County-Level Trends in Vaccination Coverage Among Children Aged 19--35 Months --- United States, 1995--2008

Philip J. Smith, PhD

James A. Singleton, MS

National Center for Immunization and Respiratory Diseases

Corresponding author: Philip J. Smith, PhD, CDC, National Center for Immunization and Respiratory Diseases, MS E-32, 1600 Clifton Road, NE, Atlanta, GA 30333, Telephone: 404-639-8729; Fax: 404-639-3266; E-mail: pzs6@cdc.gov.

Abstract

Problem/Condition: Estimated trends in county-level vaccination coverage compared with national health objectives and associated with other variables (e.g., access to care, economic conditions, and demographic characteristics) have not been reported previously.

Reporting Period: 1995--2008.

Description of System: The National Immunization Survey (NIS) is an ongoing, random-digit--dialed telephone survey that gathers vaccination coverage data from households with children aged 19--35 months in 50 states and selected urban areas and territories.

Results: During 1995--2008, 185,336 children aged 19--35 months sampled by NIS had adequate provider data and lived in one of the 257 counties where the combined sample size for at least one of the seven biennial periods during 1995--2008 was ≥35. Statistically significant increases in estimated vaccination coverage occurred in 27 of 233 counties (12%) with ≥4 doses of diphtheria and tetanus toxoids and acellular pertussis (DTaP); for 38 of 233 counties (16%) with ≥3 doses of polio vaccine; eight of 233 counties (3%) with ≥1 dose of measles, mumps, and rubella (MMR); nine of 233 counties (4%) with ≥3 doses of Haemophilus influenzae type B (Hib) vaccine; 193 of 233 counties (83%) with ≥3 doses of hepatitis B vaccine; 228 of 232 counties (98%) with ≥1 dose of varicella vaccine; and 187 of 192 counties (97%) with ≥4 doses of 7-valent pneumococcal conjugate vaccine (PCV7). Six of 233 (2%) counties had significant decreases in vaccination coverage for Hib. During the 2007--2008 biennial period, the percentage of 193 counties with estimated vaccine coverage that achieved the Healthy People 2010 objective of 90% vaccination coverage was 8% for DTaP/DTP vaccines, 93% for polio vaccine, 86% for MMR vaccine, 71% Hib vaccine, 94% for hepatitis B vaccine, 50% for varicella vaccine, and <1% for PCV7. Among 104 counties, the estimated percentage of children aged 6--23 months who were administered ≥1 dose of the seasonal influenza vaccine during the 2007--2008 influenza vaccination season was 39.0% (range: 22.2%--68.8%). For most vaccines and vaccine series, higher levels of county-level vaccination coverage correlated with a higher number of pediatricians per capita, a higher number of people living in group quarters (e.g., college residence halls, residential treatment centers, skilled nursing facilities, group homes, military barracks, correctional facilities, workers' dormitories, and facilities for persons experiencing homelessness) per capita, higher per capita income, a higher number of Hispanics per capita, and having a service-dependent economy. Lower levels of county-level vaccination coverage correlated with higher number of persons in poverty per capita, a higher percentage of black children among children aged <5 years, higher levels of housing stress (i.e., ≥30% income for rent or mortgage and certain inadequate housing characteristics), a higher number of pediatric intensive care beds per capita, and designation as a nonmetropolitan county with an economy dependent on recreation activities.

Interpretation: During 1995--2008, significant increases in vaccination coverage for individual vaccines occurred in many counties for the newly recommended vaccines, varicella and PCV7.

Public Health Actions: In counties that did not meet the Healthy People 2010 vaccination coverage objectives, states should evaluate strategies to achieve these objectives. The Guide to Community Preventive Services provides a summary of interventions that increase community vaccination coverage, including provider reminder-recall systems that remind parents to return to clinics to administer missed doses to children and assessment and feedback on the performance of vaccination providers. In counties where significant decreases in Hib vaccination coverage occurred, additional research is warranted to determine whether the recent shortage in the Hib vaccine was the sole cause of these decreases. In counties with a high proportion of children living in poverty, interventions to increase vaccination coverage among these children are needed. Additional research is required to understand potential barriers to increased coverage with these vaccines, the role of vaccination providers and their resource constraints, and factors associated with access to health care among children.

Introduction

County-level vaccination coverage estimates are important, both because public health issues often originate in small geographic areas and because certain public health actions are most effective at the local level. For example, a study conducted during 1996--2001 in inner-city locations in Atlanta, Georgia, found a high rate of unvaccinated children among racial/ethnic minorities living in the surveyed neighborhoods (1). In Sullivan County, New York, an investigation by the New York State Department of Health identified a mumps outbreak of among 31 persons (nine not vaccinated, 16 administered 2 doses, four administered 1 dose, two with unknown vaccination history) in a population with 96% vaccination coverage (2). Clusters of children who did not receive any vaccine doses were found in specific geographic areas, with the largest numbers of unvaccinated children living in counties in California, Illinois, New York, Washington, Pennsylvania, Texas, Oklahoma, Colorado, Utah, and Michigan (3). In Ashland, Oregon, 12.3% of all children attending public schools and 18.8% of children attending day-care facilities in 2002 claimed an exemption from mandatory vaccination laws, compared with 2.4% for the entire state (4,5).

Children of parents who refuse vaccine doses have an increased risk for acquiring and transmitting measles and pertussis (6), and in small geographic areas with a preponderance of children whose parents have refused vaccine doses, an increased risk for pertussis and measles exists among members in the community (6--8). In 2008, the index cases for three of four measles outbreaks outbreaks in small geographic areas were in children whose parents had refused vaccine doses (9,10). Measles is not endemic in the United States, and index cases in the measles outbreaks were imported from other counties.

Although a particular state estimate of vaccination coverage might be high overall, geographic areas with underimmunized children can exist within a state when unvaccinated children or children with lower vaccination coverage live in the same geographical location. In this case, vaccination coverage in those areas might not be sufficient to inhibit the widespread transmission of a vaccine-preventable disease (11).

To characterize trends in estimated vaccination coverage among children aged 19--35 months in 257 selected U. S. counties, CDC analyzed 1995--2008 data from the National Immunization Survey (NIS) to determine achievement of Healthy People 2010 objectives of 90% vaccination coverage for individual vaccines and 80% for vaccine series of certain routinely recommended childhood vaccines (12). Data were analyzed to identify increases and decreases in vaccination coverage among children aged 19--35 months in selected counties and associations between county-level vaccination coverage rates and county-level indices of access to care, economic conditions, and demographic composition. Estimated trends in vaccination coverage in small geographic areas such as counties might assist local public health officials with determining the current and past level of vaccination coverage and whether the coverage is likely to increase or decrease if current trends persist. This type of information might help local public health officials improve vaccination coverage.

Methods

NIS is an annual survey conducted by CDC to monitor vaccination coverage rates among U.S. children aged 19--35 months. During 1995--2008, NIS was conducted in all 50 states and the District of Columbia and selected urban areas. Data are collected in NIS in two phases. The first phase is a list-assisted random-digit--dialed telephone survey of households with landline telephones and an age-eligible child (aged 19--35 months). At the conclusion of the interview, the interviewer requests parental permission to contact the children's vaccination providers. The second phase is a mail survey of doctors and other vaccination providers named by the respondent when permission is given. In all survey years during 1995--2008, provider-reported vaccination histories obtained from the mailed survey were used to evaluate vaccination status. During 1995--2008, the response rate (13) of the telephone survey of NIS ranged from 65% to 76%, and the percentages of sampled children with complete telephone interviews having a sufficiently detailed vaccination history returned from vaccination providers to accept as a complete report ranged from 62% to 73%. The response rate of the telephone portion of NIS is the product of the percentage of telephone numbers sampled from NIS list-assisted sampling frame that were determined to be working residential telephone numbers, the percentage of sampled residential telephone numbers that were successfully screened to determine whether the household had children aged 19--25 months, and the percentage of sampled households with children aged 19--35 months that had completed the NIS telephone interview. Detailed descriptions of the statistical methods used by NIS have been published previously (14,15).

Measures of Childhood Vaccination Status

In 2009, 10 vaccines were recommended for routine use among children aged 0--24 months (16). The Advisory Committee on Immunization Practices (ACIP) recommendations for routine administration of childhood vaccines were used to evaluate the vaccination status of sampled children for all seven biennial periods during 1995--2008, regardless of vaccine shortages during which ACIP issued interim recommendations (Box 1). Sampled children were determined to be up to date with respect to selected vaccines if, by the date of the NIS telephone interview, they had been administered ≥4 doses of diphtheria and tetanus toxoids and acellular pertussis or diphtheria and tetanus toxoids and pertussis (DTaP/DTP) vaccine; ≥3 doses of polio vaccine; ≥1 dose of measles, mumps, and rubella (MMR) vaccine; ≥3 doses of Haemophilus influenzae type B (Hib) vaccine; ≥3 doses of hepatitis B vaccine; ≥1 dose of varicella vaccine; ≥2 doses of hepatitis A vaccine; and ≥4 doses of the 7-valent pneumococcal conjugate vaccine (PCV7) (Box 2). For a given survey year, sampled children aged 6--23 months were considered to be up to date for the seasonal influenza vaccine if they were administered 1 dose of the seasonal influenza vaccine during the most recent seasonal influenza season, unless they were not administered an influenza vaccine dose before the most recent influenza season, in which case they required 2 doses of the vaccine during the most recent influenza season (17). Other measures of vaccination coverage that were evaluated include ≥3 doses of DTaP/DTP, ≥3 doses of PCV7, ≥1 dose of the seasonal influenza vaccine, and ≥1 dose of hepatitis A vaccine.

Sample Sizes and Number of Counties with Coverage Estimates

To ensure the confidentiality of survey information during each of the seven biennial periods (1995--1996, 1997--1998, 1999--2000, 2001--2002, 2003--2004, 2005--2006, and 2007--2008), vaccination coverage was estimated only for counties where the sample size of children with sufficient vaccination and health-care provider data from the mailed survey portion of NIS was ≥35 for the combined survey years during the biennial period and ≥14 for each year during the biennial period.

For each of the 257 counties (among the 3,141 counties in the United States) that met the sample size requirement for at least one biennial period, an averaged estimate over the biennial period was calculated. The estimate of the number of children aged 19--35 months as of July 1, 2007, living in the 257 counties was determined from U.S. Census data (18), as well as the rank of those 257 counties according to their population of children aged 19--35 months as of July 1, 2007 (Table 1). In addition, the number of counties for which estimates were reported for each biennial period and vaccine or vaccine series type were calculated (Table 2). Because the sample size in counties fluctuated from year to year, not all counties met the sample size requirement in every biennial period. If not achieved for a biennial period, estimates were reported for fewer than 257 counties (Table 2). During 1995--2008, 185,336 children aged 19--35 months sampled by NIS had adequate provider data and lived in one of the 257 counties where the combined sample size for at least one of the seven biennial periods during 1995--2008 was ≥35.

Statistical Analysis

Coverage rates for the seven biennial periods were estimated using James-Stein statistical methods (19) to obtain estimates with greater precision. James-Stein estimates of vaccination coverage were obtained for each vaccine and vaccine series by averaging the two county-level annual estimates for each biennial period and then, for each biennial period, modeling the averaged estimates using multivariable regression as a function of 35 variables derived from the U.S. Department of Health and Human Services Area Resource File. Estimates were compared with the Healthy People 2010 vaccination coverage objectives of 90% vaccination coverage for individual vaccines and 80% for vaccine series to determine which counties achieved the Healthy People 2010 objectives.

Variables from the Area Resource File, including county-level indices of access to care, economic conditions, and demographic composition (Table 2), were used to evaluate the association between averaged county-level vaccination coverage rates and county-level indices of access to care, economic conditions, and demographic composition. Forward stepwise regression was used to select predictors among the 35 variables to explain variation in the log odds of averaged county estimates.

Among the 257 counties that achieved the sample size requirement for at least two biennial periods, trends were evaluated across time by comparing the difference of the James-Stein composite estimate for the earliest biennial period with the most recent biennial period. A p value of <0.05 from a one-sided statistical z score test was considered statistically significant.

Although comparison of the 1995--1996 and 2007--2008 biennial periods was preferred, certain counties did not have large enough sample sizes to protect the confidentiality of survey respondents during these periods. Therefore, the trend analysis included counties that met the sample size requirement for at least two biennial periods at any time during 1995--2008 (comparing the earliest and most recent periods), and the overall number of counties used to examine trends differed from the number of counties studied (Table 3).

Results

Representation of Selected Counties

During the 2007--2008 biennial period, a total of 20,552 children aged 19--35 months who were sampled by NIS had 1) sufficient provider data returned from the mail survey sent to the vaccination providers of the sampled children and 2) lived in the 193 counties where the sample size was ≥35 for the combined years and ≥14 for each survey year. During the 2007--2008 biennial period, the percentage of children aged 19--35 months living in these 193 counties comprised 49% of all children aged 19--35 months living in the United States. County-level estimates from the 2007--2008 biennial period includes 106 of the 150 counties with the greatest number of children aged 19--35 months as of July 1, 2007, as estimated from U.S. Census data.

Estimated County-Level Vaccination Coverage for Individual Vaccines During 2007--2008

During the 2007--2008 biennial period, estimated vaccination coverage ranges were as follows for individual vaccines: ≥3 doses of DTaP/DTP ranged from 91.1% in Clark County, Nevada, to 97.4% in New York County, New York (Table 4); ≥4 doses of DTaP/DTP ranged from 72.2% in Clark County, Nevada, to 95.6% in Grafton County, New Hampshire (Table 5); ≥3 doses of polio ranged from 86.8% in Pierce County, Washington, to 96.8% in Washington County, Rhode Island (Table 6); ≥1 dose of MMR ranged from 86.4% in Clark County, Nevada, to 96.6% in Suffolk County, Massachusetts (Table 7); ≥3 doses of Hib ranged from 79.2% in San Juan County, New Mexico, to 96.5% in Suffolk County, Massachusetts (Table 8); ≥3 doses of hepatitis B vaccine ranged from 85.8% in Clark County, Nevada, to 96.2% in New Castle, Delaware (Table 9); ≥1 dose of varicella vaccine ranged from 74.9% in Flathead County, Montana, to 95.0% in Hartford County, Connecticut (Table 10); ≥4 doses of PCV7 ranged from 57.8% in Hinds County, Mississippi, to 91.0% in Howard County, Maryland (≥3 doses also was measured) (Tables 11 and 12); ≥1 dose of the seasonal influenza vaccine ranged from 22.2% in El Paso County, Texas, to 68.8% in New York County, New York (Table 13); the required number of doses to be up-to-date on the seasonal influenza vaccine ranged from 10.2% in El Paso County, Texas, to 55.7% in New York, New York (Table 14); ≥1 dose of hepatitis A vaccine ranged from 24.2% in Gallatin County, Montana, to 88.6% in Dona Ana County, New Mexico (Table 15); and ≥2 doses of hepatitis A vaccine ranged from 17.5% in Sedgwick County, Kansas, to 47.6% in El Paso County, Texas (Table 16).

County-Level Trends

During the first and last biennial periods for which James-Stein estimates were available, statistically significant increases in estimated vaccination coverage occurred in 27 of 233 counties (12%) with ≥4 doses of DTaP; 38 of 233 counties (16%) with ≥3 doses of polio vaccine; eight of 233 counties (3%) with ≥1 dose of MMR; nine of 233 counties (4%) with ≥3 doses of Hib; 193 of 233 counties (83%) with ≥3 doses of hepatitis B vaccine; 228 of 232 counties (98%) with ≥1 dose of varicella vaccine; and 187 of 192 counties (97%) with ≥4 doses of PCV7. For the vaccination series, statistically significant increases occurred in 29 of 233 counties (12%) with 4:3:1; in 30 of 233 counties (13%) with 4:3:1:3; in 159 of 233 counties (68%) with 4:3:1:3:3; in 139 of 213 counties (65%) with 4:3:1:3:3:1; and in 181 of 192 counties (94%) with 4:3:1:3:3:1:4.

In six of 233 counties, statistically significant decreases occurred in the estimated percentage of administered doses or vaccination coverage rate for certain vaccines and vaccine series: ≥4 doses of the DTaP/DTP vaccine (one county: Collin County, Texas); ≥3 doses of the Hib vaccine (five counties: Hartford County, Connecticut; Pennington County, South Dakota; King County, Washington; Laramie County, Wyoming; Sweetwater County, Wyoming); the 4:3:1 series (one county: Collin County, Texas); and the 4:3:1:3 series (one county: Collin County, Texas).

Achievement of Healthy People 2010 Objectives During 2007--2008

For individual vaccines recommended during the 2007--2008 biennial period, the number of counties that achieved the Healthy People 2010 objective of 90% coverage for individual vaccines were as follows: 15 of 193 counties (8%) for ≥4 doses of DTaP (Table 5), 179 of 193 counties (93%) for ≥3 doses of polio vaccine (Table 6), 166 of 193 counties (86%) for ≥1 dose of MMR (Table 7), 137 of 193 counties (71%) for ≥3 doses of Hib (Table 8), 182 of 193 counties (94%) for ≥3 doses of hepatitis B vaccine (Table 9), 97 of 193 counties (50%) for ≥1 dose of varicella vaccine (Table 10), and one of 193 counties (<1%) for ≥4 doses of PCV7 (Table 12). Among the counties that did not achieve the Healthy People 2010 vaccination coverage objective of 90% coverage for DTaP, the estimated coverage rates ranged from <1 to 9.9 percentage points lower than the coverage objective; for MMR, coverage ranged from <1 to 3.6 percentage points lower than the objective; for Hib, coverage ranged from <1 to 7.6 percentage points lower than the objective; for hepatitis B vaccine, coverage ranged from <1 to 4.2 percentage points lower than the objective; and for varicella vaccine, coverage ranged from <1 to 9.0 percentage points lower than the objective.

County-level estimates of the 3:3:1 series (Table 17) provide an estimate of what coverage would be in those counties had all children who received 3 doses of DTaP also received the recommended fourth dose. For vaccine series recommended during the most recent 2007--2008 biennial period, the number of counties achieving the Healthy People 2010 objective of 80% coverage was 151 of 193 counties (77%) for the 4:3:1 series (Table 18), 113 of 193 counties (58%) for the 4:3:1:3 series (Table 19), 77 of 193 counties (40%) for the 4:3:1:3:3 series (Table 20), 30 of 193 counties (16%) for the 4:3:1:3:3:1 series (Table 21), and two of 193 counties (1%) for the 4:3:1:3:3:1:4 series (Table 22). Among the counties that did not achieve the Healthy People 2010 objective for the 4:3:1 series, vaccination coverage rates ranged from <1 to 9.2 percentage points lower than the coverage objective; for the 4:3:1:3 series, coverage ranged from <1 to 9.2 percentage points lower than the objective; for the 4:3:1:3:3 series, coverage ranged from <1 to 9.5 percentage points lower than the coverage objective; and for the 4:3:1:3:3:1 series, coverage ranged from <1 to 9.8 percentage points below the coverage objective.

Association of County-Level Factors with Estimated County-Level Vaccination Coverage Rates During 2007--2008

In fitting the forward stepwise regressions to obtain modeled estimates of the log-odds of county-level vaccination coverage rates for the James-Stein estimates, counties with lower estimated vaccination coverage rates had higher per capita persons in poverty, had higher percentages of black children among children aged <5 years, had a higher number of pediatric intensive care beds per capita, were designated as a nonmetropolitan county with an economy dependent on recreation activities, or were designated as experiencing housing stress (Table 2). Counties with housing stress are defined by the U.S. Department of Agriculture Economic Research Service (USDA ERS) as those in which ≥30% of households meet one or more of the following housing conditions: lack of complete plumbing, lack of complete kitchen, pay ≥30% percent of income for owner costs or rent, or more than one person per room (20). In general, counties with higher estimated vaccination coverage rates were designated as having a higher number of pediatricians per capita, having higher per capita income, having a higher number of Hispanic persons per capita, being designated as having a service-dependent economy, or having a higher number of persons per capita living in group quarters (e.g., college residence halls, residential treatment centers, skilled nursing facilities, group homes, military barracks, correctional facilities, workers' dormitories, and facilities for persons experiencing homelessness) (Table 2).

Discussion

During 2007--2008, the 193 counties included in this report included 49.0% of the population of children aged 19--35 months in the United States and 106 of the 150 most populous counties. With the exception of the hepatitis A vaccine (for which estimates of county-level vaccination coverage were available for only one biennial period), significant increases in county-level estimated vaccination coverage occurred for all the individual vaccines and vaccine series from the earliest and latest biennial periods for which estimates were available. County-level increases for individual vaccines ranged from 3.6% for MMR to 71.4% for varicella vaccine. The percentage of counties with significant increases in county-level vaccination coverage rates was greater for the more recently recommended vaccines, varicella and PCV7, and for the vaccine series that included varicella vaccine and PCV7. However, significant increases in estimated county-level vaccination coverage rates occurred in a moderate number of counties for individual vaccines that were recommended before 1995.

The Healthy People 2010 vaccination coverage objective of 90% for individual vaccines was achieved for a majority of the 193 selected counties during the 2007--2008 biennial period for the recommended number of doses for polio, MMR, Hib, hepatitis B, and varicella vaccines. In December 2007, a shortage of the Hib vaccine began and continued until mid-2009 (21,22). Research is underway to evaluate the extent to which Hib coverage decreased during the shortage.

Since licensure of hepatitis A vaccine during 1995--1996, the hepatitis A childhood immunization strategy has been implemented incrementally, beginning with the 1996 recommendation to vaccinate children living in communities with the highest disease rates, continuing in 1999 with recommendations for vaccination of children living in states, counties, and communities with consistently high hepatitis A rates, and culminating in 2006 with the recommendation for routine hepatitis A vaccination of children nationwide (23). Healthy People 2010 vaccination coverage objectives were not specified for hepatitis A vaccine. However, a vaccination coverage objective of 60% has been proposed for the Healthy People 2020 objective (24) for national hepatitis A vaccine coverage.

The findings in this report indicate that higher county-level estimated vaccination coverage is associated with greater access to primary care (e.g., more pediatricians per capita). In addition, high county-level vaccination coverage rates were found to be positively associated with per capita living in group quarters. Because the per capita living in group quarters variable used measures the extent to which a county provides housing, custodial, medical, and other services to residents who live in group quarters, the variable might provide an indirect measure of the extent to which the infrastructure of a county is organized to provide access to other medical services for other county residents (e.g., primary care services to infants). At the county level, per capita income was found to be positively associated with higher vaccination coverage rates. In other research, family income has been found to be a significant predictor of whether individual children have received all recommended childhood vaccines (25,26).

County-level factors associated with lower county-level estimated vaccination coverage include correlates of poverty. In particular, lower county-level vaccination coverage rates were found to be associated with higher levels of housing stress. Because the housing stress variable measures the extent to which ≥30% of the residents in a county live in substandard housing conditions or pay a disproportionate amount of their income for rent or mortgage, the variable can be considered an indicator of county-level poverty. Poverty is significantly correlated with lower levels of vaccination coverage (25,26). In counties where housing stress is pervasive, lower county-level estimates of vaccination coverage among infants are likely (27).

Lower county-level coverage also was associated with a higher number of pediatric intensive care beds per capita. This finding might be a result of counties allocating resources for care at hospitals rather than for primary care at medical facilities where vaccinations are administered.

The Vaccines for Children Program, implemented during 1994, was designed to mitigate the effect of the cost of childhood vaccines by providing vaccines at no cost to children entitled to the program. In 2009, VFC children had vaccination coverage levels that were similar to but slightly below coverage levels attained by fully insured children. However, determining what the coverage level would be without the VFC program is not possible.

Because the data in this report were obtained from a large national survey, the data were pooled for 2 consecutive years, and the James-Stein estimation method was used, precise estimates of vaccination coverage for many counties were obtained, making increasing trends in counties easier to detect with higher statistical power (27). In addition, ACIP definitions were used for up-to-date vaccination status for routine vaccine administration. During a vaccine shortage, ACIP usually suspends its recommendations for the routinely recommended number of doses and issues interim recommendations that require fewer doses. Therefore, the up-to-date status of children in this report was determined using ACIP routine recommendations that are used when no shortages exist, rather than ACIP interim recommendations that temporarily decrease the number of vaccine doses until a shortage has been resolved.

The findings in this report are subject to at least five limitations. First, NIS has moderate response rates, and nonresponse bias might have resulted in an overestimation of vaccination coverage. However, the statistical methods of NIS (15) are designed to reduce potential nonresponse bias, and some research suggests that the bias in NIS might be small (28). Second, NIS is a survey of children living in households with landline telephones. Therefore, the results in this report might not be generalizable to children who live in households with no telephone service or in households with cellular telephones but no landline. However, recent evidence suggests that bias in surveys that only sample households with landline telephones might be small (29,30). Third, counties with ≥90% vaccination coverage for individual vaccines and ≥80% vaccination coverage for vaccine series were categorized as meeting the Healthy People 2010 vaccination coverage objectives. In some counties, the results have wide confidence intervals, which indicates that those estimates are less precise. Fourth, because the county coverage estimates for a biennial period are reported only when the sample size during the biennial period is ≥35, the number of biennial periods used to track trends in coverage varied from county to county. From two to seven biennial estimates are available to track trends depending on the sample size for each biennial period. Because the number of estimates available for a trend analysis was small, a simple method for evaluating whether coverage had increased or decreased over the time period covered by those biennial estimates was used, in which the two biennial estimates at the beginning and end of that period were compared. Those two biennial estimates do not necessarily correspond to the first and last biennial periods during 1995--2008. As additional years of NIS data become available, more sophisticated trend analyses might be conducted to allow the evaluation of differential trends between counties for the same vaccine and to assess whether vaccination coverage increased at certain points but decreased thereafter. Finally, of the approximately 2,300 of 3,141 counties in the United States that were sampled by NIS during the 2007--2008 biennial period, only 193 counties met the sample size requirement. However, those counties represented approximately 49% of all children aged 19--35 months in the United States, and the estimates of vaccination coverage among children living in all of those counties were very close to the estimated vaccination coverage for the United States. However, a more careful analysis of counties that did not meet the sample size requirement is merited when sufficient sample size has accrued over time.

A goal of a vaccination program is to maintain a sufficiently high level of vaccination coverage to minimize the effects of vaccine-preventable diseases. The Healthy People 2010 and Healthy People 2020 (31) vaccination coverage objectives are milestones toward achieving high levels of vaccination coverage. Those milestones must be achieved in small areas (e.g., counties and communities) to eventually achieve high levels of vaccination coverage for all 50 states and the District of Columbia. Surveillance of vaccination coverage in small geographic areas might be useful for local health authorities to identify areas with low vaccination coverage and in need of policy or program intervention and might further decrease actual and potential morbidity from vaccine-preventable diseases. County estimates from NIS can supplement other local area methods of assessment, including Immunization Information Systems, retrospective school assessment surveys, and area-specific telephone or household surveys. The Guide to Community Preventive Services (32) provides a summary of interventions that have been evaluated to increase community vaccination coverage. Interventions that have been found to be effective include provider reminder-recall systems that recall parents to clinics to administer missed doses to children, assessment and feedback for vaccination providers that retrospectively evaluates the performance of providers in delivering one or more vaccinations to a client population, and use of Immunization Information Systems at the point of clinical care and to aid in surveillance and investigations of vaccination coverage (33).

References

1. LeBaron CW, Starnes DM, Rask KJ. The impact of reminder-recall interventions on low vaccination coverage in an inner-city population. Arch Pediatr Adolesc Med 2004;158:255--61.
2. CDC. Mumps outbreak in a summer camp---2005. MMWR 2006;55;175--7.
3. Smith, PJ, Chu SY, Barker LE. Children who have received no vaccines: who are they and where do they live? Pediatrics 2004;114:187--95.
4. Jackson County Department of Health and Human Services; Immunization Program of the Oregon Department of Human Services. What Ashland parents told us about vaccines and religious exemptions. Jackson County, OR; Jackson County Department of Health and Human Services; 2003. Available at http://www.co.jackson.or.us/files/ashland%20cvs%20final%20report.pdf. Accessed March 9, 2011.
5. Jackson County Department of Health and Human Services; Immunization Program of the Oregon Department of Human Services. School exemptions and disease risk in Ashland, Oregon. Jackson County, OR; Jackson County Department of Health and Human Services; 2004. Available at http://www.co.jackson.or.us/files/school%20exemptions%20and%20disease%20risk%20-%20final.pdf. Accessed March 9, 2011.
6. Omer SB, Salmon DA, Orenstein WA, et al. Vaccine refusal, mandatory immunization, and the risks of vaccine-preventable diseases. N Engl J Med 2009;360:1981--1988.
7. Omer SB, Enger KS, Moulton LH, et al. Geographic clustering of nonmedical exemptions to school immunization requirements and associations with geographic clustering of pertussis. Am J Epidemiol 2008;168:1389--96.
8. Salmon DA, Haber M, Gangarosa EJ, et al. Health consequences of religious and philosophical exemptions from immunizations laws: individual and society risk for measles. JAMA 1999;282:47--53.
9. CDC. Measles---United States, January 1--April 25, 2008. MMWR 2008;57:494--8.
10. CDC. Update: measles---United States, January--July 2008. MMWR 2008;57:893--6.
11. Fine P. Herd immunity: history, theory, practice. Epidemiol Rev 1993;15:265--302.
12. US Department of Health and Human Services. Healthy people 2010, 2nd ed. Understanding and improving health and objectives for improving health (2 vols). Washington, DC: US Department of Health and Human Services; 2000.
13. American Association for Public Opinion Research. Standard definitions. Ann Arbor, MI. American Association for Public Opinion Research; 1998. Available at http://www.aapor.org/standard_definitions/3049.htm. Accessed March 9, 2011.
14. Smith PJ, Battaglia MP, Huggins VJ, et al. Overview of the sampling design and statistical methods used in the National Immunization Survey. Am J Prev Med 2001;20(Suppl 4):17--24.
15. Smith PJ, Hoaglin DC, Battaglia MP, et al. Statistical methodology of the national immunization survey: 1994--2002. National Center for Health Statistics. Vital Health Stat Series 2; 2005;2(138).
16. CDC. Recommended immunization schedules for persons aged 0--18 years. United States, 2009. MMWR 2009;57(51,52):Q1--Q4.
17. CDC. Influenza vaccination coverage among children aged 6--23 months---United States, 2007--2008 influenza season. MMWR 2009;58:1063--6.
18. US Bureau of the Census. Population estimates: county population estimates annual estimates of the resident population for counties: April 1, 2000--July 1, 2009. Available at http://www.census.gov/popest/counties/co-est2009-01.html. Accessed March 9, 2011.
19. James W, Stein C. Estimation with quadratic loss. Proceedings of the fourth Berkeley symposium on mathematical statistics and probability. Berkeley, CA: University of California Press; 1961:361--79.
20. US Department of Agriculture, Economic Research Service. Measuring rurality: 2004 county typology codes. Housing stress counties, 2000. Washington, DC: US Department of Agriculture, Economic Research Service. Available at http://www.ers.usda.gov/briefing/rurality/typology/maps/housing.htm. Accessed February 14, 2011.
21. CDC. Continued shortage of Haemophilus influenzae type B (Hib) conjugate vaccines and potential implications for Hib surveillance---United States, 2008. MMWR 2008;57:1252--5.
22. CDC. Licensure of a Haemophilus influenzae type B (Hib) vaccine (Hiberix) and updated recommendations for use of Hib vaccine. MMWR 2009;58:1008--9.
23. CDC. Prevention of hepatitis A through active or passive immunization. Recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR 2006;55(RR07).
24. US Department of Health and Human Services. Healthypeople.gov. 2020 topics & objectives. IID-7.8. Washington, DC: US Department of Health and Human Services; 2011. Available at http://www.healthypeople.gov/2020/topicsobjectives2020/objectiveslist.aspx?topicid=23. Accessed March 9, 2011.
25. Klevens RM, Luman ET. Children living in and near poverty. Am J Prev Med 2001;20(4 Suppl):41--6.
26. Smith PJ, Jain N, Stevenson J, et al. Progress in timely vaccination coverage among children living in low-income households. Arch Pediatr Adolesc Med 2009;63:462--9.
27. Smith PJ, Singleton JA. Assessment of vaccination coverage for 181 selected counties in the U.S.: attainment of the healthy people 2010 goals and association with county-level indices of access to care, economic conditions, and demographic composition. Public Health Rep 2008;123:1--25.
28. Molinari NA, Wolter K, Skalland B, et al. Quantifying bias in a health survey: modeling total survey error in the National Immunization Survey. Stat Med 2011;30:505--14.
29. Blumberg SJ, Luke JV. Coverage bias in traditional telephone surveys of low-income and young adults. Public Opin Q 2007;71:734--49.
30. Blumberg SJ, Luke JV. Reevaluating the need for concern regarding noncoverage bias in landline surveys. Am J Public Health 2009;99:1806--10.
31. US Department of Health and Human Services. Healthypeople.gov. 2020 topics & objectives. Immunization and infectious diseases. Washington, DC: US Department of Health and Human Services; 20111. Available at http://www.healthypeople.gov/2020/topicsobjectives2020/overview.aspx?topicid=23. Accessed March 9, 2011.
32. CDC; Task Force on Community Preventive Services. Guide to community preventive services: the community guide. Atlanta, GA: Task Force on Community Preventive Services. Available at http://www.thecommunityguide.org. Accessed December 15, 2010.
33. Briss PA, Rodewald LE, Hinman AR, et al. Reviews of evidence regarding interventions to improve vaccination coverage in children, adolescents, and adults. Am J Prev Med 2000;18:97--140.

BOX 1. Biennial periods in which recommended childhood vaccines were included in the National Immunization Survey during 1995--2008

BOX 2. Vaccine components of selected vaccine series