Feikema SM, Klevens RM, Washington ML, Barker L. Extraimmunization Among US Children. JAMA. 2000;283(10):1311-1317. doi:10.1001/jama.283.10.1311
Author Affiliations: Medical Management Department, Children's Healthcare of Atlanta (Ms Feikema); and Assessment Branch (Dr Klevens) and Statistical Analysis Branch (Drs Washington and Barker), Data Management Division, National Immunization Program, Centers for Disease Control and Prevention, Atlanta, Ga.
Context Little is known about the extent of extraimmunization, ie, vaccine doses
given in excess of the recommended schedule, and whether it should be a public
Objectives To determine the extent and cost of extraimmunization in children and
to identify its associated factors.
Design, Setting, and Participants United States 1997 National Immunization Survey, in which telephone
interviews were conducted with parents of 32,742 19- to 35-month-old children
and vaccination histories were collected from health care providers for 22,806
of these children (overall response rate, 68.5%). Estimates were weighted
to represent the full sample.
Main Outcome Measures Frequency of extraimmunization compared by vaccine type as well as with
adequate immunization; factors associated with extraimmunization; and vaccine
and visit costs associated with extraimmunization.
Results Frequency of extraimmunization was less than 5% for each vaccine considered
except poliovirus (14.1%). Overall, 21% of children were extraimmunized for
at least 1 vaccine vs 31% underimmunized for at least 1 vaccine. In a multivariate
model, the strongest contributors to extraimmunization were having more than
1 immunization provider (odds ratio [OR], 2.8; 95% confidence interval [CI],
2.4-3.2) and having multiple types of providers (eg, private and public health
department; OR, 2.0; 95% CI, 1.6-2.4). Children seen only in public health
department clinics were significantly less likely to be extraimmunized (OR,
0.3; 95% CI, 0.2-0.3). Annual costs associated with extraimmunization for
this cohort of children were estimated conservatively at $26.5 million.
Conclusions These data indicate that extraimmunization can be costly. The challenge
is to reduce extraimmunization without interfering with more important efforts
to combat underimmunization. Improvements in immunization record keeping and
sharing practices may help reduce extraimmunization.
Achieving and sustaining high immunization rates among US children is
an important public health goal that has been vigorously pursued in the last
decade. Extensive research on the factors that contribute to underimmunization1- 3 has led to a variety
of interventions, including provider education to assess practice coverage
levels and to reduce missed opportunities,4,5
mass media campaigns to educate parents about the importance of immunization,6 and efforts to reduce financial and other barriers
to immunization.4 National immunization coverage
levels are now higher than ever before.7
However, little attention has been paid to extraimmunization, ie, vaccine
doses given in excess of the recommended schedule. One possible consequence
of aggressive immunization interventions and the use of untargeted educational
campaigns is the unintentional administration of extra vaccine doses. Other
factors that may contribute to extraimmunization include suboptimal record-keeping
practices and the enforcement of minimum ages and intervals for vaccine doses
as a requirement for school entry.
When complete provider vaccine history records and the parent-held "shot
card" are lacking at the time of a health care visit, the physician may be
guided by recommendations to give age-appropriate immunizations.8,9
Such action may result in the administration of extra, unneeded doses. Extra
vaccine doses also may be required to "fix" earlier errors in vaccine spacing.
If 2 vaccine doses are given too close together or if a vaccine dose is given
earlier than the minimum age, the dose may be appropriately repeated. The
extent of such errors has not been well documented, but a small study in 4
Los Angeles, Calif, public health clinics found that 22% of the children studied
had received "inappropriately timed" immunizations by ages 25 to 36 months.10
Little is known about the effects of receiving extra vaccine doses.
The Advisory Committee on Immunization Practices (ACIP) recommends that children
not receive more than 6 doses each of diphtheria and tetanus toxoids before
the age of 7 years because extra doses may cause adverse local or systemic
effects.11,12 No limits are recommended
for other routine childhood vaccines because there is no similar evidence
of harm from extra doses,13 although it has
been postulated that extra doses of some vaccines are more likely to induce
hypersensitivity to vaccine components.14 However,
even if medically safe, extraimmunization is inefficient and unnecessarily
costly, and thus, undesirable.15
A few small studies have examined the question of extraimmunization.
A population-based study of 187 children born in Dallas, Tex, in 1986 and
1987 found that by age 72 months 18% had received 1 or more extra vaccine
doses.15 Another population-based study of
2048 children born in 1992 and 1993 and living in one Minnesota county found
that 5% of the children had been given extra immunization doses by age 24
months.16 These studies were conducted when
the recommended immunization schedule was less complex.
This study sought to determine the extent and associated costs of extraimmunization
among US children aged 19 to 35 months and to identify factors associated
with extraimmunization in the population under study. The findings can be
used to determine if extraimmunization should be a public health concern and,
if so, to identify ways to prevent it.
We analyzed data from the 1997 National Immunization Survey (NIS), a
representative survey of children aged 19 to 35 months. Information is collected
continuously in 2 steps. First, a random-digit dialing sample of telephone
numbers in each of the 50 states and in 28 selected urban areas is generated.
Approximately 1.6 million telephone numbers are contacted annually to reach
a targeted 440 age-eligible children in each study area, for a total target
sample size of 34,320. A screening questionnaire is administered to adult
respondents to identify households with 19- to 35-month-old children. In households
with an eligible child, a parent is interviewed to collect demographic information,
the child's immunization history, and consent to contact the child's immunization
provider(s). In the second step, the child's immunization history is requested
from the identified immunization provider(s). Where information is received
from more than 1 provider for the same child, duplicate values are eliminated
and a composite vaccination history is created. Only children for whom provider
information was obtained were included in this analysis.
Adjustment weights are calculated for each child to adjust for households
with multiple telephone numbers, household nonresponse, and lack of a telephone
through poststratification using the National Health Interview Survey. Then,
the sample is adjusted to reflect US Census Bureau population totals by race/ethnicity,
mother's education, and age of the child. Finally, adjustments for infant
mortality, immigration, and migration rates are conducted based on natality
files from the National Center for Health Statistics.17,18
Adequate immunization was defined according to the 1995 ACIP recommended
childhood immunization schedule, before the 1996 recommendation for varicella
Intervals between doses were not considered; only the number of doses received
was counted. Between birth and 18 months of age, 14 or 15 vaccine doses are
recommended: 3 doses of hepatitis B vaccine, 4 doses of diphtheria and tetanus
toxoids and pertussis vaccine (DTP/DTaP [acellular pertussis]), 3 or 4 doses
of Haemophilus influenzae type b (Hib) vaccine, 3
doses of poliovirus vaccine, and 1 dose of measles-containing vaccine.19 Extraimmunized children were those who received more
than the number of recommended doses for any vaccine(s). Underimmunized children
were those who received fewer than the number of recommended doses for any
We first calculated the frequency of adequate immunization, extraimmunization,
and underimmunization by vaccine type. Then we calculated the frequency of
children who were adequately immunized, extraimmunized (without being underimmunized),
underimmunized (without being extraimmunized), and the remainder (both extraimmunized
and underimmunized). Children who were underimmunized for any vaccine were
then removed from the sample.
Further analyses compared only adequately immunized with extraimmunized
children. First, we evaluated the frequency of available child, family, and
provider characteristics in a bivariate analysis. Then, using the characteristics
that were significantly different (P<.05) in the
bivariate analysis, we modeled the association of being either extraimmunized
or adequately immunized using logistic regression. Because we were seeking
to build a descriptive model that simultaneously controlled for all variables
rather than a "best fit" model, we included all of the significant variables
from the bivariate analysis. The initial frequency calculations and bivariate
analysis were conducted using SAS software, version 6.12 (SAS Institute, Cary,
NC). The logistic regressions were conducted using SUDAAN, release 7.5.2 (Research
Triangle Institute, Research Triangle Park, NC).
To estimate the cost of the extra vaccine doses administered, we generated
a frequency distribution for each vaccine type, assigned a price from the
1997 Centers for Disease Control and Prevention vaccine price list (unpublished
data, January 15, 1997), and calculated the total vaccine cost. We assumed
that all extra vaccines administered in a public setting were purchased by
public funds. The number of extra vaccines administered in other settings
was weighted to achieve an overall distribution of 61% publicly and 39% privately
purchased vaccine doses, approximating the 1997 US funding distribution of
vaccine doses (R. Snyder, Centers for Disease Control and Prevention, written
communication, June 15, 1999). Since we were not able to determine particular
vaccine brands or combination vaccines administered, we used prices for noncombination
vaccines only. Where more than 1 product was available, we assumed the products
were equally distributed and calculated average public and private prices.
We priced the following vaccine products: hepatitis B, pediatric dosage (average
of 2 brands); DTP/DTaP (average of 3 brands); Hib (average of 3 brands); inactivated
poliovirus vaccine and oral poliovirus vaccine; and measles, mumps, and rubella
vaccine. Where poliovirus vaccine type was unknown, we assumed that the oral
formulation was administered.
It is estimated that for every 95 vaccine doses used, 5 doses are wasted.22 We assumed an equal distribution of waste among vaccines.
To calculate vaccine waste, we divided the total number of extra vaccine doses
administered by 95% to determine the estimated number of extra purchased vaccine
doses. We then subtracted the extra doses administered from the extra doses
purchased and multiplied the result by the average cost per extra vaccine
Finally, we estimated the number of extra visits made to receive extra
vaccine doses. An extra visit was defined as a visit to a provider where only
1 or more extra vaccine doses were received and no recommended, "nonextra"
doses were received. Using 1994 figures and a 5% discount rate, we calculated
the combined average visit cost and the cost to administer a vaccine in 1997
as $11.58 for public clinics and $20.26 for private offices.23
For visits to other provider types, we averaged the public and private visit
In the 1997 NIS, information on 32,742 children was collected from parents
(93.8% interview completion rate) and provider information was collected for
22,806 (70%) of these, for an overall response rate of 68.5% (Table 1). No differences were observed between children with and
without provider data by sex or age; however, significant differences were
noted between groups by race, household size, mother's education, mother's
marital status, household income, and parent-reported 4:3:1 immunization coverage
By vaccine, the frequency of extraimmunization was less than 5% for
all vaccines except poliovirus, for which 14.1% of children were extraimmunized
(Table 2). Overall, about half
(53%) the children in this cohort were adequately immunized, 27% were underimmunized
for at least 1 vaccine but were not extraimmunized for any vaccine, 17% were
extraimmunized for at least 1 vaccine but were not underimmunized for any
vaccine, and 4% were both underimmunized and extraimmunized for at least 1
vaccine. Therefore, 21% were extraimmunized for at least 1 vaccine.
In the bivariate analysis, extraimmunized children were more frequently
male. Significant differences also were noted between adequately immunized
and extraimmunized children by age, race/ethnicity, household income, geographic
region, and provider facility type (Table
3). Children whose parents reported vaccination history from a shot
card were more frequently extraimmunized, as were children with more than
1 immunization provider. The following family characteristics showed no significant
differences: mother's age and education, number of children in the household,
and birth order.
Results of the multivariate logistic model (Table 4) indicate that the available child, family, and provider
characteristics explain approximately 11% of the variance between adequately
immunized and extraimmunized children and that the model is significantly
better than random (χ2 = 1865.71; P<.001).
Hispanic or Asian/Pacific Islander race/ethnicity and older age (30-35 months)
were the only child or family characteristics associated with extraimmunization
that were statistically significant in the multivariate model. Children with
more than 1 provider were almost 3 times more likely to be extraimmunized
than children with only 1 provider (odds ratio [OR], 2.8; 95% confidence interval
[CI], 2.4-3.2). Children who saw multiple types of providers (eg, health department
and private) were twice as likely as those who saw only private providers
to be extraimmunized (OR, 2.0; 95% CI, 1.6-2.4). Children immunized only in
health department public clinics (OR, 0.3; 95% CI, 0.2-0.3) and only in hospitals
(OR, 0.6; 95% CI, 0.5-0.9) were less likely to be extraimmunized.
The cost analysis found a total of 1.8 million extra vaccine doses administered
nationwide at an average cost of $9.90 per dose (Table 5). This represents an excess cost of approximately $18.2
million. In addition, an estimated 96,795 vaccine doses were wasted at a cost
of almost $1 million and an extra 412,569 clinic/office visits were made to
receive only extra vaccine doses, at a cost of $7.3 million. Annual costs
associated with extraimmunization for this cohort of children were estimated
conservatively at $26.5 million.
With about 1 in 5 (or 900,000 of 3.9 million) US children receiving
at least 1 extra vaccine dose by age 19 to 35 months, extraimmunization is
clearly widespread and consequently quite costly. The extent of extraimmunization
identified in this study represents a national excess cost of at least $26.5
million for the 19- to 35-month-old population. This is a conservative estimate
that does not consider the cost of vaccine storage, handling, and distribution;
parents' travel time; loss of wages; treatment for adverse events (if any)
associated with extraimmunization; or other indirect costs.
While we cannot expect extraimmunization to be eliminated completely
because extra doses are sometimes necessary to ensure that a child is fully
immunized, reducing the extent of extraimmunization is desirable. It is particularly
important that extra doses of diphtheria and tetanus toxoids be avoided to
prevent potential adverse events.
Understanding the factors that contribute to extraimmunization will
be important in reducing its incidence. In our multivariate logistic regression
model, individual and family characteristics offered little explanatory power.
Instead, provider characteristics were most strongly associated with extraimmunization.
In particular, children with more than 1 immunization provider were more likely
to be extraimmunized. The likelihood of extraimmunization also varied by type
of provider: whereas 56% of children seen by multiple provider types were
extraimmunized, 8% of those seen only in health department clinics received
extra vaccine doses.
Another study showed similar differences in extraimmunization rates
by providers, ranging from 5% for most providers to 33% for children ever
seen in a particular system of public clinics. The authors were able to determine
that the records of the public clinic system in question often did not reflect
prior immunizations received at that clinic or elsewhere.15
Lack of ready access to complete and accurate immunization records seems
to be the likeliest explanation here as well. When children see new providers
or are referred for immunizations outside their source of primary care, particularly
if the providers are of different types (eg, public health department clinic
vs private practice), their immunization records may not follow them. One
solution is to use parent-held shot cards.4
Community- and state-based immunization registries24
represent an alternative to relying on shot cards. These computer databases
keep track of individuals' immunization histories and are accessible from
providers' offices. In addition to facilitating record sharing between providers,
immunization registries can help providers avoid vaccine spacing errors by
determining when shots are due. Registries also are useful for implementing
strategies shown to be effective at reducing underimmunization. They can be
used to assess coverage levels in practices and to generate reminder and recall
notices when immunizations are due or late.
While provider characteristics are important in explaining extraimmunization
in this study, they, along with the individual and family characteristics
that were significant in the bivariate analysis, explained only a small percentage
of the variation between extraimmunized and adequately immunized children.
Clearly, other factors must contribute to extraimmunization. Further research
on immunization record-keeping practices across different provider types may
offer some insight.
When examined by antigen, poliovirus vaccine is the largest contributor
to extraimmunization. It is the only antigen for which the percentage of extraimmunized
exceeds the percentage of underimmunized children. There are several possible
explanations for this discrepancy. First, the recommended time frame for administering
the third dose of polio vaccine spans a full year and overlaps with both the
third and fourth doses of DTP/DTaP vaccine. Providers may unwittingly administer
an extra polio dose with the fourth dose of DTP/DTaP because they are accustomed
to administering these 2 vaccines together. Second, the predominant oral formulation
administered to these children is easier to give than a shot and is relatively
inexpensive. As recommendations for the increasing use of inactivated poliovirus
vaccine25 are adopted over time, the percentage
of children extraimmunized for poliovirus may decrease. This trend should
continue to be monitored.
Other factors that might explain the differences in extraimmunization
by vaccine type include the number of recommended doses, the age ranges for
which the vaccine is recommended, the complexity of the schedule, whether
changes have been made in the schedule, how long the vaccine has been on the
recommended schedule, and the number of types of vaccine available, including
combinations, and whether they follow the same or different schedules. Increased
complexity, changes, and multiple choices may lead to confusion and increase
the chance of extra doses. More research is warranted to elucidate the impact
of these factors, to find other explanatory factors, and then to identify
strategies for reducing extraimmunization. In the meantime, vaccine manufacturers
and authoritative bodies should consider these factors and their potential
impact on extraimmunization when developing new vaccines and making changes
in the recommended schedule.
While we are confident in our estimates of the extent of and costs attributable
to extraimmunization, some potential limitations to our study should be considered.
If our methods for resolving duplicate entries were inadequate when creating
composite immunization histories, the extent and cost of extraimmunization
might be overestimated. However, because most children have only 1 provider,
the potential impact of this limitation is small. Extraimmunization may be
slightly overestimated for hepatitis B vaccine and underestimated for Hib
vaccine because multiple products and vaccine schedules are acceptable.26,27
Excluding children without provider information may limit the generalizability
of this study, since those children had significantly different demographic
characteristics from children with provider information and, by parent report,
were less likely to be up-to-date with the 4:3:1 vaccination series. We also
do not have information on the characteristics of children who were both underimmunized
and extraimmunized. In part because estimates were weighted to be representative
of all US children aged 19 to 35 months, these limitations should not substantially
affect our estimate of extraimmunization.
In our cost calculations we had to make assumptions regarding vaccine
type administered, payment source, and "extra" visits. However, we still believe
our cost results to be a conservative estimate.
For the first time, to our knowledge, the extent of extraimmunization
has been estimated on a national scale and found to be substantial and costly.
The challenge now will be learning how to reduce extraimmunization without
interfering with the more important efforts to combat underimmunization and
achieve adequate immunization.