Adding interventions to mass measles vaccinations in India

Abstract Objective To quantify the impact on mortality of offering a hypothetical set of technically feasible, high-impact interventions for maternal and child survival during India’s 2010–2013 measles supplementary immunization activity. Methods We developed Lives Saved Tool models for 12 Indian states participating in the supplementary immunization, based on state- and sex-specific data on mortality from India’s Million Deaths Study and on health services coverage from Indian household surveys. Potential add-on interventions were identified through a literature review and expert consultations. We quantified the number of lives saved for a campaign offering measles vaccine alone versus a campaign offering measles vaccine with six add-on interventions (nutritional screening and complementary feeding for children, vitamin A and zinc supplementation for children, multiple micronutrient and calcium supplementation in pregnancy, and free distribution of insecticide-treated bednets). Findings The measles vaccination campaign saved an estimated 19 016 lives of children younger than 5 years. A hypothetical campaign including measles vaccine with add-on interventions was projected to save around 73 900 lives (range: 70 200–79 300), preventing 73 700 child deaths (range: 70 000–79 000) and 300 maternal deaths (range: 200–400). The most effective interventions in the whole package were insecticide-treated bednets, measles vaccine and preventive zinc supplementation. Girls accounted for 66% of expected lives saved (12 712/19 346) for the measles vaccine campaign, and 62% of lives saved (45 721/74 367) for the hypothetical campaign including add-on interventions. Conclusion In India, a measles vaccination campaign including feasible, high-impact interventions could substantially increase the number of lives saved and mitigate gender-related inequities in child mortality.


Introduction
Measles vaccination made an important contribution to the millennium development goal to reduce under-5 mortality (MDG4), 1 accounting for 23% of the estimated worldwide decline in all-cause child mortality from 1990 to 2008. 2,3 A cornerstone of the strategy was that all children be offered a second opportunity to receive a dose of measles-containing vaccine, either through routine immunization services or through mass vaccination campaigns (known as supplementary immunization activities). 4 Supplemental immunization targets all children, to reach those who have been missed by routine services and also those who may have failed to develop an appropriate immune response after vaccination. 4 The strategy has been widely implemented in sub-Saharan Africa over the last decade, with measurable success in reducing mortality. 5 India delayed implementing supplementary immunization, and this may have contributed to the slower decline in measles mortality as compared with sub-Saharan Africa. India's share of global measles mortality increased from 16% of 535 300 deaths (95% confidence interval, CI: 347 200-976 400) in 2000 to 47% of 139 300 deaths (95% CI: 71 200-447 800) in 2010. 6 In 2010, India introduced a second opportunity to receive measles-containing vaccine through routine immunization programmes in states with 80% or higher coverage of the first dose of measles-containing vaccine, and elsewhere through supplementary immunization activities. India's first supplementary mass measles vaccination campaign took place from 2010 to 2013 in 14 states 7 containing 59% of India's 113 million under-5 children (authors' calculations based on census data). 8 These 14 states have relatively weak health systems compared with the national average 9 and poorer progress towards MDG4. 10 The supplementary immunization activity reached 119 million children aged nine months to 10 years, achieving 91% coverage of the target population of 130 743 905. 11 India's first round of supplementary mass measles vaccination delivered only a measles-containing vaccine dose. Planning is underway for a larger measles-rubella vaccine introduction campaign targeting children aged 1-15 years. 12 Campaign-style delivery has two key advantages over routine services; it can achieve high coverage even in areas where the reach of routine services is weak 2 and it reduces access barriers. On the other hand, a weakness of campaign delivery is that it represents a one-time or cyclic event. Some countries have made strategic use of mass vaccination campaigns to offer additional health interventions such as vitamin A supplements, insecticide-treated bednets and deworming medicines. 2 Therefore, vaccination campaigns could serve as an important platform to extend the reach of health services to underserved groups and improve maternal and child survival.
To date, India has largely not included add-on interventions with its mass vaccination campaigns and Indian health planners have expressed concerns over the potential challenges of implementing these, while agreeing that add-ons could be beneficial in principle. 9 To inform the design of future supplementary immunization activities in India and elsewhere we Objective To quantify the impact on mortality of offering a hypothetical set of technically feasible, high-impact interventions for maternal and child survival during India's 2010-2013 measles supplementary immunization activity. Methods We developed Lives Saved Tool models for 12 Indian states participating in the supplementary immunization, based on state-and sex-specific data on mortality from India's Million Deaths Study and on health services coverage from Indian household surveys. Potential add-on interventions were identified through a literature review and expert consultations. We quantified the number of lives saved for a campaign offering measles vaccine alone versus a campaign offering measles vaccine with six add-on interventions (nutritional screening and complementary feeding for children, vitamin A and zinc supplementation for children, multiple micronutrient and calcium supplementation in pregnancy, and free distribution of insecticide-treated bednets). Findings The measles vaccination campaign saved an estimated 19 016 lives of children younger than 5 years. A hypothetical campaign including measles vaccine with add-on interventions was projected to save around 73 900 lives (range: 70 200-79 300), preventing 73 700 child deaths (range: 70 000-79 000) and 300 maternal deaths (range: 200-400). The most effective interventions in the whole package were insecticide-treated bednets, measles vaccine and preventive zinc supplementation. Girls accounted for 66% of expected lives saved (12 712/19 346) for the measles vaccine campaign, and 62% of lives saved (45 721/74 367) for the hypothetical campaign including addon interventions. Conclusion In India, a measles vaccination campaign including feasible, high-impact interventions could substantially increase the number of lives saved and mitigate gender-related inequities in child mortality.
aimed to project the impact on mortality of a hypothetical set of technically feasible, high-impact interventions for maternal and child survival, delivered during India's 2010-2013 mass measles vaccination campaign.

Methods
For states participating in the supplementary immunization activity, we conducted a mathematical modelling study to quantify: (i) the number of lives saved by a supplementary immunization activity delivering measles-containing vaccine alone, and (ii) the number of lives that could be saved by a supplementary immunization activity package delivering measles-containing vaccine plus a set of six hypothetical add-on interventions. The analysis baseline reflected existing coverage levels for all interventions offered through routine services. Within each state we also assessed the impact of the interventions on mortality by child's sex. Ethics approval was not required for this study as it used only secondary data with no personal identifiable information. A technical appendix containing full details of the methods is available from the corresponding author.

Selection of interventions
We selected add-on interventions for modelling through a literature review and expert consultation. First, we used two systematic reviews to identify maternal and child health interventions that had been linked to routine immunization or vaccination campaigns (but not specific to measles) in a lowor middle-income country, identified from two systematic reviews. 13,14 Then we updated the literature search from these reviews to 15 May 2015, and consulted supplementary sources. 2,[15][16][17] Further suggestions were contributed by programme experts, including administrators and managers involved in India's 2010-2013 measles supplementary immunization activities. 2,9 From these inputs we prepared a comprehensive list of potential add-on interventions. Next, we condensed the list based on a review of the evidence of the feasibility of interventions, matched to target population and effectiveness, in the context of a supplementary immunization activity. 3 Finally, three experts engaged with India's immunization programme at central and state levels prioritized the interventions to create a shortlist of interventions for analysis based on criteria of programmatic and technical feasibility and policy relevance (Table 1). A total of six interventions -generally offered in India through the routine health system -were selected: (i) nutritional screening of children linked to services for complementary feeding; (ii) vitamin A supplementation for children; (iii) preventive zinc supplementation for children; (iv) free distribution of insecticide-treated bednets; (v) multiple micronutrient supplementation for pregnant women (iron, folic acid, vitamin A); and (vi) calcium supplementation for pregnant women.

Decision modelling
We modelled the impact of the interventions on maternal and child mortality over the period 2009-2013 using the freely available Lives Saved Tool (LiST), version 4.7 (Johns Hopkins Bloomberg School of Public Health, Baltimore, United States of America). LiST is a mathematical model that synthesizes evidence on the causes of maternal and child mortality and the effectiveness of interventions to combat them. 18 The structure of the model has been described elsewhere. 19 LiST can be used to project the impact that multiple interventions may have on survival. LiST was chosen because its target populations are similar to those of India's measles supplementary immunization activities. In addition, validation studies comparing actual measured mortality with modelled mortality showed that LiST provided accurate predictions in diverse geographical settings, including northern India. 20

Model parameters
Of the 14 states targeted for supplementary immunization, two were excluded from our analysis, as data on health services coverage (Nagaland, population 1 978 502) 8 and population structure (Arunachal Pradesh, population 1 383 727) 8 required for model parameterization were unavailable. LiST developers have made available parameterized models representing India and the state of Bihar in 2008 and we created LiST models for 11 additional supplementary immunization activity states by tailoring the Indian LiST module.
We used recent demographic projections for India to create age-and sex-structured populations for modelling. 10 Estimates for the effectiveness of add-on interventions were taken from the child health epidemiology reference group (CHERG) systematic reviews incorporated in LiST, with the exception of vitamin A supplementation for which we used a more recent metaanalysis incorporating findings from the Deworming and Enhanced Vitamin A Trial (DEVTA) in Uttar Pradesh, India. 21 We developed state-specific proportional mortality estimates by mapping cause-of-death data from India's Million Deaths Study (MDS) 22,23 to the LiST model categories. The MDS is a nationally representative longitudinal study of premature mortality monitoring 14 million people in India, which assigns cause of death by physician-reviewed verbal autopsy. For Manipur, Meghalaya and Tripura, state-specific mortality data were not available and for these states we used regional proportional mortality estimates. The MDS did not evaluate pertussis deaths as a separate category due to the difficulty of distinguishing pertussis from causes of death such as pneumonia when using verbal autopsy techniques. We imputed pertussis deaths using CHERG methods. 24,25 To characterize immunization coverage before the supplementary immunization activity, values for other parameters were derived from Indian household surveys. The principal data source was India's 2007-2008 district level household and facility survey; 26 data were collected just before the measles supplementary immunization. The technical appendix with illustrations of parameter values and data sources for a sample state are available from the corresponding author. Coverage data for the 2010-2013 measles supplementary immunization were provided by the Government of India.

Integrated vaccination campaign package
We modelled the supplementary immunization activity as occurring in all states in a single year (2010). The campaign would confer a one-time increase in measles vaccination coverage. Some of the hypothetical interventions, such as delivering vitamin A supplements and carrying out nutritional screening, could be completed at the time of vaccination. For these interventions, increases in coverage were modelled as a function of measles-containing vaccine coverage we modelled the causal chain between being offered in the supplementary immunization activity and increased intervention coverage as depending on measles-containing vaccine achievement and compliance. For these interventions, we used an average compliance scenario of 70% and considered two additional scenarios bounding reasonable ranges of low (50%) and high (90%) compliance. We assumed that 73% of freely distributed long-lasting insecticide-treated bednets would be used. 27 The analytic assumptions are outlined in

Sensitivity analyses
Additional analyses explored the effect of using different sources of data for proportional mortality (i.e. comparing proportional mortality data for India from CHERG and state-specific proportional mortality data from the MDS). We also quantified the impact of parameter uncertainty related to the effectiveness of vitamin A supplementation on diarrhoea mortality for children aged 6-59 months. To do this we contrasted the DEVTA meta-analysis midpoint estimate of 11% 21 with the 47% mortality benefit incorporated in LiST. 18 Finally, we developed the Dynamic Measles Immunization Calculation Engine, a transmission dynamic measles model 31 that enabled us to consider factors not captured in LiST, such as age-specific SIA: supplementary immunization activity; DTP: diphtheria-tetanus-pertussis. a Interventions selected for modelling in this analysis. b Interventions recommended as appropriate but lower priority for this analysis due to low impact on mortality or lack of evidence. c With the exception of combination vaccines, offering additional vaccines was viewed as challenging due to issues of logistics, safety and human resources. d Scope for pregnancy interventions depends on the proportion of children brought by mothers to receive measles vaccine and the proportion of pregnant women. e Interventions judged to be potentially valuable but lower priority for this analysis due to the need for empirical investigation. Note: Further details of the appraisal are available from the corresponding author.

Research
Maternal and child health interventions in India Mira Johri et al.
vaccine efficacy for measles first and second doses, and herd immunity, to model the impact on mortality of the supplementary immunization activity delivering only measles-containing vaccine.

Equity impact
Equity analyses were done on a state-bystate basis and assumed that increases in supplementary immunization coverage reached both sexes equally. To quantify the impact of the mass vaccination campaign on sex differences in child mortality, we used sex-specific proportional mortality data from the MDS 22 and sex-specific coverage data from the third round of the district level household and facility survey 26 and other sources (further details are available from the corresponding author). We assessed the impact on gender equity of the campaign delivering measles vaccine only and the hypothetical campaign delivering measles vaccine and add-on interventions by comparing the proportion of hypothetical lives saved by the campaign for girls versus boys, and the under-5 mortality rate per 1000 live births for girls versus boys in the years before (2009) and after (2010) the measles campaign.

Overall results
India's decision to introduce a second opportunity for measles vaccination via mass vaccination campaign saved the lives of an estimated 19 016 under-5 children in the 12 states included in our analysis, of whom 11 121 (58%) were in the state of Uttar Pradesh (Table 3). Table 4 shows the projected lives saved in these states by a hypothetical supplementary immunization package that included measles vaccine and high-impact add-on interventions for children and pregnant women. This was based on a scenario of 70% compliance with interventions (when applicable) and on mortality data from India's MDS. Maternal lives saved were due to calcium supplementation in pregnancy; all other lives saved represented under-5 children. Summing over all states, including maternal and child health interventions in the measles supplementary immunization campaign would have saved around 73 900 (range: 70 200-79 300), preventing 73 700 child deaths (range: 70 000-79 900) and 300 maternal deaths (range: 200-400). The hypothetical campaign offering measles vaccine with addon interventions was therefore projected to increase the number of lives saved by a factor of 3.89 (range: 3.69 to 4.17) compared with offering measles vaccine alone. The benefits of the add-ons were also distributed among states more closely to the proportion of the population (Table 3 and Table 4; supplementary data are available from the corresponding author). For example, Uttar Pradesh, which had 30.7% of the under-5 target population for the supplementary immunization, gained 58.5% of lives saved from the measles-only supplementary immunization (Table 3) and 36.0% of lives saved from the package of measles supplementary immunization with addons (Table 4).
We explored which add-on interventions in the whole package contributed most to the anticipated reductions in mortality (Table 4). Summing over all states for the average (70%) compliance scenario, the effectiveness of the interventions in descending order were: insecticide-treated bednets (24 933 lives saved), measles vaccine (18 235), preventive zinc supplementation (15 529), complementary feeding (4284), vitamin A supplementation (2093), multiple micronutrients supplementation during pregnancy (3761) and calcium supplementation during pregnancy (271). Some of the lives saved by the supplementary immunization with add-ons could not be attributed to specific interventions and are presented instead by syndrome, including diarrhoea (1838 lives saved), pneumonia

Sensitivity analyses
Using different assumptions about proportional mortality affected the findings shown in Table 4. Proportional mortality based on the MDS data attributed greater importance to malaria deaths and hence a higher projected number

Equity analyses
The campaign mitigated pre-existing inequalities in mortality between girls and boys. For under-5 children, the supplementary immunization delivering only measles-containing vaccine were projected to save the lives of 12 712 (66%) girls and 6 635 (34%) boys. The hypothetical campaign delivering measles vaccine and add-on interventions, based on the 70% compliance scenario, saved an expected 45 721 (62%) girls and 28 647 (39%) boys (Table 5; additional supplementary data are available from the corresponding author). For the eight states in which under-5 mortality per 1000 live births was initially greater in girls than in boys, the gender disparity was reduced by the supplementary immunization activity offering measlescontaining vaccine only and further reduced by the hypothetical campaign delivering measles vaccine and add-on interventions ( Table 6).

Discussion
Measles vaccination is important for reduction of child mortality, yet global coverage of the first dose of measlescontaining vaccine has been stagnant since 2009. 1 Mass vaccination campaigns are resource-intensive and planners must assess their value among a range of options for health improvement and resource expenditure. Our analysis demonstrated that India's introduction of a second opportunity for measles vaccination through large-scale campaigns from 2010 to 2013 made an important contribution to reducing mortality from measles. Our model-based analysis of 12 of the 14 participating states found that India's measles supplementary immunization activity likely saved the lives of approximately 19 000 under-5 children, corresponding to roughly 29% (range: 24% to 35%) of India's annual measles mortality. 6 We also found that a hypothetical supplementary immunization package delivering measles vaccine and a set of additional interventions of known effectiveness would increase the impact on mortality of the mass measles vaccination campaign more than threefold. Despite variation among states, the most important interventions in the package overall were insecticide-treated bednets, measles vaccine and preventive zinc supplementation. This reflects the high burden of infectious disease and undernutrition among Indian children, the impact of malaria in some areas and the relatively low coverage of these key interventions. [24][25][26] Child mortality in India differs markedly by sex, with higher mortality rates recorded for girls. 10,22 Caregiver bias associated with preference for a male child likely contributes to the mortality differentials due to lower use of regular health services for girls. Vaccination campaigns, however, show a more equal pattern of use. 30 We found

Research
Maternal and child health interventions in India Mira Johri et al.
that, due to the high coverage achieved in states with weak health systems, supplementary immunization helped to mitigate gender-related inequities in child mortality. We also studied differences in mortality for subgroups defined by household wealth status (quintiles) and area of residence (rural/urban). However, the results were uninformative due to the absence of state-and stratum-specific proportional mortality data (available from the corresponding author).
The components of this integrated health package for a supplemental mass vaccination campaign were designed based on a systematic appraisal of the evidence and a quantitative projection of likely impact. Add-on interventions were systematically selected using the published scientific literature and expert guidance. 9 A key advantage of this approach is that it presents the evidence and assumptions in a transparent framework that permits alternatives to be explored.
We placed particular emphasis on defining options for analysis that were evidence-based, feasible and relevant to the Indian context. Cause-of-death data from India's MDS 22 enabled us to show state-and sex-specific mortality patterns, while data from recent household surveys 26 facilitated an accurate portrayal of health services coverage. LiST is a validated policy model that enables competing mortality risks to be considered. 19 In addition, we validated the LiST projections of the impact of measles-containing vaccine using a transmission dynamic model that takes into account both herd immunity and age-specific vaccine efficacy. As anticipated, the dynamic model results were consistent with LiST results but showed a somewhat higher impact on mortality for the supplementary immunization with measles vaccine.
We highlight five limitations of the analysis. First, the LiST model focuses only on mortality. Many of the add-on interventions studied also reduce morbidity, and some potentially important interventions, such as antihelminthic drugs, were not considered as their direct impact is exclusively on morbidity. 2 Second, limited availability of data forced us to exclude two states of less than 2 million inhabitants each. 3,8 Third, uncertainty concerning parameter values for vitamin A effectiveness, proportional mortality due to malaria, and compliance were found to influence mortality projections. 4 Fourth, our mortality projections represent a specific point in time, whereas child survival and health services coverage are changing rapidly in India. 5 Finally, due to constraints of logistics, it may not be possible in practice to offer as many add-on interventions as we have modelled for this analysis. Our primary purpose was to demonstrate the potential benefits of bundling proven interventions with a vaccination campaign. We also showed the utility of an evidence-based approach for planning add-ons for supplementary immuniza-  a Add-on interventions were nutritional screening linked to complementary feeding; vitamin A supplementation for children; preventive zinc supplementation for children; free distribution of insecticide-treated bednets; multiple micronutrient supplementation in pregnant women (iron, folic acid, vitamin A); and calcium supplementation in pregnant women. b State totals represent the sum of sex-specific models and thus differ from Table 3. Note: Including 12 states participating in India's 2010-2013 measles supplementary immunization activity. Analyses use proportional mortality from the Million Deaths study 22 and vitamin A effectiveness from the Deworming and Enhanced Vitamin A Trial. 21 Lives saved were calculated for the period 2010-2013. Data are based on the 70% compliance scenario (see Table 2) and apply to multiple micronutrients and calcium during pregnancy, and preventive zinc for children. Table 6.  a Add-on interventions were nutritional screening linked to complementary feeding; vitamin A supplementation for children; preventive zinc supplementation for children; free distribution of insecticide-treated bednets; multiple micronutrient supplementation in pregnant women (iron, folic acid, vitamin A); and calcium supplementation in pregnant women. b Results were calculated without applying background trends in acquired immune deficiency syndrome mortality. Note: Including 12 states participating in India's 2010-2013 measles supplementary immunization activity. Analyses use proportional mortality from the Million Deaths Study 22 and vitamin A effectiveness from the Deworming and Enhanced Vitamin A Trial. 21 Lives saved were calculated for the period 2010-2013.

Research
Maternal and child health interventions in India Mira Johri et al.
tion activities. The integrated package studied in this analysis should be seen as aspirational. In practice, the impact on mortality will depend on the actual interventions offered and may be less than illustrated here.
Mass measles vaccination campaigns in many countries have offered additional interventions, 2 but the choice of which interventions to include has generally been made in an ad hoc way rather than through a systematic analysis such as the one illustrated here. Although the interventions we examined were all deemed by Indian health planners to be technically feasible to incorporate into mass vaccination campaigns, implementation research is needed to assess the feasibility and impact on health systems of offering these interventions. We also need to assess the cost-effectiveness of supplementary immunization activities that include measles vaccine and add-on maternal and child health interventions in India. As vaccination campaigns must be repeated periodically, this research has the potential to revitalize political support for accelerated measles control strategies, as well as for other vaccines delivered through mass campaigns, such as rubella and polio.

Adición de intervenciones a las vacunaciones antisarampionosas en masa en la India
Objetivo Cuantificar el impacto en la mortalidad del hecho de ofrecer un conjunto hipotético de intervenciones técnicamente viables y de alto impacto para la supervivencia de madres e hijos durante la actividad de inmunización suplementaria antisarampionosa en la India entre 2010 y 2013. Métodos Se desarrollaron los modelos de la herramienta "Live Saved Tool" para 12 estados indios participantes en la inmunización suplementaria, en base a información según el estado y el sexo sobre la mortalidad recopilada en el estudio "Million Deaths Study" de la India y sobre la cobertura de servicios sanitarios de las encuestas domésticas realizadas en la India. Se identificaron posibles intervenciones adicionales a través de un análisis de documentos y consultas a expertos. Se cuantificó el número de vidas salvadas durante una campaña que ofrecía la vacuna antisarampionosa frente a una campaña que ofrecía la