Measles vaccination of young infants in China: A cost-effectiveness analysis

Introduction

Measles is a highly communicable disease that can lead to severe illness, permanent disability and death.¹ In the pre-vaccine era, frequent measles epidemics resulted in an estimated 30 million cases and 2 million deaths annually worldwide.² Since the inclusion of measles vaccine in universal vaccination programs in the 1960s, global measles incidence has declined by over 95%, with associated deaths reduced by over three-quarters.^1,2 In 2012, the World Health Assembly adopted the goal of achieving measles elimination in 5 of the 6 World Health Organization (WHO) Regions by 2020 as part of the Global Vaccine Action Plan (GVAP).² Despite significant investment to reach high measles vaccination coverage globally, progress towards regional measles elimination goals has experienced significant setbacks in recent years.³ While measles endemic countries continue to be an important source of exported measles cases to other areas of the world, countries where measles has historically been controlled, such as Australia, the UK and the USA, are also observing an increased frequency of measles outbreaks and, in some instances, sustained transmission.⁴ There is an urgent need to reevaluate strategies for measles control to get back on track with regional and global measles elimination goals.

In 2006, the Chinese government established a goal of measles elimination by 2012.⁵ A series of impressive supplementary immunization activities (SIAs) implemented in 27 of the 31 mainland provinces from 2003 to 2009 was followed by a nation-wide measles vaccination campaign in 2010.⁵ Directly following the intensified efforts to accelerate elimination, the incidence of measles decreased from an average 68 cases per 1 million population in the 2000–2009 period to 4.6 cases per 1 million population in 2012, the lowest level ever reported in the country.⁶ However, measles incidence spiked again in 2013, where cases originating in China were eventually responsible for 44–80% of the total measles cases reported across the Western Pacific Region through 2017.⁶

China’s measles vaccination strategies have evolved from administering measles vaccine exclusively via mass immunization campaigns in the late 1960s and early 1970s to incorporating one dose of measles vaccine, administered at 8 months of age, into the Expanded Program for Immunization (EPI) recommended schedule in 1978.⁵ Subsequently, the Chinese immunization program adopted a two-dose measles schedule with administration at 8 months and 7 years of age in 1986, and then moved the second of the two doses to 18–24 months in 2005 to ensure broader protection among young children who fail to seroconvert following the first dose.⁵ Since inclusion of measles vaccine in the EPI, there had been a consistent bimodal age-distribution of measles cases, with the largest proportion occurring in children under-5 and in young adults aged 15 to 40 based on cases reported to China’s notifiable disease surveillance system.^5,7 However, more recent outbreaks in China have demonstrated a marked shift in the epidemiologic profile of measles with a disproportionately high number of cases occurring in younger infants. During 2009–2013, for example, approximately 20% of measles cases in Tianjin, China occurred in children under the age of 8 months who are age-ineligible for vaccination based on China’s recommended immunization schedule at the time.⁸ Another study using surveillance data reported a steady increase in measles incidence among children under-1 year of age in Shanghai over the past 30 years.⁹

Several clinical and field studies have revealed rapid decay of maternal antibodies transferred to young infants and, in the Chinese context, eventual absence of protective antibodies prior to 8 months when the child first becomes eligible for measles vaccination. A recent investigation¹⁰ among 605 infants in Tianjin, China found that at 3 months of age, nearly all children in the sample lacked adequate levels of protective measles antibody (<200 ug/mL); while a study¹¹ from Belgium showed evidence of maternal antibody waning among infants born to mothers in the vaccine era as early as 1 month following birth.

The WHO recommendations for initiating the measles-containing vaccine (MCV) dose series at 9 or 12 months aim to minimize the risk of primary vaccine failure due to interference from maternal antibodies acquired via passive immunity, the possibility of which is greater prior to 9 months, balanced against the desire to immunize the infant for measles at as early an age as possible since severe disease and death are more likely among young infants.² However, in settings such as China, where there is evidence of waning or absent maternal antibody protection among infants under 8 months of age combined with a disproportionately high burden of disease in this age group, consideration may need to be given to adding a measles vaccine (MV) dose to the immunization schedule for infants younger than 8 months of age. Importantly, however, this would represent additional costs to the immunization program. One of several criteria for making changes to the immunization schedule is the cost-effectiveness of such a policy change.¹² By considering the cost to health benefit trade-offs, policymakers are able to make informed decisions about how to optimize scarce health resources. This study aims to assess the cost-effectiveness of adding a single MV dose to the infant schedule in China <8 months of age compared to the current, status quo strategy of initiating the measles vaccination at 8 months.

Materials and methods

A cost-utility analysis was conducted to evaluate the costs and health effects of recommending an additional MV dose for younger infants compared to the existing strategy of initiating measles vaccination at 8 months of age in Chinese children. The proposed policy change would add a new, ‘0’, dose of MV, thus shifting the initiation of MV from 8 months to younger infant ages. However, the younger infant age dose does not replace future doses, as recommended by WHO, and therefore the comparison is between dose ‘0’ plus dose ‘1’ versus only dose ‘1’ among children < 9 months of age. Age at administration for the ‘0’ dose was considered for children aged 4, 5, 6 and 7 months. The projected costs and health benefits associated with adding the ‘0’ dose to the schedule were compared to the existing costs and health benefits of the current schedule. Disease and economic burden were assessed only among children 4 – 8 months, assuming that no changes in costs and health effects would be observed in children 9 months and older subject to the existing vaccine schedule where a second dose of measles is recommended at 18–24 months of age. The analysis perspective took into account direct healthcare costs and immunization program costs. This perspective excludes costs borne to families and caretakers related to indirect medical and non-medical costs. Primary outcomes included quality-adjusted life years (QALYs) gained, costs borne to the program following the policy change, costs associated with changes in healthcare use, and the incremental cost-utility ratio (ICER). We chose to evaluate the health impact in terms of QALYs gained in order to measure the effect of earlier vaccination on the combination of both clinical cases and deaths associated with measles. While ICERs are generally reported as the cost per unit change in healthy-adjusted life years¹³, we additionally reported the total annualized costs, clinical cases and deaths associated with the current and alternative immunization strategies in a supplementary table. All costs are reported in 2018 US dollars. Costs and health outcomes are discounted at 3%, as is recommended by the WHO.¹³

Model

A decision analytic Markov model was developed in Microsoft Excel with monthly cycles. Individual infants started off healthy at 4 months old. Each month, depending on the policy, they could receive a measles vaccine and potentially acquire a measles infection. We modeled mutually exclusive outcomes associated with infection as healthy, non-severe health complications due to infection (e.g. diarrhea), acute otitis media, and more severe complications: pneumonia, encephalitis, and subacute sclerosing panencephalitis (SSPE). Fatal outcomes were only possible for pneumonia, encephalitis and SSPE cases. Individuals could recover from pneumonia and encephalitis, though often the primary cause of measles-associated death among children, but SSPE was modeled as exclusively resulting in death. Non-fatal encephalitis cases were assumed to have a probable chance of long-term neurologic sequelae (Figure 1). We simulated until the individuals reached 9 months of age but accounted for remaining lifetime costs and QALYs based on average life expectancy and the individuals being alive with no sequelae, alive with long-term sequelae, or dead.

Figure 1.

Pathways for costs and health outcomes in the model utilized for the analysis of adding a measles dose to the Chinese infant vaccination schedule at 4, 5, 6 or 7 months versus the current schedule among children 4–8 months of age

Inputs and assumptions

Data sources, inputs and assumptions for the base case scenario and parameter ranges are described below and summarized in Table 1.

Table 1.

Model parameters, estimates, ranges and information sources for the analysis

Variable	Estimate	Lower limit	Upper limit	Source (Range)
Measles burden and health outcomes
Risk of measles infection (monthly)*
4 months	0.000129	0.000116	0.000142	Wang et al. 2015 (+/− 10%)
5 months	0.000132	0.000119	0.000145	Wang et al. 2015 (+/− 10%)
6 months	0.000223	0.000201	0.000245	Wang et al. 2015 (+/− 10%)
7 months	0.000308	0.000277	0.000339	Wang et al. 2015 (+/− 10%)
8 months	0.000487	0.000438	0.000535	Wang et al. 2015 (+/− 10%)
Clinical outcomes among infected**
Non-severe (e.g. diarrhea)	0.08	0.072	0.115	US CDC 2015 (−10%/Perry & Hasley 2004)
Otitis media	0.07	0.063	0.14	US CDC 2015 (−10%/Perry & Hasley 2004)
Pneumonia	0.06	0.054	0.086	US CDC 2015 (−10%/Perry & Hasley 2004)
Death	0.02	0.018	0.03	US CDC 2015 (−10%/Perry & Hasley 2004)
Encephalitis	0.001	0.0009	0.002	US CDC 2015 (−10%/Perry & Hasley 2004)
Death	0.15	0.135	0.25	US CDC 2015 (−10%/Perry & Hasley 2004)
Neurological or audiological sequelae	0.25	0.225	0.33	US CDC 2015 (−10%/Perry & Hasley 2004)
Subacute sclerosing panencephalitis	0.0000075	0.00000675	0.00001	US CDC 2015 (−10%/Perry & Hasley 2004)
Death	1	-	-
One-time utilities***
Non-severe (e.g. diarrhea)	0.984	-	-	Deng et al. 2019
Otitis media	0.995	-	-	Che et al. 2014
Pneumonia	0.9921	-	-	Bennet et al. 2000
Encephalitis	0.9768	-	-	Bennet et al. 2000
Subacute sclerosing panencephalitis (death)	0	-	-
Annual utilities for sequelae
Neurological sequelae due to encephalitis	0.7393	-	-	Bennet et al. 2000
Vaccination program
Efficacy****
4 months	50%	29%	91%	Lochlainn et al. 2019
5 months	67.0%	51%	81%	Lochlainn et al. 2019
6 months	76%	71%	82%	Lochlainn et al. 2019
7 months	72.0%	56%	87%	Lochlainn et al. 2019
8 months	85%	69%	97%	Lochlainn et al. 2019
Coverage
<8 months	100%	-	-	Assumption
Cost per dose
Vaccine (0 dose) administered at 4–7 months	$ 1.00	$ 0.50	$ 4.00	Zheng et al. 2018
Delivery	$ 3.30	$ 1.00	$ 6.00	Yu et al. 2018
Healthcare costs*****
One-time direct costs
Non-severe (e.g. diarrhea)	$ 791.97	$ 593.98	$ 989.96	Zeng et al. 2019
Otitis media	$ 1,037.55	$ 778.16	$ 1,296.94	Zhou et al. 2004
Pneumonia	$ 1,579.21	$ 1,184.41	$ 1,974.01	Zhou et al. 2004
Encephalitis	$ 2,470.25	$ 1,852.68	$ 3,087.81	Zhou et al. 2004
Subacute sclerosing panencephalitis	$ 1,292.17	$ 969.13	$ 1,615.21	Zhou et al. 2004
Annual direct costs
Neurological sequelae due to encephalitis	$ 50,677.54	$ 47,809.00	$ 95,198.00	Zhou et al. 2004

^*Calculated by converting the cumulative proportion of measles infections among children 4–8 months reported in Tianjin, China for the period 2009–2013 to annual risk of infection by month.

^**Clinical outcomes of measles infection were derived from CDC estimates of common sequelae.

^***One time and annual utilities were used from reference literature and incorporated into the model as disutilities.

^****Vaccine efficacy at different ages was estimated by calculating the average loss of protection by month after 8 months of age and using uncertainty ranges from Lochlainn et al. No adjustments for sub-optimal uptake or adherence were made.

^*****Healthcare costs considered direct medical costs associated with hospitalization and outpatient visits for minor complications.

Risk of measles infection

Using data from China’s Tianjin Centers for Disease Control and Prevention surveillance system for notifiable diseases, the cumulative age-specific risk of infection was calculated for the period between 2009 and 2013 among infants under 8 months of age and then converted to average annual rates of infection for the period.¹⁴ According to the 2010 Chinese Census, an estimated 13,645 infants are born each month in Tianjin, which was used as the denominator for calculating the monthly average cumulative risk of measles infection. The epidemiological trends of measles in Tianjin has been similar to mainland China as a whole, and therefore were considered sufficiently representative of the mainland population to extrapolate findings from Tianjin.⁵

Clinically relevant health states

Databases including PubMed, Google Scholar and Embase were used to identify sources for the probability of clinical complications due to measles infection as well as one-time disutility implications for the calculation of Healthy-related Quality of Life. Additional references regarding the health effects associated with measles infection were identified via reference list reviews of the relevant cost-effectiveness literature.

Vaccination effectiveness

Although China uses a domestically produced vaccine, which contains a different MV strain (Shanghai (Lu) −191) compared to the strains used for vaccines in most regions of the world (Edmonston), studies have shown that its immunogenicity and efficacy is similar to that of other measles vaccines.^15,16 Few studies have assessed the efficacy profile of MV when administered to children <9 months of age due to the documented potential for vaccine failure due to interference between vaccine-derived antibodies and maternal antibodies¹⁷ as well as the more limited ability of young infants to mount strong immune responses as compared to older infants. The WHO recently summarized the sparse evidence available on the immunogenicity, efficacy and effectiveness of immunizing children under < 9 months against measles.¹⁸ Using the WHO commissioned report, we conservatively assume 85% protection against measles infection at 8 months of age, with a 7.5% loss in protection for each month when the vaccine is administered to infants younger than 8 months of age.^2,18 Estimates from the recently published meta-analysis on the immunogenicity, effectiveness and safety of vaccinating young infants against measles were used in sensitivity analysis to assess the influence of the reported upper and lower bound ranges.¹⁹

China consistently reports high infant vaccination coverage for freely available vaccines recommended in their Expanded Program on Immunization²⁰. We would expect this to also apply to an additional ‘0’ dose of MV since it would not require an additional vaccination visit for most age target groups assessed because there are other vaccines administered at 4, 5 and 6 months.²¹

Vaccination program costs

Incremental costs to the vaccination program included costs associated with procuring the additional measles vaccine stock and injection supplies as well as associated delivery costs for adding a new dose to the routine infant schedule. Recent cost analysis of vaccine purchase conducted by Zhejiang CDC reported the MV ranged in price from US$ 0.94 to US$ 3.39 per dose.²¹ Another study on the cost of routine immunization program at the national-level in China found that the total cost, including costs borne at the provincial, prefecture, county, township and village level, totaled to US$ 92.42 per child vaccinated, which represents an administrative cost of US$ 4.20 per child per dose received²² We used this delivery cost portion of this estimate (US$ 3.30) and the costs of vaccine and related supply purchase to project the incremental cost of including another dose of vaccine into the schedule.

Healthcare costs

Data on the cost of healthcare resource use and out-of-pocket expenditures due to measles infection and related-complications in China is limited in the literature. Direct medical costs for hospitalization were collected and reported in a recent cost-effectiveness analysis.²³ This analysis used costs collected from their work in Zheijiang Province and aggregated their collected costs to complication costs reported in the US.²⁴ All costs were converted to 2018 US dollars.

Sensitivity analysis

One-way sensitivity analysis was assessed, varying each parameter within the defined value ranges, to determine influential parameters. Multi-way sensitivity analyses were also conducted to represent a number of ‘what if’ scenarios, including a ‘favorable’ and an ‘unfavorable’ scenario. We varied vaccine price and effectiveness, vaccine program delivery costs, and risk of measles acquisition by age to assess the influence of the range of plausible estimates for these parameters on the ICER. In addition to assessing parameter uncertainty, we also explicitly evaluated separate ‘base cases’ to calculate the differential ICER of initiating the measles vaccination series with a ‘0’ dose at 4, 5, 6 or 7 months of age, with the aim of selecting the age at which the policy change offers the most favorable cost to health benefit tradeoff.

Results

From 2009 to 2013, a total of 588 and 30,120 measles cases were reported in Tianjin and throughout China mainland, respectively, among children under eight months of age who, under the current recommendations, are not eligible for vaccination. Using this surveillance data²⁵, our model estimated that the annual occurrence of measles cases in children 4 to 8 months of age would decrease by 57–72% in the presence of recommendations to incorporate a ‘0’ MV dose into the routine schedule compared to the current strategy. The projected number of clinical cases, deaths, QALYs and associated costs, using the underlying surveillance data, per infant are reported for both the current scenario (initiating MV at 8 months) and the alterative scenario (initiating MV as a ‘0’ dose at 4–7 months of age followed by dose 1 at 8 months) in Table 2. While vaccine effectiveness is assumed to be lower when MV is administered to children under 8 months of age, this strategy still results in a significant absolute reduction in cases of measles. Likewise, the cost to treat clinical measles cases is reduced in the presence of vaccination strategies at ages under 8 months. In all age scenarios from our analytic perspective, total treatment-related healthcare costs would be reduced by US$ 2.8 million or more annually for the mainland Chinese population (Supplemental Table 1) due to reducing disease in children under 8 months of age in China. Without the prevention provided by vaccination of young infants, these cases would generally require case management and further health resource use to treat complicated cases and life-long sequelae.

Table 2.

Base case results on costs and health effects (cases, deaths and QALYs) per infant of incorporating a measles dose in China at 4, 5, 6 or 7 months of age compared to ‘do nothing’^┼

Per individual	Age at administration	Without additional dose	With additional dose	Difference due to policy change
Measles cases	4 months	-	0.00036	−0.00092
	5 months	-	0.00038	−0.00090
	6 months	-	0.00042	−0.00086
	7 months	-	0.00055	−0.00072
	8 months (comparator)	0.0012776	-	-
Measles deaths	4 months	-	0.0000005	−0.0000013
	5 months	-	0.0000005	−0.0000012
	6 months	-	0.0000006	−0.0000012
	7 months	-	0.0000008	−0.0000010
	8 months (comparator)	0.0000017	-	-
QALY	4 months	-	30.8435	0.01630
	5 months	-	30.8431	0.01591
	6 months	-	30.8423	0.01515
	7 months	-	30.8400	0.01279
	8 months (comparator)	30.8272	-
Vaccination program costs (US$)	4 months	-	$ 4.30	$ 4.30
	5 months	-	$ 4.29	$ 4.29
	6 months	-	$ 4.28	$ 4.28
	7 months	-	$ 4.27	$ 4.27
	8 months (comparator)	-	-	-
Healthcare costs (US$)	4 months	-	$ 0.20	$ (0.51)
	5 months	-	$ 0.21	$ (0.50)
	6 months	-	$ 0.23	$ (0.47)
	7 months	-	$ 0.31	$ (0.40)
	8 months (comparator)	$ 0.70	-	-
Net costs, i.e. the sum of program and healthcare costs (US$)	4 months	-	$ 4.50	$ 3.79
	5 months	-	$ 4.50	$ 3.79
	6 months	-	$ 4.51	$ 3.81
	7 months	-	$ 4.57	$ 3.87
	8 months (comparator)	$ 0.70	-	$ -

^┼Cost and health effects compare adding a ‘0 dose’ to the current Chinese measles vaccination schedule at 4, 5, 6 or 7 months of age to the current Chinese measles vaccination schedule alone (2 doses at 8 and 18 months). Costs and health effects are measured for children 4–8 months with and without the alternative policy change of vaccinating with a ‘0 dose’ at 4, 5, 6 or 7 months. Health and economic consequences associated with the alternative scenarios (adding a measles dose at 4, 5, 6 or 7 months) are compared to the observed number of cases at 8 months of age in the presence of the current schedule and the modeled number of clinical events and costs associated with the existing burden.

Administering an additional dose of MV to children under 8 months of age cost US$ 232.70 to US$ 302.39 per QALY gained when compared to the current strategy of initiating vaccination at 8 months of age (Table 3). Incremental health gains, ranging from 0.013–0.016 unit gain in QALY per child in the scenario when MV is administered at 4–7 months of age, are driven by the prevention of severe complications which may affect a child’s functional abilities and developmental trajectory over the course of his or her lifespan (Table 2). Vaccinating children at the ages of 4 or 5 months yields the lowest cost per unit of health gain at US$ 232.70 and US$ 238.39 per QALY gained, respectively (Table 3). This cost could be further decreased given that other vaccines are recommended at this age which would reduce the programmatic delivery costs and therefore overall incremental net costs. There is no universal consensus regarding how to define a country-specific threshold cutoff that deems an intervention cost-effective or not according to its cost per QALY gain, nor is there a publicly available, explicit threshold for China^26,27 Preventive measures where the cost per QALY gained is reported at or below the per-capita Gross Domestic Product (GDP) is generally accepted favorably for vaccine policy considerations. Many studies of the cost-effectiveness of vaccines in China use per-capita GDP as a cutoff.^28–30 With the inclusion of a MV dose at ages under <8 months, despite annual incremental costs reaching upwards of US$ 18 million, which is driven by the cost of procuring and administering additional vaccine, (Supplemental Table 1), the ICER is well below the reported Chinese per-capita GDP (2018) at US$ 11,311.

Table 3.

Base case results for the cost-effectiveness of administering a '0' MCV dose compared to initiating the measles vaccination schedule at 8 months in China

Age of Vaccination	Costs per infant (US$)	QALYs per infant	ICER (US$)┼
Adding a measles dose for infants
4 months	$ 4.50	30.8435	$ 232.70
5 months	$ 4.50	30.8431	$ 238.39
6 months	$ 4.51	30.8423	$ 251.22
7 months	$ 4.57	30.8400	$ 302.39
Comparator (no additional dose)
8 months	$ 0.70	30.8272	-

^┼ICER = Incremental cost-effectiveness ratio compares the net costs and QALYs gained per infant when a ‘0 dose’ is added to the current measles vaccination schedule in China to the net costs and QALYs gained per infant under the current strategy alone, where the ICER = (cost at X months – cost at 8 months)/(QALYs at X months – QALYs at 8 months) for each “age of vaccination” scenario for adding a measles dose at X months.

Assessing the sensitivity of the ICER to changes in key parameters suggests that the policy change to administer measles at a younger infant age would consistently result in additional health gains at a minimal cost. Two of the ten scenarios evaluated were nearly cost-saving, with ICERs < US$ 100 per QALY gained, while higher delivery administration costs (where the cost-per-dose to administer ‘0’ MV dose increases from US$ 3.3 to US$ 7.3) and discounting the future health benefits and costs at 10% instead of 3% resulted in modest increases to the ICER of the alternate strategies (Figure 2). The highest ICER in these scenarios “unfavorable” to the policy change (e.g. assume lower vaccine effectiveness, lower disease risk, and lower healthcare costs) was US$ 713 per QALY gained (Figure 2). One-way sensitivity analysis of the range of plausible estimates for the effectiveness of measles vaccine to protect against clinically relevant disease in children 4–8 months showed ICERs would range from US$ 356 per QALY when considering the lower bound assumption for effectiveness of vaccination at 7 months (45%) to US$ 170 per QALY when considering the upper bound assumption for effectiveness of vaccination at 4 months (91%) (Figure 3). Across the range of plausible assumptions for vaccine effectiveness estimates, administering an additional dose of measles vaccine <8 months of age yields a favorable incremental cost-effectiveness ratio.

Figure 2.

Sensitivity of the incremental cost-effectiveness ratio for vaccinating infants with a ‘0’ dose at 4 months to changes in key parameters

Figure 3.

Sensitivity of the incremental cost-effectiveness ratio for vaccinating infants with a ‘0’ dose to the lower bound, mid-point and upper bound values^┼ considered for vaccine effectiveness by age

Discussion

Ongoing outbreaks and an increasing burden of measles cases in infants who are currently not age eligible for vaccination may slow or even thwart China’s plans for achieving measles elimination, a goal first announced by their government in 2006. While measles elimination in China would be a significant domestic public health achievement, it would also have the potential to catalyze global elimination and eradication efforts given China’s status as the most populous country in the world. In 2017, WHO recommended measles vaccination to children under 6 months of age in outbreak settings.² A complicating factor is that the definition of “outbreak” is not clearly defined and is contextual depending on the epidemiology of measles in a particular country (e.g. the US CDC defines an outbreak as a single case in the U.S.).³¹ Available evidence suggests that sustained, secondary case chain transmission of measles can erupt into a national crisis if outbreaks, or substantial increase in the number of cases above what is generally expected, are not quickly and adequately addressed.³² Including an additional dose of MV at 4–7 months of age in China, particularly in settings with a high burden of disease in early aged infants, is one proposed strategy for rekindling the Chinese elimination plan and getting it back on track.

The cost to health benefit tradeoff of adding a ‘0’ dose to the infant measles schedule in China was shown to be a favorable option for reducing the public health burden of measles among children under 8 months. While vaccination at all ages in the range from 4–7 months offered incremental health gains at a low cost increase, vaccination at ages 4 or 5 months offered the most favorable cost per unit of health gain. Our model simulations estimate that approximately 5000 cases or more each year could be prevented when extrapolating to all of mainland China (Supplemental Table 1). Additionally, measles vaccine is a relatively cheap vaccine compared to newer vaccines or to alternative preventive measures. Given that other vaccines are recommended at ages 4 and 5 months in the Chinese routine immunization program, the addition of a measles doses at these same ages would not be anticipated to result in substantial additional resource needs.

Epidemiological studies have consistently shown that high vaccination measles coverage combined with low disease transmission in the pre-elimination setting may pose challenges to ensuring maternal antibody transfer effectively provides protection to young infants. Several factors determine antibody protective levels in infants following their birth, including gestational age, maternal age and nutritional status, levels of natural boosting in the population due to presence of measles circulation, and the rate of decay of maternal antibodies.^33–35 A study from Belgium, where measles is currently endemic, found that at 6 months of age, 5% of children born to non-vaccinated mothers had sustained immunity against measles whereas only 1% of children born to vaccinated mothers were protected.³⁴ Reports from longitudinal sero-surveillance of mother-infant pairs in China suggest that infants experience rapid decline of maternal antibodies by as young as 3 months, leaving a significant interval of susceptibility prior to the first scheduled vaccination at 8 months.¹⁰ Extrapolating from these findings, infants in China may go unprotected for 1 to 2 months or more before receiving MV at the age of 8 months under the current recommendations. Additionally, infants who may be susceptible as early as age 3 months represent a potential source of disease transmission, particularly in sub-populations without adequate vaccination coverage levels. Infants who contract measles are also more likely to experience serious disease given that their immune systems are more immature and not able to fully respond to infection. The utility of an earlier measles vaccine dose may depend on childcare characteristics. In China <2% of families use childcare outside the home or hire some sort of childcare provider to come inside their home³⁶. Most infants and children are cared for by parents or grandparents. In settings with more mixing of infants and young children in daycare settings, then administration of an earlier measles vaccine dose may be even more beneficial.

Shifting the current two-dose measles vaccination strategy recommended in China to an earlier age, i.e. moving the first dose at (8 months) to a younger age, may appeal to policymakers from a cost perspective where costs at the 8 month dose are just shifted to an earlier vaccination visit.. To the authors’ knowledge, no other studies in the English or Chinese literature have reported the cost-effectiveness of adding an additional ‘0’ dose to the measles infant schedules in spite of the substantial policy interest in addressing China’s measles burden in children <9 months of age.

However, there is not a clear understanding of the effect that the earlier vaccine dose may have on the epidemiology of measles. Few studies have examined the efficacy of the measles vaccine at the youngest infant ages, and even fewer have assessed long-term protection from early infant vaccination.^19,37 The potential of vaccine failures due to maternal and vaccine-derived antibody interference in infants and/or the inability of the immune system in young infants, which may not yet be fully developed at a young age, to mount a protective response to vaccination, makes the addition of a ‘throw away’ dose (i.e. an additional dose prior to the recommended age of initial receipt) the only potential option for vaccinating infants without the risk of waning protection or eventual absence of protection over time as the infants age.

While adding a ‘0’ or ‘throw away’ MV dose to the routine schedule would require an additional investment in the national immunization program, the healthcare savings associated with preventing clinical complications may make for a good return on investment. We note that we have not considered time costs related to productivity in this conservative analysis, which would likely increase the favorability of the ICERs. Non-complicated cases of measles require parents to lose time otherwise dedicated to income generating activities. Decisions regarding further resource allocation to measles control would likely consider the ongoing affordability of routine measles vaccination among children <8 months when other resource demands on the health budget may exist. While some countries incorporate the aspect of affordability in setting an explicit willingness-to-pay threshold and others use the per-capita GDP as a proxy, considering their willingness-to-pay threshold equivalent to 1–3 times the per-capita GDP, our use of the proxy measure may not be entirely appropriate for the Chinese context. This may raise additional considerations around the long-term sustainability of investing in earlier measles vaccination compared to competing resource demands. Considerations about other competing resource needs in the health sector will need to be made locally.

Measles vaccination among children 4 to 9 months is considered safe, supported by evidence from a number of developing countries, as well as potentially extra-beneficial to overall mortality outcomes among infants. A number of studies have explored and demonstrated the hypothesized non-specific effects of measles vaccination of reducing all-cause childhood mortality.^38,39 A recent study in the Netherlands found no severe adverse reactions to early vaccination against measles following a shift in recommendations which promoted early vaccination at 6 instead of 14 months in the wake of a measles outbreak.⁴⁰ Without population data on sero-prevalence of maternal antibodies for all Chinese infants, targeted recommendations are challenging. However, infants who experience primary vaccine failure at the ‘0’ dose administration, i.e. interference between maternal and vaccine derived antibodies, would have the opportunity to achieve full immunity with the follow-up doses which already make up the routine measles vaccination recommendations at 8 and 18–24 months. At such a low cost, adding the ‘0’ dose of MV to the existing recommendations is an appealing policy option for addressing the burden among infants <8 months without placing undue demands on immunization programming.

Because the measles vaccine has already been available in China for several decades, there are not many recent studies of its cost-effectiveness. A study from Zhejiang province found measles vaccination in the existing schedule highly cost-effective, with a cost of 170 renminbi ($24) per disability-adjusted life year saved between 1999 and 2007.⁴¹ In another study based on the existing vaccination schedule, it was estimated that 85 renminbi ($12) could prevent one case, and 4,707 renminbi ($667) could prevent one death.²⁸ These studies have shown the vaccine to be highly cost-effective.

There are limitations to this study inherent to the lack of data readily available to estimate the impact of measles vaccination strategies. Surveillance data provides information on the epidemiological trends of measles in the presence of current vaccination strategies. Without data on the counterfactual (i.e. absence of measles vaccination) in China, we restricted our study population to infants who currently are not eligible for measles vaccination. There may be epidemiological implications associated with primary vaccine failure in infants who do not receive follow-up doses in the country and the resulting contribution these infants have to further propagating transmission. Additionally, evidence suggests that opposition to co-administration of vaccines at the same visit is more commonplace among the public in China⁴², which may hinder uptake of MV added as an additional vaccination when other routinely recommended vaccinations are already being administered. An alternative to co-administration could be incorporating a ‘0’ dose measles campaign into existing catch-up campaigns, but this would implicate other program costs than those included in our analysis and require more complex modelling to capture the disease transmission dynamics between younger infants and older infants who already receive vaccination under the current recommendations. Earlier vaccination in infants may result in reduced protection over the long-term, although there is limited evidence to support this.³⁷ Assuming that China continues to achieve high measles vaccination rates at 8 months and 18 months, there should not be substantial concern for early waning of protection among infants subject to the proposed alternate strategy. Nonetheless, conclusions regarding the cost-effectiveness of the alternate strategy of vaccinating with a ‘0’ dose at 4, 5, 6 or 7 months were not changed when the effectiveness of the vaccine was assumed to reach only 50%. Due to limited data on efficacy and long-term effectiveness of administering MV < 9 months, our model relies on broad assumptions regarding incremental gains in protection up to 8 months. Further research is needed to understand the tradeoffs between vaccinating at younger ages to prevent early disease risk and spending resources on a vaccine strategy that may fail in the presence of sufficient maternal antibodies. This could include both adding to the existing observational evidence to confirm that protection against measles for Chinese infants derived from maternal antibodies rapidly wanes in the population as well as more complex modeling to project the impact on long-term population protection when adding an early measles dose to the infant schedule in China, particularly considering the existing concerns about measles transmission in older age groups.

Conclusion

A disproportionate burden of measles among young infants in China and other countries endeavoring to achieve elimination status exists which poses challenges to measles control since these infants fall below the recommended age threshold for measles vaccination. Although transfer of measles antibodies from mother to infant via passive immunity has been depended upon to provide protection to young infants until they are vaccine eligible, this has been repeatedly proven to be an unreliable assumption. Consequently, earlier administration of the measles vaccine to infants < 8 months age may be one potential strategy to achieve measles control and, eventually, elimination. We estimate that earlier administration of the initial measles vaccine between 4 and 7 months would be cost-effective and could significantly reduce the burden of measles-related morbidity and mortality in infants in China and in other countries pursuing measles elimination.