA 30333.

PEDIATRICS (ISSN 0031 4005). Copyright © 1995 by the American Acad-emy of Pediatrics.

children are infected with RV in the first 3 to 5 years

of life; neonatal and early infant infections can be

asymptomatic, but first infections after 3 months of

age more commonly are symptomatic.1 Based on

data from 1979 through 1985 for the United States,

RV infections cause an estimated 3.1 million cases of

diarrhea, 65 000 to 70 000 hospitalizations, and 125

deaths annually among infants and young

chil-dren.24 A routine RV immunization program could

prevent much of this morbidity and mortality, but no

cost-effectiveness analysis of such a program has

been reported.

RV vaccines currently being developed are live

attenuated oral preparations that would be

adminis-tered in three doses to infants at ages 2, 4, and 6

months, concurrent with routine childhood vaccines

(diphtheria-tetanus-pertussis [DTP], oriti poliovirus vaccine, and Haeinophilus influenzae type b).9 Since

1989, multicenter trials have demonstrated efficacy

rates of 50% to 80% in the prevention of RV

diar-rhea.1013 An RV vaccine may be ready for licensing

by 1996, and recommendations will be needed to

guide the decision of whether to include the vaccine

in the US national immunization program. When the

vaccine becomes available, such a decision will likely be based not only on demonstration of clinical effec-tiveness and safety, but also on cost effectiveness.1417

We performed a cost-effectiveness analysis to

ad-dress two questions: (1) given the partial efficacy of

RV vaccine achieved in recent trials and current DTP

vaccine coverage rates, would immunization of

in-fants younger than 12 months of age in the United

States be cost effective; and (2) how would cost

ef-fectiveness change if the United States achieved the

90% vaccine coverage goal of the Comprehensive

Childhood Immunization Initiative for I996?t8

METHODS

Decision Analysis Model

We constructed a decision tree to compare two alternative

options, “vaccination program” or “no vaccination program,” with the vaccination program option representing inclusion of an RV vaccine in a national immunization program (Fig 1). Currently,

in the absence of a vaccine, a child’s risk of having RV diarrhea by

5 years of age-.-shown in the tree as the node labeled “ill”-is the cumulative incidence rate. Illness may be mild, requiring no

med-ical attention, or severe, leading to dehydration and its complica-tions that require physician visits or hospitalization, or rarely, may result in death. In the vaccination program option, the probability that a child will be vaccinated with one, two, or three doses is based on current rates of DIP vaccination coverage in US children at age 12 months.tv Because vaccine efficacy is less than 100%, some vaccinated children also will develop RV diarrhea. Adverse

at Viet Nam:AAP Sponsored on September 1, 2020

www.aappublications.org/news

(2)

No Vaccination

Program

: Choice node

0 Chance node

Full III

E:J Outcome

Fig 1. Decision tree for an RV immunization program in the United States. The vaccination program option represents inclusion of an RV vaccine in a national immunization program, for a birth cohort of 4.1 million children (followed from birth to 5 years of age). With a vaccination program, some children may not be vaccinated, and some of those vaccinated may have RV diarrhea. Children who become ill (node labeled ill) with RV diarrhea may have mild symptoms or severe symptoms resulting in a physician visit, hospitalization, or death. The no vaccination program option represents the current situation with no RV vaccine. The decision tree was analyzed with SMLTREE software (see “Methods”).

reactions in RV vaccine trials have been negligible, consisting of runny nose, lassitude, and low-grade fever in a small number of children, and were not included in our model.’#{176}#{176}The model was analyzed with SMLTREE decision analysis software (version 2.9, Jim Hollenberg, Roslyn, NY).

Study Design

We performed a cost-effectiveness analysis from two perspec-tives: (I ) the health care system perspective, which includes only the direct outpatient and inpatient medical costs associated with RV vaccination and RV diarrhea; and (2) the societal perspective, which includes productivity costs attributable to premature loss of life and time lost from work to care for a sick child.2021 This analysis is based on a 1-year time frame in which the intervention is applied and a 5-year period during which the costs and benefits

of health outcomes resulting from the intervention are realized. We analyzed the hypothetical experience of the 1991 US birth cohort of 4.1 million children from birth to 5 years of age, because virtually all RV morbidity and mortality occur in this age group.

We calculated a summary measure of cost effectiveness through an incremental analysis, in which the additional costs that one program imposes over another are compared with the addi-tional benefits it delivers. In this case, the incremental cost effec-tiveness is the ratio of the difference between total costs attributed to RV diarrhea with and without a vaccine program and the difference between number of cases that occur without and with a vaccine program.

Probability Estimates

The probabilities of events in the decision tree were derived from published studies, unpublished data from national sources,

and unpublished reports from double-blind, placebo-controlled vaccine efficacy trials (Table 1). When estimates varied widely or were unknown, we chose the estimate that would bias against

immunization.

Serologic surveys indicate that nearly all children are infected with RV in the first few years of life, and more than 90% have

antibodies by 3 years of age.1 The cumulative incidence of RV

diarrhea by 5 years of age is estimated to be 75%. In prospective longitudinal studies of the natural history of RV infection, 72% to

88% of children have RV diarrhea in the first 5 years of 1ife.26

Lower rates (22% to 35%) have been observed in the placebo groups of vaccine trials, but these have followed children for 2

years or less, and multicenter studies have great variability in

disease rates between sites depending on the intensity of surveil-lance, which would tend to underestimate the true incidence of RV diarrhea. Recent vaccine trials in the United States have dem-onstrated reduction in the incidence of RV diarrhea by approxi-mately 50% in the vaccinated group;10”t therefore, the estimated risk of RV diarrhea in a vaccinated cohort during the first 5 years of life is estimated to be .38 (50% of .75). Children who receive only one or two doses of vaccine are assumed to have an intermediate risk of RV diarrhea of .56.

RV diarrhea can be severe, requiring the patient to seek medical care. The probability of severe illness was estimated to be .28 (28 severe episodes per 100 cases of RV diarrhea) based on five nonvaccine studies, with rates ranging from 20.4 to 40.9 severe episodes per 100 cases of RV diarrhea.2’3’24’2628 Most longitudinal

studies were too small to measure rates of hospitalization or

death. Based on our previous national estimates for these out-comes (ie, 500 000 physician visits, 67 500 hospitalizations, and 100 deaths), we estimated the relative proportions of severe episodes

(3)

TABLE 1. RV Diarrhea-related Probability Estimates

Variable Base Case Events/lO 000 Low and High Reference

Estimate Children* Estimates

Probability of RV diarrhea, age 0-5 y

Unvaccinated (0 doses) 0.75 7,500 0.50,0.88 1,23-26

Partially vaccinated (1-2 doses) 0.56 5,600 0.38,0.66 See text

Vaccinated (3 doses) 0.38 3,800 0.25,0.44 10-12

If ill, probability of severe RV diarrhea

Unvaccinated 0.28 2,100 0.20,0.40 2,23,24,26-29

Vaccinated 0.14 532 0.10,0.20 10-12

If severe RV diarrhea, probability of: Physician visit

Unvaccinated 0.8809 1,850 0.8809 2-4,28-30

Vaccinated 0.8809 470 0.8809

Hospitalization

Unvaccinated 0.1189 250 0.1189

Vaccinated 0.1189 63 0.1189

Death

Unvaccinated 0.0002 0.42 0.0002

Vaccinated 0.0002 0.11 0.0002

Vaccine coverage ratet 0.79 0.66,0.90 See text; 18,19

* Number of events per 10 000 children age 0-5 years with base case estimates.

1At least one dose of rotavirus vaccine by age 12 months.

that would lead to physician visits (500/500 000 + 67 500 + 100 = Disease Control, personal communication, March 1994). The goal

0.8809), hospitalizations (67/500 000 + 67 500 + 100 = 0.1189), or of the Comprehensive Childhood Immunization Initiative is a

deaths (100/500 000 + 67 500 + 100 0.0002).2 vaccine coverage rate of 90% by 1996.18

Two recent field trials indicate that RV diarrhea is less severe

among those who have been vaccinated than those who have

received a placebo, measured as symptom severity and rates of Cost Estimates

physician visits)011 Based on these studies, we estimated the prob- All costs (Table 2) are estimated in 1993 dollars,31 and future

ability of severe illness (ie, illness resulting in an outpatient visit, costs and benefits have been discounted to present value at an

hospitalization, or death) in the vaccinated group to be half that of annual rate of 4%. Direct medical costs include costs of outpatient the unvaccinated group. Thus, the efficacy of the RV vaccine is clinic visits, hospitalization, and treatment of a dying child.32’ In reflected both in a 50% decreased probability of illness and in a this model, the four categories of health outcomes-mild illness

50% decrease in the probability of severe illness. The relative requiring home treatment or severe illness resulting in outpatient

proportions of outpatient visits, hospitalizations, or death among visits, hospitalization, or death-are assumed to be mutually ex-children with severe RV diarrhea are assumed to be the same for clusive. The cost of an outpatient medical visit includes one office the vaccinated and the unvaccinated group (0.8809, 0.1189, and physician visit, laboratory tests, and medications and is derived

0.0002, respectively). from national data on medical practice costs.M Hospitalization

Because the RV vaccine would be administered concurrently costs include daily room, inpatient physician visit, medications, with the DTP vaccine, we assumed that the RV vaccine coverage intravenous fluids, laboratory tests, and one outpatient visit after rate would approximate the rate of DIP coverage.19 In the United discharge (R. E. Lapp, PhD, Blue Cross and Blue Shield

Associa-States, 79% of children will have received at least one dose of DIP tion, Chicago, IL, written communication, January 1994). The

vaccine by age 12 months, and of those, 68% receive all three doses duration of hospitalization for diarrhea has declined in recent

(S. Hadler, MD, National Immunization Program, Centers for years, reflecting national trends toward shorter hospital stays.

TABLE 2. RV Diarrhea-related Cost Estimates (1993 Dollars)

Variable Medical Costs Reference

Cost, $ Cost, $

Base Case Estimate Low and High Estimates

Outpatient clinic visit (physician, lab, 94 51, 94 34-36

medication)

Hospitalization (room, physician visit, medications, intravenous fluids, lab tests)

Unvaccinated child (3.75 d) 3615 3615,5050 R. Lapp; 10,11,38,39,42

Vaccinated child (2.5 d) 2410 2410,3367 R. Lapp; 10,11,38

Postdischarge outpatient visit 51 51 35,36

Emergency department visit (dying child)

Unvaccinated or vaccinated child 696 569, 696 R. Lapp; see text

Other direct costs of RV diarrhea episode (oral 45 20, 45 34

rehydration solution, transportation, child care, extra diapers)

Oral RV vaccine, one dose 20 5, 25 See text

Vaccine administration, one dose 10 10

Productivity Costs

Foregone earnings of parent/care giver of 66 0, 66 40

child with RV diarrhea (per d)*

Lifetime productivity loss, child with death 601,857 0, 601,857

due to RV diarrhea

* Days of work loss of parent or care giver with: child with mild RV diarrhea (vaccinated, 2.5 days; unvaccinated, 3.5 days); hospitalized

child (vaccinated, 2.5 days; unvaccinated, 3.75 days); dying child (vaccinated or unvaccinated, 4 days).

(4)

Therefore, we estimated duration of hospitalization in the unvac-cinated group to be 3.75 days, based on the most recent National Center for Health Statistics data,35 rather than the estimates of 4.2 and 5.3 days observed in earlier studies.24 Because vaccine trials have demonstrated a shorter duration of illness in the vaccine

group than the placebo group (mean, 2.5 vs 3.5 days), a hospital

stay of 2.5 days was assumed for the vaccinated group. The cost

for treatment of a dying child includes ambulance transportation and 30 minutes of critical care in an emergency department (R. E.

Lapp, PhD, written communication, January 1994). The estimated

cost of vaccine is based on a weighted average of public and

private sector purchase and administration of comparable

child-hood vaccines (R. H. Snyder, MA, National Immunization

Pro-gram, Centers for Disease Control, written communication,

De-cember 1993). Administration costs do not include the cost of an

outpatient visit, because the vaccine would be administered along

with other scheduled routine childhood vaccines. Productivity

costs include the foregone earnings of one wage earner caring for

achild with RV diarrhea and the discounted foregone wages of a

child dying before the age of I year.#{176}

Sensitivity Analysis

Sensitivity analyses were conducted on the variables in the

model for which values were uncertain. These variables include the probability of RV diarrhea, the probability of severe illness

requiring medical intervention, vaccine cost, medical treatment

costs, vaccine efficacy rates, and vaccine coverage rates. The dis-count rate was also varied from the base case of 4%, across a range

from 2% to 8%. In a worst case scenario, the extreme estimates

(Tables I and 2) that would bias the model against an

immuniza-tion program were selected for the probabilities and costs of

RV-related illness; in a best case scenario, estimates that would

bias the model in favor of an immunization program were

selected.

RESULTS

The health outcomes and costs of RV diarrhea for

a cohort of 4.1 million children followed from birth to

5 years of age have been calculated without and with

an immunization program (Table 3). With base case

estimates, immunization is predicted to prevent

I 015 000 of the cases of RV diarrhea that would be

expected without a vaccine, 433 000 physician visits,

58 000 hospitalizations, and 82 deaths during the 5

years that the cohort is followed. At a vaccine cost of

$30 per dose, an immunization vaccine program

would cost $243 million for the purchase and

admin-istration of the vaccine, but it would save the health care system $322 million in medical costs of illness

averted, yielding a net savings of $79 million in

discounted costs. From the societal perspective,

when the value of productivity gained is added to

direct medical costs, an immunization program is

expected to have a net savings to society of $466

million in discounted costs.

Cost-effectiveness Ratios

From the perspective of the health care system, the

incremental cost effectiveness of immunization is a

savings of $78 per case prevented; that is, costs to the

health care system are decreased by $78 for each

additional case of RV diarrhea prevented. From the

perspective of society, the incremental cost effective-ness is a savings of $459 for each case prevented.

Sensitivity Analysis

Univariate sensitivity analyses were conducted on

the probability and cost estimates. Estimates were

varied over the ranges shown in Tables 1 and 2,

demonstrating no substantial change in the outcome

of the analysis.

Vaccine Efficacy

Because it is unlikely that vaccine efficacy will

decrease below levels seen in current field trials, and becuase our base case efficacy estimate was selected to bias the analysis against immunization, we did not

vary the vaccine efficacy rate below that used in the

base case. Because it is possible that efficacy rates

may increase with technological improvements, we

calculated cost savings and number of cases of RV

diarrhea prevented with increased vaccine efficacy.

If an efficacy rate of 70% in the prevention of RV

diarrhea could be achieved, net savings to the health

care system would increase by 44%, from $79 million

to $1 14 million, and from the perspective of society,

net savings would increase by 23%, from $466

mu-lion to $575 million; health outcomes would improve

by 40%, from I 015 000 to 1 420 000 cases of RV

di-arrhea prevented by immunization.

Vaccine Cost

The threshold vaccine cost below which

immuni-zation saves money from the perspective of the

health care system is $40 under base case estimates (see “Threshold Analysis”).

TABLE 3. RV-related Health Outcomes and Costs Without and With an Immunization Program*

Health Outcomes No Vaccine Vaccine No. Prevented by

Vaccine

Cases of RV diarrhea 3 075 000 2 060 000 1 015 000

Physician visits Hospitalizations Deaths

775 000 104 000 164

343 000 46 000

82

433 000

58 000 82

Discounted costs Net savings

(costs) of immunization vs no immunization Medical costs ($ in millions)

RV disease costs 564 242 322

Vaccine and administration 0 243 (243)

Total medical costs 564 485 79

Productivity costs

Total medical and productivity RV-related savings (costs)

817 I 381

430 915

387 466

* Birth cohort of 4.1 million children followed from birth to 5 years of age.

(5)

U) U)

0

(I)

C

0

U)

0 C.)

E

U) U) U) U)

U)

C.)

U) U) I

1,000

800

600

400

200

0

(200)

(400)

Netcost

Base Case

Worst Case

Best Case

Vaccine Efficacy

70%

40 80

Vaccine Cost (Dollars per Dose)

Vaccine Coverage

If national vaccine coverage were increased from

the current rate of 79% receiving recommended

rou-tine childhood vaccines by age 12 months, to the 1996

Comprehensive Childhood Immunization Initiative

goal of 90% (assume 90% of children receive 3 doses

of RV vaccine by age 12 months), an additional

370 000 cases of RV diarrhea would be prevented,

and the net savings to society would increase by $74

million.

Multivariate Sensitivity Analysis

We used estimates that would maximally bias

against vaccination, including a vaccine coverage

rate of 66.3%, vaccine efficacy rate of 50%,

probabil-ity of RV diarrhea of .50 in an unvaccinated child,

probability of severe diarrhea of .20 in an unvacci-nated child, decreased outpatient and inpatient med-ical costs, vaccine cost of $35, and productivity costs

of $0. With these worst case estimates, the

immum-zation strategy results in net savings to the health

care system as long as the cost of the purchase and

administration of the vaccine is less than $17 per

dose.

Threshold Analysis

We determined from the health care system

per-spective the threshold vaccine cost given base case

estimates (from Tables I and 2) as well as estimates

that would bias against vaccination (worst case) and

in favor of vaccination (best case) (Fig 2). Under base case estimates (Tables 1 and 2), the threshold cost per dose of vaccine is $40; that is, at a vaccine cost of less

than $40 per dose, the program would save money

from the health care system perspective, and at a

vaccine cost of more than $40 per dose, the program

would not save money. At a vaccine cost of $40 per

dose, all costs of the vaccine program are offset by

reductions in the cost of RV illness. The threshold

Net Savings

vaccine cost under worst case estimates is $17 per

dose and under best case estimates is $74 per dose of

vaccine. If the vaccine efficacy were increased to

70%, with all other probabilities and costs remaining

at base case estimates, the threshold vaccine cost

would be $44 per dose.

DISCUSSION

This analysis suggests that, despite a vaccine effi-cacy rate of 50% in the prevention of all RV diarrhea

and 75% in the prevention of severe RV diarrhea

resulting in physician visits, an RV immunization

program would be cost effective from the

perspec-tive of both the health care system and society.

Sen-sitivity analysis demonstrates that the model is not

sensitive to changes in the estimates of costs and

epidemiologic parameters. When cost and

probabil-ity estimates that would maximally bias against

im-munization are evaluated, a vaccine program

re-mains cost effective as long as the cost and

administration of the vaccine is less than $17 per

dose. This cost compares reasonably well with costs

of current routine childhood vaccines, which range

from $6.29 per dose (oral polio vaccine, vaccine only,

weighted public and private sector cost) to $20.31

(measles-mumps-rubella, vaccine only, weighted

public and private sector cost; R. H. Snyder, MA,

written communication, December 1993). Although

our analysis is based on the cumulative risk of RV

diarrhea during the first 5 years of life in a birth cohort of 4.1 million children, one also can view these

results as annual disease and economic burden

esti-mates for the total population of children younger

than 5 years of age.

Previous studies have estimated the disease and

economic burden of RV diarrhea. Estimates of the

annual medical costs of RV diarrhea, inflated to 1993 dollars, range from a low of $200 million noted in the Institute of Medicine study24 that assessed only

out-Fig 2. Net savings (cost) of an RV immunization program to the health care system by cost of vaccine (per dose, purchase and

administration). Base case, best case, and worst case sensitivity analyses were performed. At the threshold (break-even) vaccine cost, all

costs of the vaccine program are offset by reductions in the cost of RV illness (threshold cost: base case, $40; best case, $74; and worst case,

$17). With 70% vaccine efficacy, the threshold vaccine cost is $44.

(6)

patient medical visits and hospitalization, to a high of $664 million based on hospitalization costs alone,

using Texas Children’s Hospital as a reference.4 A

midrange estimate of $453 million, derived in 1990

from data on outpatient medical visits and

hospital-ization, approximates the 1993 estimate of medical

costs in our analysis ($427 million).#{176} The low

esti-mate from the Institute of Medicine study ($200

mil-lion) was derived by extrapolation from disease rates

for three small cohorts (together, fewer than 300

children) to the 16 million children in the United

States younger than 5 years of age at that time. This

estimate of 23 000 hospitalizations per year in the

United States was based on results from one study,

in which only one of the 126 children monitored was

hospitalized. The high estimate, $664 million, was

derived by extrapolation of rates of hospitalization for RV diarrhea (based on International

Classifica-tion of Diseases, 9th revision, discharge codes

and/or laboratory RV detection) of children

admit-ted to Texas Children’s Hospital during a 10-year

period (1979 to 1989), estimated at 110 000 hospital-izations per year in the United States. The midrange

estimate, $452 million, was derived from analysis of

national hospital discharge databases (1979 to

1985)2.3 and is supported by surveys of hospital

labo-ratory RV detections.4’ We used the midrange

esti-mate of 65 000 to 70 000 hospitalizations per year to

derive an estimate for the probability of

hospitaliza-tion for RV diarrhea, adjusted for increase in size of

the US birth cohort from approximately 3.7 million in

the mid-1980s to 4.1 million in 199322,42

Unlike other economic analyses of immunization programs, this analysis depends in part on measures of vaccine efficacy taken from trials that are still in progress for a vaccine that has not yet been approved

for widespread use. A limitation of this analysis is

the uncertainty about the efficacy of the RV vaccine in decreasing hospitalization rates. In field trials, RV vaccination decreased both the incidence of RV

diar-rhea (by half) and the severity of RV diarrhea,

mea-sured as decreased symptoms, physician visits

(50%), and duration of diarrhea (2.5 days compared

with 3.5 days). We have extrapolated from these

results concerning disease severity to estimate a 50%

decrease in hospitalizations among vaccinated

chil-dren with RV diarrhea and fewer days of work loss

by care givers. Because hospitalization accounts for

the largest proportion of the costs of RV-related

ill-ness to the health care system (estimated at 60% to

80% of medical costs), it would be important to

as-sess the impact of a vaccine on rates of

hospitaliza-tion when more data become available. Also, much

interest has been focused on increasing the use of

oral rehydration therapy to treat children with acute

diarrhea and to prevent dehydration that could lead

to hospitalization. To the extent that this effort is

successful, the rate of hospitalization also might be

decreased, and the impact of the vaccine might be

diminished.

Limited experience with these vaccines did not

permit us to assess any effect from “herd immunity,”

or the effect of any waning of protection over time.

The data on which we based our probability and cost

estimates are taken from conditions and studies in

the United States, and separate studies would need

to be undertaken to assess cost effectiveness of this

vaccine in developing countries.

This cost-effectiveness analysis joins several other

economic analyses of vaccines that have been

per-formed during the past two decades. Studies have

demonstrated that immunization programs are not

only cost effective, but they also reduce health care

costs in industrialized and developing

coun-tries.16’447 Analyses of vaccines for pertussis,

mea-sles-mumps-rubella, DTP, varicella, and others

con-sistently have shown positive ratios of reduction in

the costs of disease to the costs of the vaccination

program, with benefit-to-cost ratios ranging from 2:1

to 15:115,4857 Although RV vaccine efficacy rates

reach only 50% to 75% in the prevention of severe

diarrhea and physician visits, the dehydrating nature

and the high prevalence of RV diarrhea in the

pop-ulation younger than 5 years of age make

immuni-zation a cost-effective intervention.

RV vaccines currently being tested could be

li-censed by the Food and Drug Administration by

1996. Although a vaccine has not yet been released

for commercial use, this analysis suggests that the

use of an RV vaccine in the United States would be

cost effective as well as cost saving, and it supports

continued efforts to develop a vaccine for

wide-spread use. If a program of routine RV immunization is adopted in the United States, the cost effectiveness

of the vaccine should be reevaluated as new data

become available.

ACKNOWLEDGMENTS

We gratefully acknowledge the support provided by Phaedra

A. Shaffer, MPA, Prevention Effectiveness Activity, Epidemiology Program Office, Centers for Disease Control and Prevention; and Robert E. Lapp, PhD, Center for Health Economics and Policy Research, Blue Cross and Blue Shield Association, Chicago, IL; and the editorial assistance of John O’Connor.

REFERENCES

1. Kapikian AZ, Chanock RM. Rotaviruses. In: Fields BN, Knipe DM,

Chanock RM, et al, eds. Virology. 2nd ed. New York: Raven Press;

1990;2:1353-1404

2. Glass RI, Lew JF, Gangarosa RE, et al. Estimates of the morbidity and mortality from diarrheal diseases in American children. I Pediatr. 1991;

118:S27-S33

3. Ho M-S, Glass RI, Pinsky PF, et al. Rotavirus as a cause of diarrheal

morbidity and mortality in the United States. I Infect Dis. 1988;158:

1112-1116

4. Matson DO, Estes MK. Impact of rotavirus infection at a large pediatric

hospital. IInfect Dis. 1990;162:598-604

5. Clark HF, Borian FE, Plotkin SA. Immune protection of infants against rotavirus gastroenteritis by a serotype 1 reassortant of bovine rotavirus

WC-3. Jinfect Dis. 1990;161:1099-1104

6. Kapikian AZ, Flores J, Hoshino Y, et al. Rotavirus: the major etiologic agent of severe infantile diarrhea may be controllable by a “Jennenan” approach to vaccination. IInfect Dis. 1986;153:815-822

7. Kapikian AZ, Flores J, Vesikari T, et al. Recent advances in development of a rotavirus vaccine for prevention of severe diarrheal illness of

infants and young children. In: Mestecky J, ad. Immunology of Milk and

the Neonate. New York: Plenum Press; 1991:255-264

8. Bishop RF. Development of candidate rotavirus vaccines. Vaccine. 1993;

11:247-254

9. Kapikian AZ. Rhesus rotavirus-based human rotavirus vaccines and

observations on selected non-Jennerian approaches to rotavirus vacci-nation. In: Kapikian AZ, ed. Viral Infections of the Gastrointestinal Tract,

2nd ed. New York: Marcel Dekker, Inc; 1993:443-470

10. Bernstein DI, Glass RI, Rogers G, et al. Evaluation of rhesus rotavirus

(7)

monovalent and tetravalent reassortant vaccines in U.S. children.

JAMA. 1995;273:1191-1196

11. Dennehy PH. Safety and efficacy of an oral tetravalent rhesus rotavirus

vaccine (RRV-TV) in healthy infants. Pediatr Res. 1994;35(program issue):1052

12. Vesikari T. Clinical trials of live oral rotavirus vaccines: the Finnish experience. Vaccine. 1993;1 I :255-261

13. Glass RI, Gentsch J, Smith JC. Rotavirus vaccines: success by

reassort-ment? Science. 1994;265:1389-1391

14. Finkler SA. The distinction between costs and charges. Ann Intern Med.

1982;96:102-109

15. Hinman AR. Public health considerations. In: Plotkin SA, Mortimer EA

Jr. eds. Vaccines. Philadelphia: WB Saunders; 1988:587-611

16. Willems JS, Sanders CR. Cost-effectiveness and cost-benefit analysis of

vaccines. IInfect Dis. 1981;144:486-493

17. Mooney G, Crease A. Appendix C: Priority setting for health service efficiency: the role of measurement of burden of illness. In: Jamison DT,

ed. Disease Control Priorities in Developing Countries. World Bank; 1993:

731-740

18. Centers for Disease Control and Prevention. Comprehensive Childhood Immunization Initiative. Atlanta: National Immunization Program, Centers for Disease Control and Prevention; 1993

19. Centers for Disease Control and Prevention. Vaccination coverage of

2-year-old children-United States, 1991-92. MMWR. 1994;42:985-988

20. Drummond MF, Stoddard GL, Torrance GW. Methods for the Economic Evaluation of Health Care Programs. New York: Oxford University Press;

1987

21. Warner KE, Luce BR. Cost-Benefit and Cost-Effectiveness Analysis in Health

Care. Ann Arbor, MI: Health Administration Press; 1982

22. National Center for Health Statistics, Vital and Health Statistics.

Monthly Vital Statistics Report. Hyattsville, MD: National Center for

Health Statistics; 1993;42

23. Rodriguez WJ, Kim HW, Brandt CD, et al. Longitudinal study of

rota-virus infection and gastroenteritis in families served by a pediatric

medical practice: clinical and epidemiologic observations. Pediatr Infect Dis J.1987;6:170-176

24. Institute of Medicine. Prospects for immunizing against rotavirus. In:

Diseases of Inportance in the United States, Appendix 0: New Vaccine

Development: Establishing Priorities. Washington, DC: National Academy Press; 1985;1:410-423

25. Kapikian AZ. Viral gastroenteritis (clinical conference). JAMA. 1993;269:

627-630

26. Gurwith M, Wenman W, Hinde D, et al. A prospective study of

rota-virus infection in infants and young children. I Infect Dis. 1981;144:

218-224

27. Koopman JS. Patterns and etiology of diarrhea in three clinical settings.

Am JEpidemiol. 1984;119:114-123

28. Hjelt K, Krasilnikoff PA, Graubelle PC. Incidence of hospitalization and

outpatient clinical visits caused by rotavirus and non-rotavirus acute gastroenteritis. Dan Med Bull. 1984;31:249-251

29. Ho M-S, Glass RI, Pinsky PF, et al. Diarrheal deaths in American children: are they preventable? JAMA. 1988;260:3281-3285

30. Glass RI, Ho MS, Lew JF, et al. Cost-benefit studies of rotavirus vaccines

in the United States. In: Sack DA, Freij L, ads. Prospects for Public Health

Benefits in Developing Countries from New Vaccines Against Enteric Infec-tions. Goteborg: Sarec Conference Report; 1990:102-107

31. US Department of Commerce Economics and Statistics

Administra-tion, Bureau of the Census. Statistical Abstract ofthe United States 1993.

Washington, DC: Bureau of the Census; 1993. Table 163, Consumer price indexes, by major groups: 1960 to 1993 (p 482); and Table 756, Consumer price indexes of medical care prices: 1970 to 1993 (p 114)

32. Hodgson TA. State of the art of cost-of-illness estimates. Adv Health Econ Health Services Res. 1983;4:129-164

33. Scitovsky AA. Estimating the direct costs of illness. Milbank Q. 1982;60:

463-491

34. Avendano P, Matson DO, Long J, et al. Costs associated with office

visits for diarrhea in infants and toddlers. Pediatr Infect Dis J. 1993;12:

897-902

35. Gonzalez, ML, ed. Socioeconomic Characteristics of Medical Practice.

Chicago: American Medical Association; 1992

36. Crane M. What your colleagues are charging. Med Economics. 1992;69:

191-211

37. American Hospital Association. Data compiled from the American

Hospital Association 1991 Annual Survey of Hospitals. In: American Hospital Association, ed. Hospital Statistics. 1992-1993 ed. Chicago: American Hospital Association; 1992

38. Miller W. Physician charges in the hospital: exploring episodes of care for controlling volume growth. Med Care. 1992;30:630-645

39. National Center for Health Statistics. Vital and Health Statistics. Detailed Diagnoses and Procedures, National Hospital Discharge Survey. Hyattsville, MD: National Center for Health Statistics, Centers for Disease Control

and Prevention; 1994. Computer database

40. Centers for Disease Control and Prevention. A Practical Guide to Preven-tion Effectiveness. Decision and Economic Analysis. Atlanta: Centers for

Disease Control and Prevention; 1994

41. Brandt CD, Kim HW, Rodriguez JO, et al. Pediatric viral gastroenteritis during eight years of study. 1Clin Microbiol. 1983;18:71-78

42. Department of Health and Human Services. International Classification of

Diseases. 9th revision. Clinical Modification. 4th ed. Washington, DC: US

Department of Health and Human Services; 1991. DHHS publication 91-1260.

43. Peter C. Childhood immunizations. N Engl I Med. 1992;327:25:

1794-1800

44. Ginsberg GM, Tulchinsky TH. Costs and benefits of a measles inocula-tion of children in Israel, the West Bank, and Gaza. JEpidemiol

Commu-nity Health. 1990;44:272-280

45. Musgrove P. Cost-benefit analysis of a regional system for vaccination against pneumonia, meningitis type B, and typhoid fever. Bull Pan Am Health Organ. 1992;26:173-189

46. Ponnighaus JM. The cost/benefit of measles immunization: a study from southern Zambia. JTrop Med Hyg. 1980;83:141-149

47. Shepard DS, Sanoh L, Coffi E. Cost-effectiveness of the expanded pro-gramme on immunization in the Ivory Coast: a preliminary assessment.

Soc Sci Med. 1986;22:369-377

48. Cochi SL, Broome CV, Hightower AW. Immunization of US children

with Hemophilus influenzae type b polysaccharide vaccine. JAMA. 1985;

253:521-529

49. Hatziandreu E, Palmer CS, Brown RE, et al. A Cost Benefit Analysis of the Diphtheria-Tetanus-Pertussis Vaccine. Washington, DC: Division of

Im-munization, National Center for Prevention Services, Centers for Dis-ease Control and Prevention. Human Affairs Research Center, Battelle

Medical Technology Assessment and Policy Research Center; 1993.

Draft report

50. Hatziandreu E, Brown RE, Halpem MT. A Cost Benefit Analysis of the Measles-Mumps-Rubella Vaccine. Washington, DC: Division of Immuni-zation, National Center for Prevention Services, Centers for Disease

Control and Prevention. Human Affairs Research Center, Battelle Mad-ical Technology Assessment and Policy Research Center; 1993. Draft report

51. Hinman AR, Koplan JR. Pertussis and pertussis vaccine: further analy-sis of benefits, risks and costs. Dev Biol Stand. 1985;61:429-437

52. Hinman AR, Koplan fl’ Pertussis and pertussis vaccine: reanalysis of

benefits, risks, and costs. JAMA. 1984;251:3109-3113

53. Koplan JP, Schoenbaum SC, Weinstein MC, at al. Pertussis vaccine-an analysis of benefits, risks, and costs. N Engl JMed. 1979;301:906-911

54. Lieu TA, Cochi SL, Black SB, at al. Cost-effectiveness of a routine

varicella vaccination program for US children. JAMA. 1994;271:375-381

55. Margolis HS, Schatz CC, Kane MA. Development of recommendations for control of hepatitis B virus infections: the role of cost analysis.

Vaccine. 1990;8(suppl):S81-585

56. White CC, Koplan JP, Orenstein WA. Benefits, risks and costs of immu-nization for measles, mumps and rubella. Am J Public Health. 1985;75:

739-744

57. Willems JS, Sanders CR, Riddiough MA, et al. Cost effectiveness of

vaccination against pneumococcal pneumonia. N EngI IMed. 1980303:

553-559

(8)

1995;96;609

Pediatrics

Jean C. Smith, Anne C. Haddix, Steven M. Teutsch and Roger I. Glass

States