Effect of combination vaccines on completion and compliance of childhood vaccinations in the United States

ABSTRACT Vaccination at age-appropriate intervals increases protection against morbidity and mortality; however, compliance rates among children remain low partly due to a complicated vaccination schedule. Use of combination vaccines reduces the number of injections per visit; however, there is limited evidence quantifying the effect of combination vaccines on vaccination rates. To examine how combination vaccines impact childhood completion (receipt of recommended doses) and compliance (receipt of age-appropriate vaccinations) rates, this study analyzed vaccination data from the 2012 National Immunization Survey (NIS), a nationally representative cross-sectional survey of caregivers of children aged 24 to 35 months in the United States. Vaccines were categorized as combination or single antigen. Vaccine completion was measured at ages 8, 18, and 24 months. Vaccine compliance and time undervaccinated were measured at 24 months. Children who received at least 1 combination vaccine (86%) had a higher completion rate (69%) and compliance with the full vaccine series (4:3:1:3:3:1:4 series) at 24 months (24%) than those who received only single-antigen vaccines (50% and 13%, respectively). Receipt of combination vaccine was associated with an increased likelihood of completing all recommended vaccinations at 24 months (odds ratio [OR] = 2.5; P < 0.001), receiving all vaccinations at age-appropriate times (OR = 2.2; P < 0.001), and less than 7 months undervaccinated (OR = 2.4; P < 0.001). Combination vaccines were associated with improved completion and compliance and should be encouraged among children who are undervaccinated or who received single-antigen vaccines only.


Introduction
Despite the Advisory Committee on Immunization Practices' (ACIP's) recommendation that children receive numerous vaccinations by 2 y of age, 1 compliance with the recommended vaccination schedule remains low. 2,3 Evidence suggests the potentially large number of injections per visit and the complicated timing of each dose in the recommended childhood vaccination schedule can be cumbersome for parents and may lead to vaccination delay. [4][5][6][7] For providers, developing a modified or catch-up schedule for children with delayed vaccination can be difficult due to the complicated timing of doses, thus resulting in missed opportunities for vaccination and further vaccination delay. 8 Therefore, the ACIP suggests the use of combination vaccines (multiple antigens administered in the same syringe) as an effective strategy to simplify vaccination administration, thusly reducing barriers associated with simultaneous injections and the complicated dosing schedule. 9 Several studies have quantified the effects of hepatitis B combination vaccines on the timely receipt of childhood vaccinations in the United States (US). 10,11 A retrospective study of administrative Medicaid claims compared completion rates between children who received at least 1 dose of the diphtheria, tetanus toxoid and acellular pertussis adsorbed, hepatitis B (recombinant), and inactivated poliovirus (DTaP-HepB-IPV) combination vaccine and children who received no doses of combination vaccine. 10 Results showed that children in the combination cohort had a higher completion rate for the recommended vaccination series as well as higher rates of individual antigens (i.e., DTaP, HepB, and IPV) compared with children in the reference cohort. Additionally, 45% of the combination cohort received all the recommended vaccinations at age-appropriate times, compared with 38% of the reference cohort. However, evidence related to the effect of other combination vaccines compared with single-antigen vaccines on completion and timeliness rates remains limited.
The objective of this study was to examine the effect of all currently available combination vaccines on completion of the recommended number of vaccine doses between birth and 2 y of age and to determine the age-appropriate receipt of each dose of childhood vaccination (compliance) among a nationally representative sample of children in the US.
vaccine. Most of the mothers of the sampled children were aged 30 y or older (57%), married (63%), residing in the South (38%), and educated beyond high school (55%). About half of children were male, 53% were non-Hispanic white, and 75% had siblings in the household. Most children had 1 vaccine provider (67%), and most providers were based in private practice (58%). About half of children were enrolled in Medicaid or the SCHIP.
Demographic differences were observed between children who received at least 1 combination vaccine compared with those who received single-antigen vaccines only. Among children who received single-antigen vaccines only, a greater percentage of mothers were aged over 30 y (64% vs. 56%; P D 0.007), were married (68% vs. 62%; P D 0.015), graduated college (40% vs. 30%; P < 0.001), and had household incomes above the poverty line (65% vs. 58%; P D 0.007).

Vaccine completion
For the majority of vaccines, children who received at least 1 combination vaccine had significantly (P < 0.001) higher completion rates compared with those who received single-antigen vaccines only (Table 2). Differences ranged from a low of a 16 percentage point difference for completion of pneumococcal conjugate vaccine (PCV) at 24 m to a high of a 20 percentage point difference for completion of IPV at 24 months. Completion of IPV, MMR, Hib, HepB, varicella, rotavirus, and PCV at 24 months was significantly lower among children who received single-antigen vaccines only compared with those who received combination vaccines. Completion of the full vaccine series (4 DTaP, 3 IPV, 1 MMR, 3 or 4 Hib, 3 HepB, 1 varicella, and 4 PCV) (4:3:1:3:3:1:4) was lower among those who received single-antigen vaccine only (50%) compared with children receiving combination vaccines (69%).

Vaccine compliance
Compliance rates for children who received combination vaccines was significantly (P < 0.001) higher from those who received single-antigen vaccines only (Table 3). For all vaccines, the proportion of children who received all doses on time was higher among those receiving combination vaccines compared with those who received single-antigen vaccines only. For example, 58% of those receiving combination vaccines received 4 doses of DTaP on time compared with 45% of those who received single-antigen vaccines only (P < 0.001). The largest statistically significant differences were observed for IPV (22%), HepB (22%), and varicella (19%). Children receiving combination vaccines had a higher compliance rate for the 4:3:1:3:3:1:4 series compared with those receiving single-antigen vaccines  All percentage values are weighted based on National Immunization Survey sampling design.
"All" is defined as having received all recommended doses by the measurement period.
"None" is defined as having received zero doses by the measurement period. Partial completion (having received at least 1 (but not all) doses by the measurement period) was a third category included in this analysis and can be calculated by subtracting the sum of "All" and "None" from 100%. only (24% vs. 13%; P < 0.001). For most vaccines, the mean number of days undervaccinated among children who received combination vaccines was significantly lower than among those who received single-antigen vaccines only.

Likelihood of completion and compliance
Findings from multivariable logistic regression on 3 completion and compliance outcomes (i.e., completed series by 24 months, received all vaccines on time, and less than 7 months of undervaccination) indicated receiving combination vaccines was significantly associated with improved outcomes after controlling for individual, household, and provider characteristics (Table 4). For the 4:3:1:3:3:1:4 series, the likelihood of completing all vaccines in the series by 24 months was 2.5 times greater for children receiving combination vaccines compared with those receiving single-antigen vaccines only (P < 0.001). Similarly, compared with children receiving single-antigen vaccines only those receiving at least 1 combination vaccine were 2.2 times more likely to receive all vaccines on time and 2.4 times more likely to have less time undervaccinated (less than 7 months).

Discussion
The present study sought to examine the effect of combination vaccines on completion and compliance of childhood vaccination in the US using a nationally representative survey of households with children. Among children who received at least 1 classifiable vaccine by 24 months of age, we found 86% received at least 1 combination vaccine. The remaining 14% of the sample received single-antigen vaccines with no evidence of a combination vaccine. Children who received combination vaccines were demographically different than those who received single-antigen vaccine only. Specifically, children with single-antigen only vaccines had a significantly higher proportion of mothers who were married, graduated college, and were living above the poverty level. Prior studies have found that parents who refuse or intentionally delay vaccines tend to belong to households with higher income, have married mothers with college educations, and are covered by private health insurance. 12,5 This suggests that children who received singleantigen vaccines are similar to those whose parents intentionally refuse or delay vaccinations. Factors contributing to parental vaccine hesitancy (i.e., concern about whether to vaccinate one's child) have been widely examined. Evidence suggests vaccine refusal and delay may be attributed to perceived risks, perceived benefits, and vaccine safety or efficacy concerns. 4,5, 13-16 As described by Salmon and colleagues, 17 the impact of vaccine hesitancy is on a continuum. Vaccine hesitant parents may intentionally refuse vaccination, delay some or all vaccines (e.g., use of an alternative schedule), or accept vaccination for their child despite concerns. A 2012 survey of parents found that among those with children aged 2 to 6 y, 5.0% of parents delayed and 5.4% refused at least one recommended vaccine. 18 Compared to the same survey conducted in 2014, researchers found the proportion of parents who delayed at least one vaccine remained constant (5.5%) while the proportion of those who refused at least one vaccine decreased significantly (3.3%, P < 0.05). Although the 2012 NIS does not include variables allowing direct measurement of vaccine refusal or delay, we anticipate a proportion of children who were undervaccinated or delayed vaccination could be attributed to vaccine hesitancy. The use of combination vaccines is suggested as an effective strategy for increasing vaccination rates. 9 In our study, we found completion rates were higher among children who received at least 1 combination vaccine. For 4:3:1:3:3:1:4 series, 69% of children receiving combination vaccines completed the recommended doses compared with only 50% of children receiving single-antigen vaccines only. A similar effect was observed in a retrospective study of administrative Medicaid claims that compared completion rates between children who received at least 1 dose of the DTaP-HepB-IPV combination vaccine and children who received no doses of combination vaccine. 10 Results showed children in the combination cohort had a higher completion rate for the recommended vaccination series as well as higher rates of individual antigens (i.e., DTaP, HepB, and IPV) compared with children in the reference cohort.
Furthermore, findings from the current study revealed that children receiving at least 1 combination vaccine had better compliance rates compared with those who received singleantigen vaccines only. Low compliance among children who received single-antigen vaccines only may be in part attributed to fears of simultaneous vaccination. Prior evidence suggests that fears concerning simultaneous vaccination may lead to vaccine avoidance or delay. 4-6 A recent systematic review found that parents raised concerns regarding pain and distress associated with simultaneous injections at a single visit, which may lead to vaccination avoidance or delays. 6 Furthermore, several studies indicate that approximately 76% to 78% of parents who delayed their child's vaccinations had vaccine safety concerns linked to multiple injections. 4,5 Adjusted logistic regression models demonstrated more than a 2-fold increase in the likelihood of being complete or compliant with the vaccine series by 24 months for children who received combination vaccines compared with those who received single-antigen vaccines only. These findings are consistent with those demonstrated in prior studies that compared compliance between children who received specific combination vaccines and single-antigen vaccines. Happe and colleagues (HAPPE 2007) 10 reported that 45% of children who received DTaP-HepB-IPV received all the recommended vaccinations at age-appropriate times, whereas 38% of those who did not receive DTaP-HepB-IPV received their vaccinations on time. Additionally, a study by Marshall and colleagues 11 found that any hepatitis B combination vaccine (DTaP-HepB-IPV or HepB-Hib) was associated with increased coverage rates for several individual vaccines and the full recommended childhood vaccination series.
Our study had several limitations, primarily related to restrictions of the NIS data set. Due to the use of telephone survey methods to identify household respondents and vaccination providers to collect information on vaccination history, estimates obtained from the NIS are susceptible to under-, inaccurate, or incomplete reporting. As completion and compliance measures rely on household-or provider-reported data, there is a possibility of bias in our measurements. Furthermore, the NIS lacks assessment of knowledge, attitudes, beliefs, healthcare seeking behavior, or other factors (e.g., reasons for vaccine refusal, availability/provider recommendation of single-antigen vaccines) related to vaccination that could provide a better understanding of reasons for administration of combination vaccines, receipt of single-antigen vaccines, and/or general vaccine hesitancy. As the 2 cohorts appear demographically different, future research should explore vaccine hesitancy as a potential confounder or moderator of combination vaccine uptake.
The present study fills a critical knowledge gap by evaluating the outcomes associated with combination vaccines. Prior studies that compared combination vaccines with single-antigen vaccines did not evaluate the effect of non-hepatitis B combination vaccines on compliance measures. Moreover, prior studies had not evaluated the effect of combination vaccines on completion and compliance of the full vaccination series recommended by the ACIP. Therefore, findings from the current study provide a more comprehensive understanding of the association between receipt of combination vaccines and completion or compliance outcomes. As vaccine compliance rates among children remain low in US, findings from this study suggest that combination vaccines should be encouraged for some children who are undervaccinated or are receiving singleantigen vaccines to increase compliance. Among the proportion of children with vaccine hesitant parents, particularly those with vaccine safety concerns, new strategies toward increasing vaccination are needed. Findings from the present study indicate differences in compliance between the combination and single-antigen cohorts for vaccines not included in a combination vaccine (e.g., rotavirus), suggesting issues other than simultaneous vaccination may impact vaccine delay/refusal. Interventions aimed at reducing vaccine hesitancy have been widely published; however, a recent systematic literature review found many interventions were not designed or powered to assess impact on vaccine refusal or delay. 19 Therefore, future research aimed at evaluating the impact of strategies aimed at improving vaccination, including increasing the use of combination vaccines (e.g., increasing awareness of the benefits of the use of combination vaccines, educating parents and physicians who may be hesitant to use combination vaccines) is warranted.

Data source
We used data from the 2012 National Immunization Survey (NIS) Public Use File. 20 The NIS, conducted by the National Center for Health Statistics and the Centers for Disease Control and Prevention, has been used extensively for tracking vaccination rates among children in the US. 2,3 The NIS consists of a nationally representative annual household survey of caregivers of young children to collect demographic and vaccination information as well as vaccination history from children's immunization providers. Demographic variables (e.g., race and ethnicity) were taken from the Household Survey as reported by the child's caregiver. Vaccination history (i.e., age in days at vaccination) was taken from the Provider Survey. Detailed NIS methods and Institutional Review Board (IRB) approval for data analysis are reported elsewhere. 21

Study sample
From an unweighted sample of 25 736 children, we selected data for children who had complete survey responses from their caregivers, resided in the US (excluding the US Virgin Islands), and had adequate vaccination data supplied by their vaccine providers. Adequate vaccination data are a variable existing in the NIS denoting whether a child had missing or incomplete provider-reported vaccination data due to a lack of parental consent to contact the child's immunization provider (s), provider nonresponse, or incomplete provider-reported data. Excluding children with inadequate provider-reported vaccinations is a method used by the CDC when calculating annual vaccination completion using NIS data. 23 Children aged less than 24 months at the time of the survey were excluded, thus capturing only children aged 24 to 35 months and allowing a consistent vaccination capture period between birth and 24 months for all children in the sample. Children who were unvaccinated or received only vaccines that could not be categorized as single antigen or combination (i.e., coded as "not otherwise specified") were also excluded (N D 149). The final unweighted sample size was 11 561, equating to a weighted sample of 4 043 159 children.

Vaccine completion
Based on ACIP's 2012 recommendations, 24 we defined vaccine completion as the accrual of the required number of doses by a specific age (8,18, and 24 months) regardless of timing of vaccine administration. Detailed lists of doses required at 8, 18, and 24 months are presented elsewhere. 25 Completion of rotavirus vaccine and Haemophilus influenzae type b vaccine (Hib) doses was assessed based on the product administered. For example, a child's vaccination for rotavirus was considered complete if 2 doses of Rotarix (GSK Vaccines) or 3 doses of RotaTeq (Merck & Co., Inc.) were administered. These algorithms are detailed in the published vaccine schedule. 24 Although influenza and hepatitis A vaccines are also recommended by the ACIP, we did not assess them in the present study. The NIS data do not provide vaccination dates, thus we were not able to determine whether the seasonal influenza vaccine was administered during the influenza season, which is a criterion for administration. The second dose of hepatitis A vaccine should be administered between 6 and 18 months following the first dose at age 12 months or later; therefore, the second dose may be administered after 24 months-falling outside of our study period. As we could not capture appropriate completion and compliance of 2 doses of hepatitis A vaccine, it was not included in the present analysis.

Vaccine compliance
Vaccine compliance is a measure of the timeliness of vaccine administration. The ACIP suggests specific ages and intervals at which vaccines should be administered to maximize effectiveness. 24 Vaccine compliance was measured as the proportion of children who received vaccine doses at age-appropriate intervals (e.g., dose 1 of DTaP at 2 months). Age-appropriate windows were based on the ACIP's recommended dosing schedule. 17,24 Additionally, we measured compliance by calculating the number of days that a child was undervaccinated. This method has been proposed in several prior vaccine compliance studies. 2,3,26 Undervaccination was assessed as having no evidence of a recommended vaccination 1 day following the end of the age-appropriate window. Each day of undervaccination was counted as 1 day, irrespective of the number of vaccines missed by that day. 27 For a child who did not receive a vaccine by age 24 months, we calculated the total number of undervaccinated days as the number of days between the first day of undervaccination and age 24 months. The final measure of vaccine compliance was the total number of days undervaccinated. Accumulation of 7 months or more of undervaccination was categorized as "severely undervaccinated." In addition to reporting total compliance with each dose, we also assessed the degree of compliance, including all, some, and no doses of recommended vaccine received on time.
The ACIP allows early vaccination as "age-appropriate" if received within a 4-day grace period before the minimum age for the dose. 9 Any vaccines administered during a 4-day grace period before the minimum age for the dose were included as on time. Any vaccines administered before the grace period were excluded from calculation of the compliance end point.

Combination vaccines
We categorized all vaccines into 2 mutually exclusive groups: combination or single antigen. Table 5 lists vaccines that were considered combination vaccines. Vaccines with an unknown type (e.g., HI D "Hib, Unknown") were not categorized as combination or single, but were included in calculating the completion and compliance endpoints. All other vaccines were considered to be single antigen.
Although vaccines such as DTaP include multiple components in a single syringe, they were grouped with single-antigen vaccines. Preliminary review of the data indicated that no participants received the individual components of DTaP as single-antigen vaccines, and a small number of participants received a single-antigen measles vaccine; therefore, we anticipated that patterns of administration of DTaP and measles, mumps, and rubella (MMR) vaccines may be closer to those of true single-antigen vaccines (e.g., IPV) than to those of combination vaccines (e.g., DTaP-HepB-IPV).
After each vaccine was coded as combination or single antigen, the vaccination history of each child was reviewed to determine whether each child received at least 1 combination vaccine or single-antigen vaccines only. Next, a binary categorical measure of vaccine type was created to flag each child into 2 mutually exclusive groups: received at least 1 combination vaccine or received no combination vaccines (i.e., received singleantigen vaccines only). This categorical measure of vaccination type was used to determine differences in vaccination completion and compliance between children who received at least 1 combination vaccine and those who received single-antigen vaccines only.

Statistical analyses
Bivariate analysis included evaluation of completion and compliance rates for the overall population and stratified by vaccination type. The Rao-Scott F-adjusted chi-square statistic was used to determine the differences between completion and compliance rates by vaccination type. This statistic is adjusted for complex survey data and provides a more conservative interpretation than the Wald chi-square statistic. 28 Multivariable analysis included logistic regressions on 3 vaccination outcomes: completed vaccines series by 24 months (completion), received all vaccines on time (compliance), and undervaccinated for less than 7 months (compliance). These logistic regression models evaluated the effect of receipt of at least 1 combination vaccine on vaccine completion or compliance after adjusting for demographic characteristics such as maternal age, maternal marital status, maternal education, poverty status, enrollment in Medicaid or the State Children's Health Insurance Program (SCHIP), and provider variables including number of vaccination providers and type of vaccination providers (e.g., public, private).
Estimated frequencies, percentages, means, standard errors, medians, odds ratios, confidence intervals, and P values were calculated using SAS statistical software, version 9.3 (SAS Institute, Inc., 2011). A P value less than 0.05 was used as an indicator that a difference between groups may exist. SAS survey procedures and domain analysis techniques were used to calculate weighted values and correct standard errors per NIS data user guidance. 23

Trademark statements
Rotarix is a trademark of the GSK group of companies. RotaTeq is a trademark of Merck & Co., Inc.