Childhood vaccination improves life expectancy, decreases healthcare costs, and reduces the spread of preventable diseases [1–4]. Routine vaccination averts an estimated 2–3 million deaths globally due to diphtheria, tetanus, pertussis, and measles every year [5]. In the United States alone, routine vaccinations prevented an estimated 322 million illnesses, 21 million hospitalizations, and 732,000 deaths, at a net savings of $295 billion in direct costs and $1.38 trillion in total societal costs, for children born during the period 1994–2013 [6]. International organizations such as the World Health Organization (WHO) implement programmatic immunization programs to ensure high vaccination coverage around the world [7].

The Federated States of Micronesia (FSM) is an island nation spread nearly 2700 km across the Pacific Ocean just north of the equator. There are four island states, Chuuk, Kosrae, Pohnpei, and Yap, and all states except Kosrae are comprised of one main “high” island surrounded by lower lying outer islands. According to the 2010 census, there were 12,073 children under the age of 5 years throughout the FSM states [8]. Researchers from the Centers for Disease Control and Prevention (CDC) partner with local immunization programs to monitor vaccination coverage among children and adults in FSM and other U.S. Affiliated Pacific Island jurisdictions. Technical assistance provided by CDC based on vaccination coverage assessments aid the local programs in assessing the level of vulnerability to vaccine-preventable diseases (VPD) in their jurisdictions, evaluating their program interventions, and developing recommendations and technical support for key stakeholders [9].

Most childhood vaccines are administered in FSM according to the Advisory Committee on Immunization Practices (ACIP) recommendations, with a few exceptions based on the increased burden of some VPDs in the region. The ACIP recommendations include guidelines for vaccination timeliness, or adherence to the recommended timing and spacing of doses, ensuring protection from VPDs as early in life as possible. Late vaccine administration increases the length of time required to obtain adequate protection from VPDs [10]. Previous studies have shown low vaccination timeliness despite high vaccination coverage in many countries [11–17].

While prior CDC assessments found moderate to high vaccination coverage in FSM, they did not account for timing of vaccination in the assessment. In 2016, CDC researchers reassessed vaccination data for 1824 children born between 2007 and 2014. The objective of this study was to examine the traditional measure of total vaccine uptake compared to on time vaccination coverage for four doses of diphtheria and tetanus toxoids and acellular pertussis vaccine (DTaP) and one dose of the measles, mumps, and rubella vaccine (MMR) among children 24–35 months. Additionally, a standard deterministic compartmental model was developed to represent transmission dynamics and estimate the potential impact of a hypothetical measles outbreak in each state based on measured levels of vaccine uptake and timeliness. The assessment of on time coverage in this study provides a more rigorous examination of the nuanced aspects of vaccination coverage than a traditional vaccination coverage assessment; typically vaccination coverage assessments only measure the number of vaccine doses administered by a certain age, regardless of timing and spacing [16]. This study highlights the importance of vaccination timeliness in the context of vaccination coverage assessments, local immunization program objectives, and community protection against the spread of VPDs, and provides information that can be used by the immunization program to improve planning and strategy of vaccination outreach activities.

At least one public health or medical record was available for each child included in the analyses; seven children in the Pohnpei dataset had no written documentation of vaccination, but vaccination events were recalled through caregiver interview. These children were excluded from the analyses. Data were analyzed for 1226 children in Chuuk, 172 children in Kosrae, 188 children in Pohnpei, and 238 children in Yap.

Coverage for the fourth dose of DTaP and the first dose of MMR varied substantially across the four states of FSM. Even when coverage was high, on time coverage was substantially lower, meaning most children who were vaccinated did not receive doses at the recommended age of 12 months. Children in all states received these doses as late as ages 26–35 months; children who were not vaccinated by the end of the data collection period may have received doses even later or not at all. Models indicated increases greater than 10% in measles cases in scenarios where vaccinations were delayed, due to an increased pool of susceptible individuals in the population.

Children who receive vaccinations later than recommended remain susceptible to VPDs until they are vaccinated, which may have a substantial impact on health systems and local economies during an outbreak. Immunization programs should consider vaccination timeliness in addition to vaccine receipt as a measure of susceptibility to VPDs in young children, particularly in areas that have experienced recent outbreaks.

Measurement of up-to-date coverage alone may sometimes obscure substantial vaccination delays that occur during the first two years of life. Risk of disease due to delay in vaccination varies, but understanding vaccination timeliness is especially important for diseases that have the potential to cause large outbreaks, such as measles [25]. Recent VPD outbreaks in FSM highlighted a need for further investigation of vaccination outcomes, such as timeliness. Using methods such as Kaplan-Meier and TSEIR modeling are newer approaches in the context of vaccination coverage and timeliness, which can be utilized by immunization programs to better understand the level of protection against VPDs, and the risk of VPD outbreaks, in their communities.

Inverse Kaplan-Meier graphs provide a useful visual method to estimate the proportion of children vaccinated at any given age, and understand the nature of under-vaccination within a cohort over the first two years of life. Many children in this study were under-vaccinated for DTaP or unvaccinated for MMR for several months after the recommended age of vaccine administration, despite being vaccinated at the end of follow-up. It is important to note that doses were delayed up to two years from the recommendation for both vaccines, even in the states with highest total coverage.

In this study, the proportion of children receiving DTaP doses on time decreased as the schedule progressed, consistent with previous studies [12,15,26,27]. Because this vaccine requires a minimum interval of four weeks between the first three doses, and six months between the third and fourth dose, children late for those first doses will remain late for subsequent doses, and may be more likely to not receive the fourth dose by the age of 24 months, even with adequate time for catch-up [28]. Transplacental immunity wanes rapidly in the first three months of life [29]. The greater the delay during the early months in life, the longer the child remains under-protected against the commonly circulating VPD, pertussis. In studies of pertussis outbreaks, incidence tended to be highest among infants younger than six months, and many cases occurred in young children who were under-vaccinated for age [30,31].

In the case of the first dose of MMR vaccine, delays measured by the Kaplan-Meier method represent a period of complete lack of protection against measles. Investigations of measles epidemics in the United States have found a primary cause was a failure to provide vaccines at the recommended age [32]. Studies of measles outbreaks in England and Wales have shown a significant increase in the reproductive number, or number of secondary infections per infection, occurring immediately following years where there was a decrease in uptake of the MMR vaccine, as a result of a greater number of susceptible individuals in the population [33]. In our study, TSEIR models parameterized according to observed delays in FSM demonstrated the potential impact an increased pool of susceptible individuals in their populations may have in the event of an outbreak.

Since introduction of measles-containing vaccines in the late 1980s, outbreaks of measles cases have been identified in FSM during three different years [34]. From 1991–1994, FSM was part of a large measles outbreak across the region, in which a combined population of about 300,000 experienced more than 1300 cases of measles and 16 deaths due to measles [35]. In 1992, approximately 35 cases were identified. In 1994, approximately 930 cases were identified. More recently, in 2014, a measles outbreak in Pohnpei, Kosrae, and Chuuk resulted in 393 suspected cases, with 140 laboratory confirmed, 244 others linked through epidemiologic data, and 9 clinically compatible. The computed incidence for this outbreak was 2226.6 per million individuals. There was one death attributed to measles in 2014 [36].

An important caveat to consider is the FSM immunization program recommends a second dose of MMR as early as 13 months of age, due to increased risk of measles in the region. The results of this study demonstrated as many as 4 out of 5 children in some regions of the country were not receiving the first dose at the recommended age, and some children were not vaccinated until well into their second to third year of life. Further, recent studies have found increased risk of complications such as febrile seizures in children who received delayed MMR vaccinations [37,38]. This highlights the value of measuring timeliness of vaccination in conjunction with total coverage, as a means of better understanding both the level of VPD protection within a population, and the risk of complications in the population associated with delayed vaccines.

The data collection methods identified in this paper represent the most accurate and reliable methods for estimating vaccination coverage in the low resource setting of FSM, but are subject to limitations. First, data were collected across different years, so results may not be generalizable across states and may not reflect current trends. Second, data collected by administrative surveys in Chuuk, Kosrae, and Yap may underrepresent children who have not accessed the healthcare system and who may be more likely to be un- or under-vaccinated. Third, the Kaplan-Meier method produces slightly higher results than conventional coverage measures as a result of censoring; this method reduces the population “at risk” at the time point when censoring occurs. Conventional estimates of coverage are provided for comparison to Kaplan-Meier results. Fourth, in the delay scenario, the TSEIR model assumes a linear increase toward total coverage actually observed, which impacts the time to completion of the hypothetical outbreak. In the event of a real measles outbreak, it is likely a mass vaccination campaign would more rapidly reduce the number of susceptible individuals in the population and accelerate this timeline; the estimated difference in cases and difference in outbreak duration between the on time and delay scenarios may be overestimated.

This study provides new information about how long children in FSM remain under-protected from commonly circulating VPDs such as measles and pertussis, which may be useful for immunization programs in FSM to set priorities for immunization activities. More generally, this study underscores an important and potentially overlooked opportunity to improve public health and vaccination coverage. The success of an intervention that addresses vaccination timeliness benefits from targeting a portion of the population that is already predisposed and has the necessary healthcare access to obtain vaccinations. Caregivers of these children potentially need only a reminder to ensure vaccinations are occurring at the best time for the protection of the child’s health. As such, timeliness-based interventions may be an inexpensive way to improve vaccination coverage and public health protection from vaccine-preventable diseases. Some evidence-based interventions to consider to improve on time vaccination include: parental education on the importance of timely vaccinations, in conjunction with the use of the existing Immunization Information System which provides reminders to schedule visits at appropriate age and aid physicians in clinical decision support to reduce missed vaccination opportunities; recall of patients who have missed scheduled appointments, or who are otherwise at risk of delayed vaccination; increasing community awareness of vaccination activities; and increasing frequency of vaccination activities in alternate community settings [39]. Immunization programs in island nations with far-reaching outer islands may also benefit from planning regular community outreach to improve timing and spacing of doses for children in hard to reach locations, for whom access to public health clinics may be more difficult.