
,




* Center for Health Studies, Group Health Cooperative, Seattle, Washington
Pritzker School of Medicine, University of Chicago, Chicago, Illinois
Department of Pediatrics, Medical College of Wisconsin, Milwaukee, Wisconsin
|| Departments of Pediatrics and Epidemiology, University of Washington, Seattle, Washington
¶ Division of Research, Kaiser Permanente of Northern California, Oakland, California
# Center for Health Research, Kaiser Permanente Northwest, Portland, Oregon
** Center for Vaccine Research, HarborUCLA Medical Center, Torrance, California

KaiserUCLA Vaccine Research Group, Southern California Kaiser Permanente, Panorama City, California

National Immunization Program, Vaccine Safety and Development Activity, Centers for Disease Control and Prevention, Atlanta, Georgia
| ABSTRACT |
|---|
|
|
|---|
Methods. A cohort study of 176134 children born between January 1, 1994, and December 31, 1997, and monitored from birth to the second birthday was performed. Subjects belonged to the Vaccine Safety Datalink Project, a study of children enrolled in 1 of 4 HMOs. Children were continuously enrolled in a HMO for the first 2 years of life. Prevailing recommendations regarding optimal ages of immunization and intervals between doses were applied to define appropriate immunization timing and immunization status. Noncompliance was defined as having a missing or late immunization or an immunization error. Immunization errors included invalid immunizations (too early to be acceptable), extra immunizations (superfluous immunizations or make-up immunizations for invalid immunizations), and missed opportunities resulting in late or missing immunizations.
Results. Although 75.4% of children in these HMOs were up to date for all immunizations at 2 years, only 35.6% of children were fully compliant with recommended immunization practices. Less than 8% of children received all immunizations in accordance with strict interpretation of recommended guidelines. Fifty-one percent of children had at least 1 immunization error by age 2 years; 29.7% had a missed opportunity with subsequent late or missing immunization, 20.4% had an invalid immunization, and 11.6% had an extra immunization. Common reasons for noncompliance included missed opportunities for the fourth Haemophilus influenzae type b vaccine (14.6%), invalid fourth diphtheria-tetanus-pertussis/acellular pertussis immunizations (11.0%), and superfluous polio immunizations (9.8%).
Conclusions. Approximately 35.6% of children were compliant with prevailing childhood immunization recommendations from 1996 to 1999. Efforts to improve compliance with guidelines are recommended, to optimize childhood infectious disease prevention.
Key Words: compliance immunization guidelines health maintenance organization
Abbreviations: HMO, health maintenance organization Hib, Haemophilus influenzae type b DTP, diphtheria-tetanus-pertussis DTaP, diphtheria-tetanus-acellular pertussis HBV, hepatitis B virus IPV, inactivated polio vaccine OPV, oral polio vaccine MMR, measles-mumps-rubella VSD, Vaccine Safety Datalink Hib-OMP, H influenzae type b meningococcal outer membrane protein conjugate vaccine CI, confidence interval
The development and use of immunizations against infectious agents have been important and successful steps toward disease prevention. Childhood immunizations have reduced dramatically the incidence rates of debilitating and sometimes lethal diseases.1,2 The high levels of immunization among US children represent a significant public health achievement.3,4
Despite the successes in the immunization of children, the system is imperfect, because many children remain unvaccinated, are vaccinated late, or are vaccinated erroneously. Measurements of overall immunization coverage levels may overlook significant deviations from recommended immunization practices. Efforts to ensure proper immunization are warranted to ensure that children are optimally protected, with minimization of unnecessary risks and costs.5 Indeed, outbreaks of measles and pertussis have been attributed to improper immunization practices.6,7
Several studies have shown that deviations from recommended immunization practices occur commonly for different vaccines.815 These deviations occur despite annually published immunization guidelines that explicitly define criteria for the timing of immunizations (Table 1). 1623 Previous studies of immunization compliance concentrated on 3 types of errors, namely, missed opportunities, extra immunizations, and invalid immunizations.914 However, little is known about which specific errors occur most frequently and which errors have the greatest impact on immunization status.
|
We therefore aimed to evaluate compliance with immunization guidelines for a large cohort of children. We also assessed the frequency of immunization errors in this cohort, with attention to the effects that these errors had on subsequent immunization status. Finally, we analyzed trends in error rates over the study period and attempted to characterize risk factors for major errors.
| METHODS |
|---|
|
|
|---|
VSD data have been used in numerous studies pertaining to childhood immunization.24 Each child possesses a unique identifier, and demographic data include birth date, gender, Medicaid status, and race (except for the Group Health Cooperative). Immunization histories are derived primarily from computerized immunization tracking systems maintained by the HMOs and include data concerning immunization date, type of vaccine administered, and clinic at which the vaccine was administered.
This study analyzed data for 176134 children born between January 1, 1994, and December 31, 1997, who were continuously enrolled for the first 2 years of life. The enrollment requirement was made to ensure completeness of immunization histories. We restricted our analyses to the following types of vaccines administered before age 2 years: polio vaccine (inactivated polio vaccine [IPV] or oral polio vaccine [OPV]), diphtheria-tetanus-pertussis (DTP)/diphtheria-tetanus-acellular pertussis (DTaP) vaccine, Haemophilus influenzae type b (Hib) vaccine (Hib conjugate vaccine, Hib-tetanus toxoid conjugate vaccine, or Hib meningococcal outer membrane protein conjugate vaccine [Hib-OMP]), hepatitis B virus (HBV) vaccine, and measles-mumps-rubella (MMR) vaccine. Varicella zoster vaccine was introduced during the study period, conjugate pneumococcal vaccine was released after the start of the study period, and rotavirus vaccine was withdrawn after only a brief period of utilization at some of the HMOs; therefore, these vaccines were not included in this study.
Because records were entered into an automated database, we considered the possibility of clerical errors when immunizations occurred close in time. Therefore, to achieve greater confidence that each automated record represented a distinct immunization, we assumed that immunizations of the same type given within 14 days of each other represented duplicate record entries, and these duplicate records (11486 immunizations, 0.5% of the total) were eliminated from our analyses.
Definitions of Immunization Errors
To define the appropriate scheduling of immunizations, minimal ages, and minimal intervals between doses, we applied the recommended childhood immunization schedules for 19941999. A compilation of these recommendations for children <2 years of age appears in Table 1. A child was considered up to date if the following immunizations were received by age 2 years: 3 polio vaccine doses (IPV or OPV), 4 DTP or DTaP vaccine doses, 4 Hib conjugate vaccine or Hib-tetanus toxoid conjugate vaccine doses (or 3 doses of Hib-OMP), 3 HBV vaccine doses, and 1 MMR vaccine dose.
An immunization error was defined as a missed opportunity, an invalid immunization, or an extra immunization. Missed opportunities were defined as visits at which a child received at least 1 immunization and was eligible to receive a different immunization that was not administered. A late immunization was defined as an immunization given later than the recommended age for that dose in the series. Immunizations were considered on time if they were administered when the child was the recommended age. If a late or missing immunization followed a missed opportunity, then the missed opportunity was considered an error that affected the child's immunization status and was included in our analysis (counted as 1 error, not 2). In cases in which an immunization was given on time despite an earlier missed opportunity, the missed opportunities were discussed separately and were excluded from aggregate error calculations.
Immunizations were considered invalid if they were administered before the minimal recommended age had been achieved or before the minimal between-dose interval had elapsed. In many cases, the minimal age requirements did not correspond to the optimal recommended age or interval for immunization. Therefore, invalid immunizations were distinguished from early (but valid) immunizations that occurred before the optimal recommended age or interval. Early (but valid) and late immunizations were not considered errors. Children vaccinated in strict accordance with recommended guidelines ("age-appropriate" or "fully compliant" immunization) received all immunizations on time, with no immunization errors.
Children who received more than the recommended number of doses of any vaccine before age 2 years were considered extraimmunized. Extra immunizations were categorized as either superfluous (administered before age 2 years after completion of the initial series) or make-up (to compensate for a prior invalid immunization). If a make-up immunization was given, then only 1 error was counted (for the invalid immunization).
Several exceptions were allowed for immunizations in excess of the recommended number. A second MMR vaccine dose could be given before age 2 years if both immunizations were given after age 1 year and the immunizations were at least 4 weeks apart, in accordance with Advisory Committee on Immunization Practices recommendations.16 Children who received HBV vaccine before age 2 months and proceeded to complete a combination Hib-HBV vaccination schedule also were not considered extraimmunized. Finally, combination vaccines that were given in excess for 1 vaccine component but not the other were not considered errors. However, if any component of a combination vaccine (such as the tetanus component of a DTP-Hib immunization) was given too early, then the immunization was considered an error. These rules were consistent with established recommendations.25,26
Immunization Schedule Definitions
Several changes to immunization recommendations that affected the definition of a valid immunization occurred between 1994 and 1999 (Table 1). Because the definition of an invalid immunization depended not only on the prevailing recommendations at the time of immunization but also on the type of vaccine being used, we categorized children accordingly, when possible, so that a single set of criteria could be used to assess the appropriateness of immunization timing.
For example, for polio immunization, children were designated as belonging to an all-OPV schedule if they received only OPV, an all-IPV schedule if they received only IPV, and an IPV-OPV sequential schedule if they received IPV for their first 2 immunizations and OPV for their third immunization. At times, a nonstandard combination of IPV and OPV was used (<2% of cases). For these cases, if IPV followed OPV or OPV followed IPV, then the interval was 4 weeks; if OPV followed OPV, then the interval was 6 weeks. These intervals were applied according to either explicit recommendations or recommendations suggested by the standard immunization schedules.
Children who received all-DTP, all-DTaP, all-DTP-Hib, or alternative combination regimens were defined as being on the correspondingly named schedule. However, the type of DTP schedule did not affect how the timing recommendations were applied. For Hib immunization, children who received their first 2 Hib immunizations as Hib-OMP were designated as being on the Hib-OMP schedule. A third Hib shot given before 12 months to children on a Hib-OMP schedule was considered an invalid immunization. Children who received a mixture of Hib-OMP and Hib-tetanus toxoid conjugate vaccine or Hib conjugate vaccine were not considered to be on a Hib-OMP schedule. Similarly, children who received their first 2 HBV immunizations as a combination Hib-HBV vaccination (disregarding HBV vaccine given at birth) were designated as being on a combination Hib-HBV schedule. Children who received a mixture of combined Hib-HBV vaccine and monovalent HBV vaccine were considered to be on a noncombination schedule.
Statistical Analyses
Confidence intervals (CIs) for estimated means were calculated with sandwich estimation of the variance, with clustering according to site and schedule type. Relative risk estimates were obtained by using the Cochran-Mantel-Haenszel method, with stratification according to site, schedule type, race, gender, and Medicaid status. Trends in immunization error rates over time were compared by using Poisson regression, with clustering according to site and schedule type. Statistical analyses were conducted with Stata (College Station, TX) and SAS (SAS Institute, Cary, NC) software.27
| RESULTS |
|---|
|
|
|---|
|
|
Among children who received an invalid immunization, 82.7% (95% CI: 73.492.0%) received no make-up dose and were underimmunized, whereas the remainder received a make-up dose and were extraimmunized. Among children who received an extra immunization, 80.7% (95% CI: 64.297.2%) of extra doses resulted from superfluous administration of the vaccine, whereas the remainder resulted from make-up doses after an invalid immunization. The number of children who were both extraimmunized and underimmunized for the same series (as a consequence of multiple invalid immunizations) was negligible (<0.1%).
Table 4 compares outcomes across the 4 HMO sites. The identities of the sites were purposely not disclosed. Significant differences in the percentages of children who experienced errors were observed among the HMO sites. The most significant difference was in the percentage of children with at least 1 invalid immunization, which appeared to be driven by a high percentage of children who received an invalid fourth DTP/DTaP immunization.
|
|
|
| DISCUSSION |
|---|
|
|
|---|
The most common errors that occurred in this sample were missed opportunities for the fourth Hib immunization, invalid administration of the fourth DTP/DTaP dose, and superfluous administration of polio vaccine. Invalid fourth DTP/DTaP immunization was typically not followed by a make-up dose, in contrast to missed opportunities for the fourth Hib immunization. Superfluous polio immunization has decreased precipitously since the introduction of IPV.12 Therefore, the error currently likely to have the greatest impact on childhood immunization status seems to be invalid administration of the fourth DTP/DTaP vaccine.
We found that late third DTP/DTaP immunization was associated with invalid fourth DTP/DTaP immunization. Administration of the fourth dose <6 months after the third caused the vast majority of invalid DTP/DTaP immunizations. When a late third immunization occurs, it may prompt providers to try to bring the child up to date too hastily, leading to an invalid fourth immunization. The median time elapsed between the third and fourth doses was 5.5 months. Although the likelihood of increased adverse consequences resulting from administration of the third and fourth DTP/DTaP doses 5.5 to 6 months apart is likely to be small, many children had the 2 doses administered <5 months apart, which could theoretically increase the risk of adverse events and compromise immunity.
We observed a high frequency of immunization errors, consistent with previous studies. Other studies found that
835% of children receive invalid immunizations.10,11,15 An analysis of the frequency and financial impact of extraimmunization among US children found that 21% of children surveyed were extraimmunized, which resulted in excess annual costs of $26.5 million.9 Finally, one review concluded that immunization coverage levels could be increased by up to 30% by avoiding missed opportunities. Major factors leading to missed opportunities were lack of provider knowledge of immunization status and misunderstanding of appropriate contraindications to immunizations.28
Perfect adherence to age and interval recommendations may lead to imperfect compliance with immunization recommendations. The implementation of immunization registries (either locally or nationally) and clinical intervention or feedback systems, especially ones paying special attention to minimal age requirements and between-dose intervals, are important steps to be considered in reducing immunization errors.2935 Poor patient record tracking and availability, insufficient provider knowledge regarding vaccine contraindications and immunization status, having >1 provider administering immunizations, and the complexity of current immunization guidelines have been shown to contribute to immunization errors.30,31,36 At each HMO involved in this study, reminder systems, registries, and other efforts to reduce immunization errors were implemented to varying degrees during the study period. Such systems could reduce our estimate of errors relative to those likely to be encountered in other clinical settings. Conversely, organized efforts to increase immunization rates (eg, to meet National Committee for Quality Assurance standards) might lead to overzealous immunization practices, resulting in more errors. Our findings suggest that the focus of such interventions may need to be modified to address specific immunization errors.
This study had several limitations. Our cohort included children who were continuously enrolled in 1 health plan for the first 2 years of life. The racial/ethnic identity of a large proportion of the sample was unknown, and the proportion of children receiving Medicaid was below the national average. Use of such a selected cohort, with relatively stable health care coverage, would likely lead to a conservative estimate of vaccination errors. All children in this study were residents of California, Oregon, or Washington, with the majority receiving health care at 2 of the 4 HMOs. Therefore, our results may not be applicable to the general population. However, these children constitute a relatively diverse population representing 4 HMO sites with varying practices.
Using a retrospective automated dataset, we were unable to verify the specific reasons for vaccinating or not vaccinating a child at a given visit. Therefore, we could not account for special circumstances such as vaccine shortages, contraindications to immunization, or parental refusal, which would lead us to overestimate the error problem. The possibility of clerical errors also could not be excluded. However, datasets such as that used in this study have been shown to have a high degree of accuracy, with minimal discrepancies between the automated and paper-based medical records.37,38 Lastly, this particular study period was marked by multiple changes in recommended immunization practices, potentially leading to greater confusion and a high error rate. It would be worthwhile to reproduce these findings for a contemporary cohort.
This study raises important questions about the public health consequences of failure to adhere to immunization guidelines. Despite poor compliance with immunization guidelines, the incidence of vaccine-preventable diseases among US children remains low.2 Generally, data are lacking to show significant direct harm to children from immunization errors. Moreover, extra immunizations in the first 2 years of life might be considered actually beneficial. Therefore, we may ask whether it is truly important to obey these guidelines strictly and whether the current efforts to refine and to enforce these guidelines are warranted, particularly in light of potential costs associated with ensuring total adherence to immunization guidelines. However, reducing errors need not be costly and may be achievable simply with greater public attention to common errors and their risk factors. Furthermore, up-to-date status may underrepresent the true immunity of children to vaccine-preventable diseases and epidemics. To the extent that better compliance with recommendations minimizes this risk, greater attention to improving compliance is warranted.
This study suggests that compliance with national immunization recommendations is low and immunization errors occur frequently, ultimately leading to improper immunization. With increasingly complicated guidelines and expanding coverage against new infectious agents on the horizon, it is important to implement strategies that will lead to improved practice patterns and childhood immunization coverage in the future.
| FOOTNOTES |
|---|
Address correspondence to Robert L. Davis, MD, MPH, Office of Genomics and Disease Prevention, Centers for Disease Control and Prevention, 4770 Buford Highway, Mailstop K-89, Atlanta, GA 30341. E-mail: rad2{at}cdc.gov
No conflict of interest declared.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
J. A. Boom, C. S. Nelson, L. E. Laufman, A. E. Kohrt, and C. A. Kozinetz Improvement in Provider Immunization Knowledge and Behaviors Following a Peer Education Intervention Clinical Pediatrics, October 1, 2007; 46(8): 706 - 717. [Abstract] [PDF] |
||||
![]() |
S. S. Kim, J. A. Frimpong, P. A. Rivers, and J. J. Kronenfeld Effects of Maternal and Provider Characteristics on Up-to-Date Immunization Status of Children Aged 19 to 35 Months Am J Public Health, February 1, 2007; 97(2): 259 - 266. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Kempe, M. F. Daley, U. D. Parashar, L. A. Crane, B. L. Beaty, S. Stokley, J. Barrow, C. Babbel, L. M. Dickinson, M.-A. Widdowson, et al. Will Pediatricians Adopt the New Rotavirus Vaccine? Pediatrics, January 1, 2007; 119(1): 1 - 10. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||