CDC's National Immunization Program: Methods Used to Identify Pockets of Underimmunized Children Not Evaluated

Published by the Government Accountability Office on 1997-08-01.

Below is a raw (and likely hideous) rendition of the original report. (PDF)

      United States
GAO   General Acconnting Office
      Washington, D.C. 20648

      Health, Education and Human Semites Division

      August 1, 1997

      The Honorable Dale Bumpers
      The Honorable Ron Wyden
      United States Senate

      The Honorable Scott Klug
      House of Representatives

      Subject:   CDC’s National Immunization Program: Methods Used to Identifv
                 Pockets of Under-immunized Children Not Evaluated

      The Centers for Disease Control and Prevention (CDC) estimates that three-
      fourths of Z-year-old children in America are up-to-date for the basic series of
      childhood immunizations. However, some specific geographic areas and
      communities are at higher risk of disease outbreaks because they harbor
      concentrations of children who have not received timely immunizations.1 The
      continued existence of such pockets of underimmunized children, sometimes
      called pockets of need, is evidenced by measles outbreaks in 1996 in Alaska
      and Utah that included young children2

      You asked us to determine what methods CDC uses to identify pockets of
      underinununized preschool children and what is known about the effectiveness
      of methods in use. To answer these questions we interviewed CDC and state
      public health officials and other experts in the field; reviewed the relevant
      literature; and, in&me and December 1996, conducted telephone surveys of
      state immunization program managers. We carried out our work in accordance

      ‘Vaccines for Children: Reexamination of Program Goals and Imnlementation
      Needed to Ensure Vaccination (GAO/PEMD-95-22, June 15, 1995).
      2“Measles Outbreak Among School-Aged Children-Juneau, Alaska, 1996,”
      MMWR 1Morbiditv and Mort&tv Weeklv Repor& Centers for Disease Control
      and Prevention, Vol. 45, No. 36 (Sept. 13, 1996), pp. 777-80, and “Washington
      County Measles Outbreak Is Over,” Utah Department of Health (Salt Lake City,
      Aug. 7, 1996).
                                  GAOAEHS-97-136R    Pockets   of Underimmunization
with generally accepted government auditing standards and completed it in May


CDC officials view identification of pockets of under-immunized children as a
state responsibility rather than a federal one; therefore, CDC does not directly
implement methods for identifying pockets of underimmunized 2-year-old
children. CDC defines pockets of need as specific geographic areas within
state or urban jurisdictions that contain large numbers of 2-year-old children
who are either under-immunized or at risk of under-immunization. CDC’s
National Immunization Program instead focuses on increasing the overall
immunization rate for the basic series and reducing disease. ln fiscal year 1997
,tidance for grant recipients, CDC, for the first time, directed states to develop
plans for identifying pockets of need. At that time, CDC suggested two
identification methods: (1) measurin g immunization coverage rates directly or
 (2) using surrogate measures that may indicate low coverage rates.

We found that neither CDC nor the states have assessed how well these or
other methods identify pockets of under-immunized children or children at risk
for underimmunization, although some appraisals have been published about
using these methods to measure or improve coverage rates. Nevertheless,
almost every state is initiating a new activity, or continuing a previous one, to
identify pockets of need. For example, 20 states are assessing immunization
coverage rates of children receiving care in public health clinics as their
primary method for identifying pockets; 15 other states are using survey
techniques such as door-to-door surveys.


Since 1963, CDC has provided grants to state and local health agencies for
planning, developing, and operating childhood immunization programs and,
beginning in 1992, delivering vaccines. These grants for immunization programs
and vaccines are intended to assist state and local health agencies in providing
services, information, outreach and community mobilization programs,
education and training, and disease surveillance and investigation. CDC’s
National Immunization Program now makes grants to each state and 28 urban
areas. These grants are intended to implement each state’s and area’s own
immunization action plan for preventing and controlling vaccine-preventable
diseases. For fiscal year 1998, CDC anticipates spending approximately $287.8
million for the immunization grant program, commonly referred to as the 317

2                       GAOLEEHS-97-136R        Pockets   of Underimmunization

program, after section.317 of the Public Health Service Act (42 U.S.C. 247b), as
amended, which authorizes project grants for preventive health services.3

In 1990, the Department of Health and Human Services (HHS) established
immunization and disease reduction goals for the end of this decade. Two of
these goals are immunizin g at least 90 percent of &i&en 2 years of age and
younger with the appropriate basic immunization series and eliminating, or
greatly reducin g, indigenous cases of preventable diseases by the year 2000.4 Jn   -.
 1997 congressional testimony, the Director of the National Immunization
Program noted that to ensure that the national coverage goal is reached,
pockets of need must be identified and activities targeted to improve coverage
in the most hard-to-reach populations5

One method CDC uses to estimate immunization coverage rates is the National
Immunization Survey @IS), a telephone household survey conducted by
random-digit dialing and including some verification by provider records.
According to NIS results for 1995, the national immunization coverage rate is 76
percent for the basic series; states’ coverage rates range from 66 to 89 percent,
with 38 states not statistically distinguishable for the NIS’s national coverage

3For fiscal year 1998, the Administration proposes to reduce the amount of state
grants by $14.4 million because the states hold unobligated funds from awards
in previous years.
‘?‘his basic series is the 4:3:1 series, where children receive four doses of
diphtheria-tetanus-pertussis vaccine, three doses of poliovirus vaccine, and one
dose of any measles-containing vaccine. For statiticsil purposes, CDC modified
the objective and tracks children aged 19 to 35 months as 2-year-olds. For
some limitations of this definition, see &P. Goldstein and R.S. Daum, “Counting
Immunisations,” Lance& Vol. 344, No. 8916 (1994), pp. 144-45, and V. Dietz and
others, “‘Vaccination Coverage in the USA [Letter],” Lancet, Vol. 334, No. 8934
(1994), pp. 143940.
5W.A . Orenstein, “Statement before the Subcommittee on Public He&h and
Safety, Senate Committee on Labor and Human Resources, U.S. Senate,” May 6,

3                      GAO/HEHS-97-136R       Pockets   of Underimmunization
rate;6 and coverage rates for the 28 urban areas range from 62 to 87 percent.’
In a recent report, we noted that the NIS survey may lend a false sense of
security by obscuring the existence of substantial pockets of under-immunized
children.s For example, a household survey of central and southeast Seattle
found an immunization coverage rate of 57 percent, in contrast to the 79
percent reported by the NIS for the King County area incorporating Seattle.

The risk of outbreaks of vaccine-preventable diseases is highest where
concentrations of nonimmunized or underin-ununized children reside. In these
geographic areas, the probability is higher that an infected child will come into
contact with and infect a susceptible child. This situation contrasts to one in
which an overall lower immunization rate occurs, but with susceptible children
spread over a larger area. In those instances, a lesser public health threat
exists because of the lower probability that an infected child will come in
contact with and infect another susceptible child.

60ur comparisons used the state and national percentage rates and confidence
intervals published by CDC. We found 8 states higher and 4 states lower than
the national rate. See “National, States, and Urban Area Vaccination Coverage
Levels Among Children Aged 1935 Months-United States, January-December
 1995,” MMWR, Vol. 46, No. 8 (Feb. 28, 1997), pp. 176-82.
7CDC also estimates national immunization rates using the National Health
Interview Survey, a nationally representative, face-to-face household survey.
This survey, last published in 1994, estimated that 73 percent of 2-year-old
children are up-to-date on their basic series of vaccinations, 2 percent fewer
than the comparable NIS estimate.

‘For this and other limitations of the survey, see CDC’s National Immunization
Survev: Methodological Problems Limit Survev’s Utitv (GAOK’EMD-96-16, Sept.
19, 1996).

4                       GAOEIEHS-97-136R        Pockets   of Underimmunization


According to CDC officials, identifying pockets of underimmunized children is
the responsibility of the states and therefore CDC does not, itself, identify
them.g Furthermore, the Director of the National Immunization Program said,
CDC immunization efforts should be directed toward improving overall
immunization coverage rates in states and urban areas with low rates. In his
view, a state with a low coverage rate throughout has more need than a state
with a higher overall coverage rate even if the latter state has, within it, specific
areas with low coverage rates. In general, CDC officials do not believe that
focusing all efforts on pockets of need is a sufficient strategy to increase
overall coverage rates.

Although CDC officials believe that focusing all efforts only on pockets of need
is an insufficient strategy for improving overall immunization coverage rates,
they consider identiIication and elimination of these concentrations as useful
and necessary for meeting coverage goals. In August 1996, for fiscal year 1997,
CDC added targeting pockets of need to its list of requirements for section 317
project grar~ts.~~CDC required each state and each of the 28 urban areas to
include in its application a separate plan to identify geographic areas in which
subpopulations are (1) at high risk for under-immunization or (2)
underimmunized and at high risk for vaccine-preventable disease.ll These

‘According to our discussion with Department of Health and Human Services
@lHS) officials, HHS’s statement on pockets of need, which follows, was in
error: “For the first time in FY 1996, CDC also will be able to help States target
resources to pockets of need because of the new National Immunization
Survey” (HHS, Office of the Assistant Secretary for Management and Budget,
U.S. Denartment of Health and Human Services: The F’iscal Year 1997 Budget
(Washington, D.C.: Mar. 19, 1996), p. 28.
l”Earlier in September 1995, the Senate Committee on Appropriations directed
CDC to develop and implement a strategy for identifying and targeting
immunization resources for high-risk populations (see S. Rept. 145, 104th Cong.,
 1st sess., 1995, accompanying the fiscal 1996 HHS appropriation, p- 53). In
comments responding to a draft of this correspondence, CDC stated that in
most years since 1991, grant guidance discussed the need to address such
“CDC uses a broad definition for pockets of need because different geographic
regions might be faced with merent problems.

5                        GAOIEEHS-97-136R        Pockets   of Underimmhation
geographic areas may consist of groupings of census tracts, distinct
communities, or BBS-designated shortage areas for health providers. CDC
pointed out that all states and the 28 urban areas are likely to include such

CDC’s grant guidance said that to identify defined geographic areas, states
should measure immunization coverage rates directly, using methods such as
provider-based assessments, community-based household cluster surveys,
random household telephone surveys, and birth certificate surveys. If coverage
data from these survey methods are unavailable, CDC recommended using
surrogate characteristics for demographic, sociological, or epidemiological
measures, such as high proportion of racial and ethnic minority subpopulations,
high poverty rate, low education status of parents, high population density, and
high incidence of vaccine-preventable diseases.

During the current fiscal year, CDC has facilitated information sharing among a
group of traditionally underserved urban areas where pockets of need probably
exist.12 In addition, CDC is supporting an immunization demonstration project
in two Chicago public housing projects that it considers likely pockets of need.


Our survey of state immunization program managers in December 1996 found
that states are using, or are beginning to use, a wide array of primary methods
to identify pockets of underinunum.zed children or children at risk for
under-immunization, including methods that CDC has recommended. In fact,
most states were already engaged in attempts to identify pockets before CDC
issued the grant guidance-l3 For the primary method of identifying pockets, we
found that 20 states are using provider-based assessments of coverage rates for
children who are seen in public health clinics and other public programs; 7
states are using retrospective school surveys; 6 states are using surrogate

121nlate 1996 and &riy 1997, from 11 of the urban areas with large numbers of
underimmunized children, CDC convened two meetings of immunization
program managers to share local plans for increasing immunization rates.
“In our June 1996 survey of state immunization program managers, 42 states
were engaged in some activity to identify pockets of under-immunization. When
we started our surveys, CDC did not have a definition, so we defined pockets of
need to the managers as concentrations of preschool children in which
immunization rates are much lower than the average in a state.

6                       GAOIHEHS-97-136R        Pockets   of Underinumnhation

characteristics; 6 states are using immunization registries; 5 states are using
birth certificate surveys; 3 states are using door-to-doqr surveys; and 3 states
are using no method at all. Many state managers use several methods for
identifying pockets of need, they said-l4 (See enclosure I for breakdown of the
states and the methods used.)

The following is a brief description of the methods states are using to identify
pockets of need:

      Provider assessments focus on improving the immunization coverage
      rates among 2-year-old children either attending public health clinics or
      enrolled in clinics serving clients of the Special Supplemental Nutrition
      Program for Women, Infants, and Children (WE). Additionally, as the
      number of children receiving care in public health clinics dwindles,
      states are considering expanding their assessments to include private
      physician offices. Immunization program officials believe this method is
      effective in increasing overall coverage rates.

      Population surveys-such as door-to-door and face-to-face household
      surveys, as well as random household telephone surveys-are intended to
      measure immunization coverage rates in specific neighborhoods. This
      method also identifies specific underimmunized children. However, this
      method is labor intensive and may not work well in areas with a
      transient population.

      Follow-back surveys for birth certificates are another form of population
      survey. When doing these, surveyors select a sample of birth certificates
      for children born 2 years earlier, then-using available information such
      as the mother’s name, telephone number, address at the time of
      childbirth, motor vehicle licensing records, and hospital and public
      health department records-trace and locate each child. After locating
      children, surveyors obtain the children’s immunization histories. As with
      other population surveys, tracing children in transient family situations
      can be difficult and time-consuming.

      School surveys to determine, retrospectively, immunization status at age
      2 are based on school record examinations of children who are entering
      kindergarten and first grade. Using these school records, public health

14Wereport here only the method that each state manager indicated is the
primary method.

7                      GAOAEHS-97-136R         Pockets   of Underimmunization
      officials look backward to determine the immunization status of the
      children when they were younger, typically when they were 2 years old.
      Following trends of underimmunization, some states are recognizing
      some neighborhoods or school districts as persistent pockets of need.

      Surrogate characteristics are demographic, social, and epidemiological
      attributes thought to be associated with the presence of underimmunized
      children. These might include population density, race and ethnicity,
      income, or incidence of vaccine-preventable disease. After identifying
      geographic areas characterized by the chosen surrogates, public health
      officials would target interventions directed toward increasing the
      immunization coverage rate of ch.ildren. However, other pockets exist,
      as evidenced by the measles outbreaks in Utah and Alaska, where the
      usual surrogate characteristics probably would not have helped to
      identify pockets of under-immunized children.

      Immunization registries track the immunization status of children from
      birth or their first encounter with a public or private provider that
      participates in registries. Registries make available to participating
      providers, by telephone or computer, current information on a child’s
      immunization status, based on provider records of children residing in or
      seeking care in a defined geog.raDhic area Registries are also used as
      reminder systems to notify parents and providers when immunizations
      are due.15 However, the usefulness of these tools to identify pockets of
      underimmunization depends on the extent to which (1) such systems
      include all children in a selected age group within a given community or
      service area and (2) public and private providers submit information on

For a more detailed discussion of our assessment of the strengths and
limitations of states’ primary methods of identifying pockets of
underimmunization, see enclosure II.

With the exception of using surrogate characteristics, none of the methods
states are using was developed to identify pockets of underimmunized children
or children at risk for underimmunization. Rather, these methods were

151nmost states, at least one imnnmization registry project is under way, many
with the financial support of private foundations. Statewide registries in
Arizona, Mississippi, North Dakota, and Rhode Island report that they have
established databases of 75 percent or more of their target populations.

8                      GAOLEEHS-97-136R       Pockets   of Underimmunization

developed to measure immunization coverage rates or to motivate changes in
provider practices. According to the Director, National Immunization Program,
CDC (1) has not conducted any evaluations of the effectiveness of these
methods for identifying pockets of need and (2) is unaware of any other such
evaluations. In late 1996, CDC made a S-year grant to a university researcher
for developing and evaluating surrogate characteristics and geographic
information systems as methods to identify pockets of underimmunized
children in 19 counties within a particular state.‘6 According to one
immunization program manager, it would be helpful if CDC provided an
inexpensive evaluation method for identifying pockets of need. According to
several state managers, CDC should do research to determine what methods of
identifying pockets are effective or how to measure the outcomes of


We provided a draft of this correspondence to CDC officials. In a letter dated
July 2, 1997, the Director of CDC said #at the agency generally agrees that
targeting efforts towards pockets of need helps to prevent outbreaks of vaccine-
preventable diseases. CDC reiterated its view that the states, not CDC, are
responsible for identifying pockets of need. According to CDC, because states
are experienced with their unique immunization circumstances, they are better
positioned to identify specific census tracts, zip codes, or other small
geographic areas where children may be at risk. Furthermore, in CD& view,
federal budgetary resources are more appropriately used for interventions that
will increase immunization coverage among preschool children rather than for
additional evaluation of the methods states are using to identify pockets of
underimmunized preschool children. CDC stated that evaluation would require
a complete census of the entire state, county, or city area relating to potential
pockets of need and would require considerable funding. In the agency’s view,
the ultimate evaluation of CDC and state efforts to address pockets of need is
whether disease outbreaks are occurring. Vaccine-preventable disease levels
nationally are generally at, or near, all-time record low levels, CDC says.

16Ageographic information system is computer software that organizes and
links data from different sources to display information on maps. For an
application to disease outbreaks, see M.L. Popovich and B. Tatham, “Use of
Immunization Data and Automated Mapping Techniques to Target Public Health
Outreach Programs,” Am&                                        Vol. 13, No. 2s
(1997), pp. 102-7.

9                      GAOLEIEXS-97-136R      Pockets   of Underimmunization
CDC appears to have misunderstood our discussion about evaluation. We note
that no evaluations have been completed for the methods, which states are
using or that CDC recommended to them, of identifying pockets of
iutderimmunization. We are not implying, however, that CDC should evaluate
the success of states in identifying pockets of underimmunization.

CDC also ,believes that surrogate methods have been adequately evaluated for
their effectiveness in identifying pockets of underimmunized children. To
support this assertion, CDC cites four studies, completed in 1993, that it funded
to find out why preschool children in Baltimore, Los Angeles, Philadelphia, and
Rochester had not been immunized on time. We disagree. Although these
studies identied several factors associated with missed opportunities for
immunization, the studies do not examine the relationships between those
factors and pockets of need.17 Therefore, it cannot be inferred that these
factors are generally associated with pockets nor can these four studies be
generalized to the 50 states.

In written comments, CDC provides examples of state activities to improve
immunization rates, such as cooperating with the WIG program, doing clinic
assessments, and developing registries. However, as we point out in this letter,
these methods were developed to improve coverage rates. They were not
developed to identify pockets of need, nor have they been evaluated for this

CDC takes issue with our illustration of recent measles outbreaks in Alaska and
Utah as examples of pockets of need because only a minority of the cases are
in the preschool age range. A CDC official emphasized that those transmitting
the disease were predominately school-age children. These facts were clear to
us. Any outbreak of disease in a defined pocket has the potential to affect
preschool children as well as school-age children and adults, as these outbreaks
did. CDC also suggests that philosophical or religious objections to
immunization may be involved in some of these cases. We believe that such
objections may play a role in undermummization in some areas and contribute
to determining a pocket of need.

17Forour previous discussion of the four studies, see Vaccines for Children
(GAO/PEMD-95-22), pp. 17-18, and for a published report of one of the studies,
see R.G. Frank and others, ‘The Demand for Childhood linmumxations: Results
from the Baltimore Immunization Study,” Inauirv, Vol. 32, No. 2 (1995), pp. 164-

 10                     GAOBEHS-97-136R         Pockets   of Underimmunization
In CDC’s view, we underemphasize the relative disadvantages of household
survey methods. As indicated in our recent report on the National
Immunization Survey,18household surveys may be comparatively less efficient
because many households must be screened to identify enough with Z-year-old
children. Accordingly, as we point out in this letter, such surveys may be labor
intensive. We discuss relative advantages and disadvantages of ah methods
used to identify pockets of underimmunized children; it is not our intention to
endorse any particular method.                                                         -.

In addition to the comments above, CDC provided some technical comments
that we incorporated into the final letter, when appropriate.

As agreed with your offices, unless you publicly announce its contents earlier,
we plan no further distribution of this letter until 30 days from its date of issue.
At that time, we will send copies of this letter to interested congressional
committees, the Secretary of HHS, the Director of CDC, and other federal and
state officials. We will also make copies available to others upon request.

This letter was prepared under the direction of Sandra K. Isaacson, Assistant
Director, (202) 512-7174. Other major contributors include Richard C. Weston
and George Bogart.

Marsha Lillie-Blanton
Associate Director of Health Services Quality
 and Public Health Issues

Enclosures - 2

‘*See CDC’s National Immunization Survev (GAO/PEMD-9616), pp. 9-17.

11                       GAO/EIEHS-97-136R       Pockets   of Underimmunization
ENCLOSURE I                                                                 ENCLOSURE I

                   Examples                      States
 Provider          Assessments of coverage       Tot.& 20
 assessments       rates for public health       Colorado, Connecticut, Georgia,
                   clinics or private providers; Hawaii, Idaho, Illinois, Iowa, Kentucky,
                   assessments of coverage       Louisiana, Maine, Michigan,” Missouri,
                   rates for the Special         Nebraska, Nevada, New Hampshire,
                   Supplemental Nutrition        New York,” Pennsylvania, Utah,
                   Program for Women,            Virginia, and West Virginia
                   Infants, and Children (WIG)
 Population        Door-to-door surveys; birth   Tot& 8
 surveys           ceticate surveys              Florida> Mississippi, North Carolina,
                                                 Ohio, Oregon, Tennessee, Texas, and
 School            Retrospective surveys          Total: 7
 surveys                                          Kansas, Maryland, Minnesota, South
                                                  Dakota, Vermont, Wisconsin, and
 Surrogate         Ethnic diversity, poverty      Total: 6
 characteristics   rate, population density,      Alabama, Alaska, California, Indiana,
                   socioeconomic status, and      Massachusetts, and Rhode Island
                   disease incidence
 Immunization      Tracking systems for           Total: 6
 registries        immunization information       Arizona, Arkansas, Delaware, North
                                                  Dakota, Oklahoma, and South Carolina

Note: Montana, New Jersey, and New Mexico reported using no method.

“Only Michigan reported planning random household telephone surveys.

bNew York and Florida each reported developing a geographic information system,
applying weighted demographic characteristics to population data derived from census
and commercial sources.

12                                 GAOLKEHS-97-136R       Pockets   of Underimmunization
ENCLOSURE II                                                                ENCLOSURE II

The following is a discussion of the strengths and limitations .of methods states use-
provider evaluations, population surveys, school surveys, surrogate characteristics, and
immunization registries-to identify pockets of under-immunized children or children at
risk for underimmunization.


Provider assessments can be an inexpensive and effective method for improving overall
immunization rates. This is because such assessments focus on immunization coverage
rates among at-risk children who are in contact with health care providers available to
intervene immediately. For example, when routine measurement of immunization
coverage rates began in Georgia’s public health clinics, the immunization rates among
children served in its approximately 220 public health clinics more than doubled, from 37
percent in 1986 to 83 percent in 1994.l’ However, this method is unlikely to be effective if
used to identify pockets of need because underimmunized children not receiving care in
clinics may go undetected.

Assessing the immunization status of children enrolled in WIC is also viewed as a
provider assessment. However, this method has limitations similar to public health clinic
provider assessments. For example, CDC investigators found that during the 1991
measles epidemic in New York City, at least 90 percent of preschool children enrolled in
WIC were up-to-date for measles by 21 months of age.” But no information was provided
on the underimmunized children in New York who were not enrolled in the WIG program.

%V. Orenstein, “Update on CDC’s National Immunization Program,” in All Kids Count
National Program Meeting: Summarv Proceedings, Savannah, Ga (Feb. 29-Mar. 1, 1996),
pp. 8-14. See E.F. Dini and others, “Information as Intervention: How Georgia Used
Vaccination Coverage Data to Double Public Sector Vaccination Coverage in Seven Years,”
Journal of Public Health Management Practice, Vol. 2, No. 1 (1996), pp. 4549, and C.W.
LeBaron and others, “Impact of Measurement and Feedback on Vaccination Coverage in
Public Clinics, 19881994.” JAMA IJournal of the American Medical Associationl, Vol. 277,
No. 8 (1997), pp. 63135.

20Suchhigh coverage for measles immunization at the peak of a major epidemic suggests
that very high rates of coverage are. necessary to prevent outbreaks of measles in
preschool populations (see C.W. LeBaron and others, “Measles Vaccination Levels of
Children Enrolled in WIC during the 1991 Measles Epidemic in New York City,” American
Journal of Public Health, Vol. 86, No. 11 [1996], pp. 1551-56).

13                                GAOIHEHS-97-136R       Pockets   of Underimmunization
ENCLOSURE II                                                                  ENCLOSURE II


Methods such as door-to-door surveys or random household telephone surveys ascertain
the immunization status of preschool children, measure immrmization coverage in specific’
neighborhoods, and might identify geographic pockets of currently underimmunized
children. Such surveys allow for identifying and collecting information on all children,
including children who are unaccounted for in provider records and children who are not
in households with telephones. In addition, door-to-door interviews, conducted in the
home, usually improve response rates and generally increase the validity of responses. In
the home, if parents cannot recall the child’s immunization status, they might have
written documentation to refer to or the surveyor can readily obtain written consent
authorizing access to health care provider records for immunization status. Another
advantage of these surveys is their potential for providing ready linkage to intervention if
vaccinations need to be administered. However, these surveys may be labor intensive.

Another form of population survey is birth certificate surveys. This form might also
locate under-immunized children who do not come into contact with providers. These
surveys rely on identification information collected when a child is born; generally,
therefore, they do not include children who have moved into or-out of the state.
Although some state vital statistics offices now have the capacity to provide data for
public health follow-up activities, tracing children from transient families can be difficult
and time-consuming.21


These surveys may be excellent for determinin g trends in neighborhood and community
coverage. Unlike some other methods, school surveys include children with no regular
providers, children who change providers, and children who may lack connections to the
health care system. However, the delay between timely preschool intrnunization and
entering school limits the utility of retrospective school record surveys compared with
some other methods. This is because timely interventions to improve coverage of those
2-year-olds who are underimmunized is impossible. Furthermore, because of the
retrospective aspect of the survey, school surveys do not include children who were

“For example, the state of Washington has a project to establish a birth data system for
birth certificates, newborn screening, child immunization, birth defects registry, and
maternal and child health referrals, see P. Starr and S. Starr, “Reinventing Vital Statistics:
the Impact of Changes in Information Technology, Welfare Policy, and Health Care,”
Public Health Renorts, Vol. 110, No. 5 (1995), pp. 534-44.

 14                                 GAO/HEHS-97-136R        Pockets   of Underimmunization
ENCLOSURE II                                                                ENCLOSURE II

residents during their preschool years but moved to another school district before


This method can be relatively inexpensive and easy to develop. If the surrogate
characteristics are well-chosen, targeting groups with these characteristics may be useful
in selecting appropriate immunization interventions. Because this method includes using
aggregate data to target groups, it may not identify individual underimmunized children.
Although race and ethnicity or income, for example, may correlate with some pockets of
children at risk for underimmunization, these surrogate characteristics may not reflect the
immunization status of individual children. Additionally, these surrogate characteristics
may not help to increase immunization coverage if they inadequately capture other
critical characteristics of local health care services.=


The usefulness of this method to identify pockets of underimmunization is determined by
the extent to which (1) tracking systems for immunization information registries include
children within a selected age group, within a given commtmity or service area, and (2)
public and private providers submit information on immunization?4 Often registries are
relatively expensive to start and require large-capacity computer hardware and a high
degree of software expertise. Establishing the databases requires maximizing the extent
and ease of public and private provider access, while maintaining confidentiality of

22SeeT V. Murphy and others, “Estimating Immunization Coverage from School-Based
Childhood Immunization Records,” Pediatric-Infectious Disease Journal, Vol. 14, No. 7
(1995), pp. 561-67, and L.E. Rodewald and others, “The School-Based Immunization
Survey: an Inexpensive Tool for Measuring Vaccine Coverage,” American Journal of Public
Health, Vol. 83, No. 12 (1993), pp. 1749-51.
?l’he characteristics of health clinics or of physicians may also be important. See N.
Rudner, “Potentials for Improving Health Department Immunization Rates: The
Relationships Between Service Delivery Factors and Immunization Completion,” Journal
of Public Health Management Practice, Vol. 2, No.1 (1996), pp. 50-58.
%SeeKM. Faherty and others, “Prospects for Childhood Immunization Registries in Public
Health Assessment and Assurance: Initial Observations from the All Kids Count Initiative
Projects,” Journal of Public Health Management Practice, Vol. 2, No. 1 (1996), pp. 1-11.

15                                GAOEIEHS-97-136R       Pockets   of Underimmunization
ENCLOSURE II                                                               ENCLOSURE II
registry records.25 For example, because of conhdentiality considerations, CDC has
recommended including a child’s address-a data element of importance if a registry is
used to identify geographic pockets of under-immunization-in state registry data sets but
not in data sets for transferring a child’s record to another registry.


2”For a discussion of privacy issues, see L-0. Go&in and Z. LazzarK, “Childhood
Immunization Registries: A National Review of Public Health Information Systems and the
Protection of privacy,” JAMA Vol. 274, No. 22 (1995), pp. 1793-99.

 16                               GAOAXHS-97-136R         Pockets   of Underimnmnhation
Ordering     Information

The first copy of each GAO report and testimony is free.
Additional   copies are $2 each. Orders should be sent to the          -.
following  address, accompanied by a check or money order
made out to the Superintendent     of Documents, when
necessary. VISA and Mastercard      credit cards are accepted, also.
Orders for 100 or more copies to be mailed to a single address
are discounted    25 percent.

Orders by mail:

U.S. General Accounting    Office
P.O. Box 37050
Washington,  DC 20013

or visit:

Room 1100
700 4th St. NW (corner of 4th and G Sts. NW)
U.S. General Accounting  Office
Washington,  DC

Orders may also be placed by calling (202) 512-6000
or by using fax number (202) 512-6061, or TDD (202)       512-2537.

Each day, GAO issues a list of newly available reports and
testimony.   To receive facsimile copies of the daily list or any
list from the past 30 days, please call (202) 512-6000 using a
touchtone phone, A recorded menu will provide information         on
how to obtain these lists.

For information on how to access GAO reports on the INTERNET,
send an e-mail message with ‘info” in the body to:


or visit GAO’s World Wide Web Home Page at:

United States
General Accounting    Office
Washington,   D.C. 20548-0001

Official   Business
Penalty    for Private   Use $300

Address    Correction    Requested