Since 2008, the Advisory Committee on Immunization Practices (ACIP) recommends all children aged 6 months to 18 years receive annual influenza vaccination, and children aged 6 months through 8 years receive two doses of influenza vaccine during their first vaccination season to achieve adequate immune response [1]. The antigenic components of the influenza vaccine often change yearly to best predict circulating influenza strains. Children who received influenza vaccines in previous seasons likely do not require two doses in the current season; however, in some circumstances, e.g., when a new strain (such as H1N1) is added to an influenza vaccine, two doses are still necessary. Therefore, ACIP also recommends two doses to some children after their first influenza seasons depending on their vaccination histories [2–6].

Poor compliance with the two-dose influenza recommendations has been reported in the literature [7–10]. In 2013, using data from population-based Immunization Information Systems (IIS) from eight sentinel sites, CDC published trends in compliance with two-dose influenza vaccine recommendations among children aged 6 months through 8 years from the 2007–08 through the 2011–12 influenza vaccination seasons [8]. Using IIS data from six sentinel sites, our objectives were: 1) to update the above-mentioned report to show trends in full influenza vaccination coverage and two-dose compliance from the 2010–11 through the 2014–15 seasons, and 2) to investigate timing of dose administration among these children, stratified by influenza vaccine doses recommended and received.

IIS are confidential, population-based databases of consolidated immunization data from vaccine providers [11]. IIS sentinel sites, which include North Dakota (ND), Michigan (MI), Minnesota (MN), New York City (NYC), Oregon (OR, six counties containing 56% of the state population), and Wisconsin (WI) in the 2013–17 period, are selected state/city-based IIS with high data quality. Among these sites, childhood immunization information could be included in the IIS without any consent and with no right to opt out in OR, ND, and NYC. Childhood immunization information could be included in the IIS with implicit consent, with a right to opt out, in the other sites. Three sites (MI, NYC, and ND) mandated all immunization providers to report childhood immunizations to the IIS, Oregon mandated public immunization providers and pharmacies to do so, and there are no mandates in MN and WI [12]. The enrollment of children under 6 years old was >95% for all sites [13]. The sentinel sites are funded by the Centers for Disease Control and Prevention for vaccination coverage evaluation in their pediatric populations, which are approximately 10% of the United States pediatric population.

We analyzed records for >2.6 million children during each influenza vaccination season (July 1 through June 30) from the 2010–11 season through the 2014–15 season. Children were included in the analysis if they were aged 6–23 months, 2–4 years, or 5–8 years throughout each influenza vaccination season. Children aging into or out of an age group were excluded to ensure all children had the same opportunity for vaccination. Children known to have moved out of the state/city jurisdiction or children known to be deceased were excluded from the analysis.

Regardless of product type (e.g., nasal spray or injectable vaccine), we considered a dose valid if it complied with ACIP’s minimum age and interval recommendations for influenza vaccination (i.e., inactivated influenza vaccine given to a child aged ≥ 6 months and the interval to the previous dose in the same influenza season was ≥ 4 weeks, with a grace period of four days allowed for both criteria; live attenuated influenza vaccine given to a child aged ≥ 2 years, administered ≥ 24 days from a previous influenza vaccine dose, and ≥ 28 days from a previous live vaccine). If a child received more than two valid doses in the same season (<0.1% of all children did so in our dataset), the third or later doses were considered extraneous and excluded from analyses. We only counted valid doses for influenza vaccination calculations.

Our primary outcomes were full vaccination coverage and two-dose compliance. Full vaccination coverage was classified as having received at least the recommended number of influenza vaccine doses (one or two), based on ACIP recommendations for that season. Two-dose compliance was defined as the percentage of children during each season who received at least two doses of influenza vaccine among those who required two doses and initiated the series. Our secondary outcomes were partial vaccination coverage, timing of vaccine dose administration in each season, and mean dose intervals between the two doses for each season. Partial vaccination coverage was classified as having received only one influenza vaccine dose when recommended to receive two doses.

For each child in the selected cohorts, we evaluated the child’s eligibility for two doses of influenza vaccines according to ACIP’s yearly recommendations [2–6] by assessing full vaccination histories in the IIS. The percentage of children requiring two doses was calculated by dividing the total number of children requiring two doses by the total number of children in the IIS, stratified by age and influenza season. To calculate vaccination coverage rate, we divided the total number of children who were fully or partially vaccinated by the latest Census population estimates in the area for the corresponding year, and further stratified by age and influenza season. Full and partial coverage combined constituted ≥1 dose coverage. To assess the timing of vaccine dose administration in each season, we calculated the cumulative percentage of all influenza vaccine doses administered by month throughout each season, stratified by children’s status, including children who: (1) required one dose and received ≥1 doses; (2) required two doses and received only one dose; (3) required two doses and received ≥2 doses. For children who required and completed two doses, we also calculated mean intervals between the two doses for each season. We used SAS®, version 9.3 (SAS Institute, Inc.) and Microsoft® Excel® 2010 (Microsoft Corp.) for all analyses.

Our analysis showed increasing trends of full influenza vaccination coverage in children aged 6 months through 8 years from the 2010–11 season to the 2014–15 season, which are consistent with the National Immunization Survey (NIS) data [14]. The overall full influenza vaccination coverage remained low. Compared with NIS data on full influenza vaccination coverage among children aged 6–23 months in the 2011–12 season [14], among the four sites that were comparable geographically, our full coverage rates were within the confidence intervals of NIS estimates for two sites; were very close (0.5% lower) to the lower boundary of the NIS confidence interval for one site; were outside (3.5% lower than the lower boundary) the NIS confidence interval for one site.

Monitoring two-dose compliance can inform vaccine programs of different issues than monitoring vaccination coverage. Non-compliance to two-dose recommendations may indicate vaccine access barriers or unawareness of the recommendations rather than vaccine hesitancy or avoidance because this metric is based on a subset of population that initiated the series. Two-dose compliance increased slightly in children aged 6–23 months from the 2010–11 season to the 2014–15 season but remained largely unchanged in older children. Therefore, the increasing trends of full influenza vaccination coverage in children aged 2–4 years and 5–8 years were likely because fewer children in later seasons were recommended to receive two doses since there was no major change in the influenza vaccine. Of note, the relatively higher two-dose compliance in the 2010–11 season in older groups may be attributed to higher awareness among providers of the new component of pandemic H1N1 strain in the seasonal influenza vaccine, which required more children to receive two doses. This effect was not obvious in children aged 6–23 months because they were less likely to have had the opportunity to receive influenza vaccine in previous seasons and thus required two doses anyway.

There are several possible reasons for low two-dose compliance: (1) the complexity of ACIP’s influenza vaccination recommendation in some years and several changes over the years may make it difficult for providers to realize which children require two doses, especially for older children; (2) providers and/or parents may not know the importance of two dose completion; (3) convenience of access to vaccination may be a barrier to parents with busy schedules, particularly with older children who have fewer wellness visits; (4) providers may not have access to full vaccination history of children to assess eligibility for two doses; and (5) late availability of influenza vaccine may delay initiating first dose and impact two-dose completion [15]. Therefore, improved messaging of the two-dose influenza vaccine recommendation, and convenient access to vaccination (i.e., complementary vaccination settings such as schools, extended office hours) are needed. One study found that campaigns providing better information regarding the vaccine and the disease, and convenient access to vaccination can facilitate influenza vaccination in children [16]. Providers are encouraged to submit data to IIS to enhance the system, and use consolidated vaccination histories in IIS to assess children’s eligibility for two doses, especially for older children, who may be perceived to not need two doses. Clinical decision support functions in IIS and other electronic health systems can also make recommendations to providers on patients’ vaccine needs. In addition, patient reminders can also be considered as they are positively associated with in-office influenza vaccination [17]. Lastly, the lags observed in the timing of dose administration among children who required two doses and received only one dose, compared with the timing of first dose administration among children who required and received two doses, might reflect a subset of the population with differing access to vaccination services, knowledge of vaccination recommendations, or motivation to receive influenza vaccination for their children, but the reasons for this delay would require further investigation beyond the scope of this analysis.

Children who require two doses of influenza vaccine are not optimally protected from influenza disease if they only receive one dose. Studies have shown that although children may develop antibodies after one dose, no protection from disease in terms of clinic visits and lab-confirmed influenza cases has been observed when compared with children who did not receive any dose [18–20]. In the United States, influenza activity can peak from October to March, but since 1976, >60% of the seasons’ peak months were in February or later [2]. It is encouraging that across all seasons, 80% of children who required two doses and received both doses did so by the end of January. However, when influenza activity peaked in late December during the 2014–15 season, only half of these children had completed both doses (Fig. 3). Moreover, the median intervals between two doses were approximately seven weeks. These results demonstrate the importance of administering two influenza vaccine doses when recommended, and completing the two-dose series early in the season whenever possible. Shortening the interval between administration of the first and second doses, within ACIP recommendations, can increase the proportion of children who are optimally protected against influenza disease early in the season.

There are at least two limitations in this analysis. First, data from the six sentinel sites may not represent the U. S. population. Although previously published IIS sentinel site data were very close to the NIS national data on influenza full vaccination coverage in children 6–23 months [8, 14], the current IIS sentinel sites cover different geographical areas and therefore may have different vaccination coverage and compliance rates. Second, incomplete vaccination histories in IIS data may underestimate vaccination coverage and two-dose compliance, especially in older children who visit their physician’s office less frequently but receive influenza vaccination in other places. However, although IIS sentinel sites are only required to have >85% children enrollment, the actual enrollment of children was >95% for each sentinel site [13], which minimizes potential underestimation for influenza vaccination rates.

In summary, two-dose compliance remained low in children aged <9 years, especially in older children (2–8 years) from the 2010–11 through the 2014–15 influenza vaccination season, which impacted full influenza vaccination coverage. In the 2014–15 influenza season, among children who required two doses and received ≥2 doses, only half completed the two-dose series before influenza activity peaked based on data from IIS. Improved messaging of the two-dose influenza vaccine recommendations, convenient access to influenza vaccination, and high levels of reporting of influenza vaccinations to IIS are needed for providers and parents. Providers are encouraged to determine children’s eligibility for two doses of influenza vaccine using consolidated vaccination histories in IIS. Moreover, early vaccination and shortened two-dose intervals within ACIP recommendations may facilitate early completion of the two-dose series so that more children are protected before influenza activity peaks.