• 1.

    Black RE et al.Maternal and Child Nutrition Study Group 2013. Maternal and child undernutrition and overweight in low-income and middle-income countries. Lancet 382: 427451.

    • Search Google Scholar
    • Export Citation
  • 2.

    Prendergast AJ, Humphrey JH, 2014. The stunting syndrome in developing countries. Paediatr Int Child Health 34: 250265.

  • 3.

    Bhutta ZA, Das JK, Rizvi A, Gaff MF, Walker N, Horton S, Webb P, Lartey A, Black RE, 2013. Evidence-based interventions for improvement of maternal and child nutrition: what can be done and at what cost?Lancet 382: 452477.

    • Search Google Scholar
    • Export Citation
  • 4.

    Chan M, 1997. The global burden of intestinal nematode infections—fifty years on. Parasitol Today 13: 438443.

  • 5.

    de Silva NR, 2003. Impact of mass chemotherapy on the morbidity due to soil-transmitted nematodes. Acta Trop 86: 197214.

  • 6.

    Taylor-Robinson DC, Maayan N, Soares-Weiser K, Donegan S, Garner P, 2012. Deworming drugs for soil-transmitted intestinal worms in children: effects on nutritional indicators, haemoglobin and school performance. Cochrane Database Syst Rev 11: CD000371.

    • Search Google Scholar
    • Export Citation
  • 7.

    Higgins JPT, Deeks, JJ, Altman DG, eds, 2011. Chapter 16: Special topics in statistics. Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions, Version 5.1.0. The Cochrane Collaboration, 2011. Available at: www.handbook.cochrane.org. Accessed February 10, 2017.

  • 8.

    Stampfer MJ, Buring JE, Willett W, Rosner B, Eberlein K, Hennekens CH, 1985. The 2 × 2 factorial design: its application to a randomized trial of aspirin and carotene in U.S. physicians. Stat Med 4: 111116.

    • Search Google Scholar
    • Export Citation
  • 9.

    Mills EJ, Thorlund K, Ioannidis JP, 2012. Calculating additive treatment effects from multiple randomized trials provides useful estimates of combination therapies. J Clin Epidemiol 65: 12821288.

    • Search Google Scholar
    • Export Citation
  • 10.

    Habicht JP, Martorell R, Rivera JA, 1995. Nutritional impact of supplementation in the INCAP longitudinal study: analytic strategies and inferences. J Nutr 125: 1042S1050S.

    • Search Google Scholar
    • Export Citation
  • 11.

    Rivera JA, Hotz C, Gonzalez-Cossio T, Neufeld L, Garcia-Guerra A, 2003. The effect of micronutrient deficiencies on child growth: a review of results from community-based supplementation trials. J Nutr 133: 4010S4020S.

    • Search Google Scholar
    • Export Citation
  • 12.

    Mora JO, Herrera MG, Suescun J, de Navarro L, Wagner M, 1981. The effects of nutritional supplementation on physical growth of children at risk of malnutrition. Am J Clin Nutr 34: 18851892.

    • Search Google Scholar
    • Export Citation
  • 13.

    Mills EJ, Gardner D, Thorlund K, Briel M, Bryan S, Hutton B, Guyatt GH, 2014. A users’ guide to understanding therapeutic substitutions. J Clin Epidemiol 67: 305313.

    • Search Google Scholar
    • Export Citation
  • 14.

    Ayoya MA, Spiekermann-Brouwer GM, Traore AK, Stoltzfus RJ, Habicht JP, Garza C, 2009. Multiple micronutrients including iron are not more effective than iron alone for improving hemoglobin and iron status of Malian school children. J Nutr 139: 19721979.

    • Search Google Scholar
    • Export Citation
  • 15.

    Sandstrom B, 2001. Micronutrient interactions: effects on absorption and bioavailability. Br J Nutr 85 (Suppl 2): S181S185.

  • 16.

    Lonnerdal B, 2004. Interactions between micronutrients: synergies and antagonisms. Pettifor JM, Zlotkin S, eds. Neslé Nutrition Workshop Series Pediatric Program Volume 54: Micronutrient Deficiencies During the Weaning Period and the First Years of Life. Basel, Switzerland: Nestle Nutrition.

  • 17.

    Mayo-Wilson E, Junior J, Imdad A, Dean S, Xhs C, Es C, Jaswal A, Bhutta Z, 2014. Zinc supplementation for preventing mortality, morbidity, and growth failure in children aged 6 months to 12 years of age. Cochrane Database Syst Rev 5: CD009384.

    • Search Google Scholar
    • Export Citation
  • 18.

    Rahman MM, Tofail F, Wahed MA, Fuchs GJ, Baqui AH, Alvarez JO, 2002. Short-term supplementation with zinc and vitamin A has no significant effect on the growth of undernourished Bangladeshi children. Am J Clin Nutr 75: 8791.

    • Search Google Scholar
    • Export Citation
  • 19.

    Berger J, Ninh NX, Khan NC, Nhien NV, Lien DK, Trung NQ, Khoi HH, 2006. Efficacy of combined iron and zinc supplementation on micronutrient status and growth in Vietnamese infants. Eur J Clin Nutr 60: 443454.

    • Search Google Scholar
    • Export Citation
  • 20.

    Dijkhuizen MA, Winichagoon P, Wieringa FT, Wasantwisut E, Utomo B, Ninh NX, Hidayat A, Berger J, 2008. Zinc supplementation improved length growth only in anemic infants in a multi-country trial of iron and zinc supplementation in south-east Asia. J Nutr 138: 19691975.

    • Search Google Scholar
    • Export Citation
  • 21.

    Fahmida U, Rumawas JS, Utomo B, Patmonodewo S, Schultink W, 2007. Zinc-iron, but not zinc-alone supplementation, increased linear growth of stunted infants with low haemoglobin. Asia Pac J Clin Nutr 16: 301309.

    • Search Google Scholar
    • Export Citation
  • 22.

    Yang YX, Han JH, Shao XP, He M, Bian LH, Wang Z, Wang GD, Men JH, 2002. Effect of micronutrient supplementation on the growth of preschool children in China. Biomed Environ Sci 15: 196202.

    • Search Google Scholar
    • Export Citation
  • 23.

    Jumbe NL, Murray JC, Kern S, 2016. Data sharing and inductive learning—toward healthy birth, growth, and development. N Engl J Med 374: 24152417.

    • Search Google Scholar
    • Export Citation
  • 24.

    HBGDki, 2017. Data Science Resources Tools & Models. Available at: http://hbgdki.org/tools-models. Accessed November 15, 2017.

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 

 

Challenges in Assessing Combined Interventions to Promote Linear Growth

View More View Less
  • 1 Department of Epidemiology, University of Washington, Seattle, Washington;
  • | 2 Department of Global Health, University of Washington, Seattle, Washington;
  • | 3 Childhood Acute Illness & Nutrition Network (CHAIN), Nairobi, Kenya;
  • | 4 Pharmactuarials LLC, Mountain View, California;
  • | 5 Department of Pediatrics, University of Washington, Seattle, Washington;
  • | 6 Department of Medicine, University of Washington, Seattle, Washington

Despite the recognition of stunting as a public health priority, nutritional and nonnutritional interventions to reduce or prevent linear growth failure have demonstrated minimal impact. Investigators and policymakers face several challenges that limit their ability to assess the potential benefits of combining available interventions into a linear growth promotion package. We use two common but very different interventions, deworming and multiple micronutrient supplements, to illustrate barriers to recommending an optimal linear growth promotion package based on the currently available literature. These challenges suggest that combining individual- and population-based as well as model-based approaches would complement existing research using systematic review, meta-analysis, and factorial randomized trials, and help integrate existing fields of research to inform the development of optimal linear growth promotion packages for children living in resource-limited settings.

INTRODUCTION

More than 165 million children worldwide are stunted (height-for-age z-score [HAZ] < −2), a marker of chronic malnutrition.1 Stunting is associated with substantial morbidity and mortality, including a 2-fold increase in the risk of death before age 5.1,2 Although stunting is a public health priority, current interventions have demonstrated minimal effect in reversing or preventing stunting. Estimates suggest that if available growth-promoting interventions reached 90% of their target population, stunting incidence would only be reduced by one-fifth.3 The combination of proximal etiologic causes of stunting, including lack of adequate nutrition, hormonal dysregulation, repeated infections, environmental enteric dysfunction, and chronic systemic inflammation, occurs within a complex network of more distal factors that underlie stunting.2 As a result of the multifactorial etiologies of stunting, successful prevention or treatment may require combined packages of interventions targeting multiple pathophysiologic pathways leading to stunting. The failure to prevent or reverse stunting may result from the complex interactions between prenatal and early childhood health insults that result in linear growth failure.

There is increasing interest in determining the effects of combination packages of interventions that may or may not have been evaluated together in clinical trials. Although many randomized trials of potentially growth-promoting interventions have been conducted, interventions are usually not tested in different combinations to establish optimal packages for linear growth promotion. Investigators using currently available literature to evaluate the potential benefit of combining interventions into useful packages to prevent or treat stunting face several methodological challenges that limit their ability to assess the potential combination benefits of available interventions. An understanding of these challenges and the limitations of traditional methods should encourage researchers and policymakers to consider the use of nontraditional methods designed to integrate data and model testing of different intervention combinations.

This article highlights two commonly delivered but very different interventions, deworming and multiple micronutrient supplementation, that are administered concurrently across a range of geographic areas. We use these examples to illustrate challenges in constructing a linear growth promotion package using currently available evidence.

DEWORMING: INTERPRETING EFFICACY VERSUS EFFECTIVENESS TRIALS

Soil-transmitted helminths are among the world’s most common infections, affecting more than 25% of the world’s population.4 Children are most affected by severe infections that can have detrimental long-term effects, including stunting.5 Anthelmintic therapy (deworming) is administered to more than 100 million children annually, often in mass drug administration (MDA) programs, and there is some evidence to suggest a potential linear growth benefit from deworming.6 However, results from a recent Cochrane review of randomized trials assessing the growth benefits of deworming interventions are mixed. Although some individual trials demonstrate beneficial effects of deworming on childhood growth, the results of pooled analyses demonstrate no statistically significant effects (mean difference −0.02 cm, 95% confidence interval [CI]: −0.17, 0.12).6 Interpreting and applying these results requires a careful understanding of the distinction between efficacy and effectiveness in the context of clinical trial design.

Efficacy trials assess how well an intervention can work under ideal circumstances. To determine the potential efficacy of deworming interventions, initial trials need to be conducted among children who are known to have helminth infection at baseline, to assess the effect of treatment in an infected population. Among the 42 trials in children included in the Cochrane review, few trials (2, N = 136) were conducted under these conditions and measured linear growth as an outcome (mean difference: 0.10 cm, 95% CI: −0.15, 0.45). Best available estimates of the efficacy of deworming should rely on studies that first screen for helminth infection and treat only those who are infected.6

Effectiveness trials, also known as pragmatic trials, assess an intervention’s effect in real world conditions and can be challenging to interpret. Effectiveness trials of deworming use MDA to treat populations independent of individual helminth infection status because it is less expensive and more readily delivered than implementing a screen and treat program. Thus, the measured effectiveness of MDA programs, the most common deworming intervention, to promote linear growth, depends on the prevalence of helminth infection in the population as any improvement in growth would only be expected to occur in children infected with helminths. Conversely, uninfected individuals are not likely to benefit from deworming and will attenuate the population effect size observed in that study. As a result, studies testing MDA deworming interventions conducted in high prevalence settings demonstrate greater growth benefit than studies in lower prevalence populations (mean difference: 0.25 cm, 95% CI: −0.10, 0.60 versus −0.26 cm, 95% CI: −0.74, 0.21).6 Although the true effect of deworming any single child infected with helminths may be fixed, the population level effects will depend on the prevalence of infection. As a result, estimates of the population benefit expected if anthelmintics administered by MDA are included in a package targeting stunting must take into account 1) the benefit expected among infected children and how that benefit is likely to translate to population level growth outcomes and 2) interaction effects of package contents.

MULTIPLE MICRONUTRIENTS: UNDERSTANDING AND EVALUATING THE COMBINED EFFECTS OF INTERVENTIONS

Results from trials of childhood multiple micronutrient interventions provide insight into the potential multiplicative or additive benefits provided by combining interventions into a package. However, although the methodological principles of evaluating the combined effects of interventions are well established, determining how combined interventions interact is complex.7,8 The chief aim of trialing multiple interventions is assessing whether the combination offers an added benefit over the single intervention, often referred to as additive effects. Additive effects assume that if treatment A has an effect EA over a control and treatment B has an effect EB over that same control, then the combination of A and B would have an effect EA + EB.9 Additive effects assume that the interventions work via independent pathways. For example, it would be reasonable to assume that a handwashing intervention and a nutritional intervention do not use the same pathway and therefore, should not interact. However, many interventions have complex or uncertain mechanisms of action, and so it is advisable to test for interactions between combined interventions using factorial randomized trials. Guidance for the assumptions of additive effects have been previously reported.9 In rare circumstances, interactions can lead to a greater-than-expected treatment effect (i.e., EAB > EA + EB), referred to as synergy. When an interaction leads to a treatment effect less than the sum of the combined treatments (i.e., EAB < EA + EB) it is termed as antagonism. A common cause of antagonistic interactions is class effect. This occurs when interventions using similar pathways have comparable individual effect sizes but negligible additive effects, because the therapeutic utility of the shared pathway is already fully exploited by a single therapeutic.

Intra-package interactions are extremely relevant in the context of micronutrient interventions which often do not target a single nutrient, but rather try to tackle multiple nutritional deficiencies.1012 In theory, by addressing several stunting-associated deficiencies, childhood micronutrient supplementation should have additive benefits. However, demonstrating additive or synergistic benefit is more challenging than proving a single intervention efficacy.13 There may be a maximum threshold of benefit provided by micronutrient supplementation that limits the total linear growth impact achievable. In addition, individual micronutrients may compete for intestinal or hematological transport mechanisms, potentially prohibiting the full effect of each individual intervention.1416

For example, a published meta-analysis of zinc supplementation in children suggests a modest but statistically significant benefit on linear growth, particularly among children who are zinc deficient.17 However, results of individual factorial design studies testing zinc in combination with other micronutrients such as vitamin A or iron are inconsistent and do not demonstrate a clear additive benefit.1822 It is possible that the micronutrients exert a class effect or that they compete to exert an effect through a single biological pathway. The difficulty in demonstrating the biological benefit of combining multiple individual micronutrients into a single treatment reinforces the need for careful consideration of potential interactions between constituent interventions of a package when designing any growth promotion package.

To directly extract efficacy and effectiveness knowledge and quantitatively delineate the complex interaction between proximal and distal determinants on growth outcomes, a large knowledgebase combining novel modeling techniques beyond systematic review, meta-analysis, and factorial randomized trials may be helpful to increase learning and knowledge generation.

DISCUSSION

Anthelmintics and multiple micronutrient interventions serve as examples to highlight the complexity of assessing interventions targeting linear growth and of determining optimal combinations of interventions to achieve maximum potential benefit. Many interventions have not been evaluated for growth benefit in well-designed efficacy trials, and many trials of combined interventions were not designed to evaluate additive effects or antagonism. In addition, the diversity of interventions suggested for inclusion in a possible package targeting linear growth promotion may require delivery through a combination of delivery platforms. Therefore, recommending an optimal linear growth promotion package based on the currently available published literature with any certainty of its beneficial effect is challenging. This, in turn, limits the ability of policymakers to set guidelines for packages of interventions. Single intervention-based recommendations may be a missed opportunity to optimize childhood health.

It may be possible to use novel modeling techniques to integrate existing datasets and find commonalities that are not apparent through systematic review and meta-analysis. However, although using meta-analysis to pool or compare results from the diverse fields of research contributing to the stunting discourse may not always be methodologically appropriate, it may be possible to use nontraditional methods specifically designed to test the integration of different interventions. For example, nonlinear effect models, Markov models, piecewise linear models, principal components models, and machine-learning decision models can be used to evaluate longitudinal growth outcomes in pooled data and explore alternative treatment strategies via clinical trial simulation.

The Healthy Birth, Growth, and Development knowledge integration initiative at the Bill & Melinda Gates foundation is using a variety of analytic methods to model longitudinal growth outcomes and test the effect of interventions using clinical trial data.23 Select examples of models developed through this initiative are summarized in Table 1.24 Although a careful understanding of efficacy versus effectiveness and potential interactions in the current literature is still required, a combination of these methods applied to pooled datasets could offer new insight into the interactions between, and importance of, growth-promoting interventions. Ultimately, these innovative techniques could be used to inform the design of growth-package interventions before clinical trial testing.

Table 1

Examples of models developed for longitudinal growth outcomes and clinical trial simulation through HBGDki initiative24

Model nameDescription
Full random effects modelParametric nonlinear model to describe standardized growth (height-for-age z-score [HAZ])
Joint distribution of length, weight, and head circumferenceJoint parametric nonlinear using nonlinear deceleration structural model
Linear models ordered categorical model for HAZOrdered categorical model for HAZ with category probabilities depending on age and other predictors
Multistate Markov model to describe longitudinal changes in HAZMultistate model allowing transitions between HAZ categories; modeling-ordered categorical outcomes
Piecewise linear model to describe longitudinal HAZ measuresPiecewise linear growth over specified age intervals. Child-specific birth size and slopes are usually included in the model
Nonlinear mixed effects (NLME) modelParametric models for pre- and postnatal growth
Bayesian NLME modelBayesian parametric models for pre- and postnatal growth
Functional principal components model to describe longitudinal measuresSemiparametric model to describe growth
Superimposition by translation and rotation modelNLME model for weight and length/height
Machine-learning modelsEnsemble of 1,000 gradient-boosted decision trees

HBGDki = Health Birth, Growth, and Development knowledge integration.

It is clear that high-quality data from interventional trials are needed to inform the development of optimal intervention packages to improve growth in children living in resource-limited settings. However, using systematic review and meta-analysis of the current evidence base to design intervention packages is likely to expend valuable resources testing packages of interventions that are not optimized. New approaches to designing linear growth promotion packages are required.

Acknowledgments:

The co-authors thank Edward Mills and David Price for providing edits and feedback on the manuscript and the Strategic Analysis, Research, and Training Center (START) at the University of Washington for funding the original research.

REFERENCES

  • 1.

    Black RE et al.Maternal and Child Nutrition Study Group 2013. Maternal and child undernutrition and overweight in low-income and middle-income countries. Lancet 382: 427451.

    • Search Google Scholar
    • Export Citation
  • 2.

    Prendergast AJ, Humphrey JH, 2014. The stunting syndrome in developing countries. Paediatr Int Child Health 34: 250265.

  • 3.

    Bhutta ZA, Das JK, Rizvi A, Gaff MF, Walker N, Horton S, Webb P, Lartey A, Black RE, 2013. Evidence-based interventions for improvement of maternal and child nutrition: what can be done and at what cost?Lancet 382: 452477.

    • Search Google Scholar
    • Export Citation
  • 4.

    Chan M, 1997. The global burden of intestinal nematode infections—fifty years on. Parasitol Today 13: 438443.

  • 5.

    de Silva NR, 2003. Impact of mass chemotherapy on the morbidity due to soil-transmitted nematodes. Acta Trop 86: 197214.

  • 6.

    Taylor-Robinson DC, Maayan N, Soares-Weiser K, Donegan S, Garner P, 2012. Deworming drugs for soil-transmitted intestinal worms in children: effects on nutritional indicators, haemoglobin and school performance. Cochrane Database Syst Rev 11: CD000371.

    • Search Google Scholar
    • Export Citation
  • 7.

    Higgins JPT, Deeks, JJ, Altman DG, eds, 2011. Chapter 16: Special topics in statistics. Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions, Version 5.1.0. The Cochrane Collaboration, 2011. Available at: www.handbook.cochrane.org. Accessed February 10, 2017.

  • 8.

    Stampfer MJ, Buring JE, Willett W, Rosner B, Eberlein K, Hennekens CH, 1985. The 2 × 2 factorial design: its application to a randomized trial of aspirin and carotene in U.S. physicians. Stat Med 4: 111116.

    • Search Google Scholar
    • Export Citation
  • 9.

    Mills EJ, Thorlund K, Ioannidis JP, 2012. Calculating additive treatment effects from multiple randomized trials provides useful estimates of combination therapies. J Clin Epidemiol 65: 12821288.

    • Search Google Scholar
    • Export Citation
  • 10.

    Habicht JP, Martorell R, Rivera JA, 1995. Nutritional impact of supplementation in the INCAP longitudinal study: analytic strategies and inferences. J Nutr 125: 1042S1050S.

    • Search Google Scholar
    • Export Citation
  • 11.

    Rivera JA, Hotz C, Gonzalez-Cossio T, Neufeld L, Garcia-Guerra A, 2003. The effect of micronutrient deficiencies on child growth: a review of results from community-based supplementation trials. J Nutr 133: 4010S4020S.

    • Search Google Scholar
    • Export Citation
  • 12.

    Mora JO, Herrera MG, Suescun J, de Navarro L, Wagner M, 1981. The effects of nutritional supplementation on physical growth of children at risk of malnutrition. Am J Clin Nutr 34: 18851892.

    • Search Google Scholar
    • Export Citation
  • 13.

    Mills EJ, Gardner D, Thorlund K, Briel M, Bryan S, Hutton B, Guyatt GH, 2014. A users’ guide to understanding therapeutic substitutions. J Clin Epidemiol 67: 305313.

    • Search Google Scholar
    • Export Citation
  • 14.

    Ayoya MA, Spiekermann-Brouwer GM, Traore AK, Stoltzfus RJ, Habicht JP, Garza C, 2009. Multiple micronutrients including iron are not more effective than iron alone for improving hemoglobin and iron status of Malian school children. J Nutr 139: 19721979.

    • Search Google Scholar
    • Export Citation
  • 15.

    Sandstrom B, 2001. Micronutrient interactions: effects on absorption and bioavailability. Br J Nutr 85 (Suppl 2): S181S185.

  • 16.

    Lonnerdal B, 2004. Interactions between micronutrients: synergies and antagonisms. Pettifor JM, Zlotkin S, eds. Neslé Nutrition Workshop Series Pediatric Program Volume 54: Micronutrient Deficiencies During the Weaning Period and the First Years of Life. Basel, Switzerland: Nestle Nutrition.

  • 17.

    Mayo-Wilson E, Junior J, Imdad A, Dean S, Xhs C, Es C, Jaswal A, Bhutta Z, 2014. Zinc supplementation for preventing mortality, morbidity, and growth failure in children aged 6 months to 12 years of age. Cochrane Database Syst Rev 5: CD009384.

    • Search Google Scholar
    • Export Citation
  • 18.

    Rahman MM, Tofail F, Wahed MA, Fuchs GJ, Baqui AH, Alvarez JO, 2002. Short-term supplementation with zinc and vitamin A has no significant effect on the growth of undernourished Bangladeshi children. Am J Clin Nutr 75: 8791.

    • Search Google Scholar
    • Export Citation
  • 19.

    Berger J, Ninh NX, Khan NC, Nhien NV, Lien DK, Trung NQ, Khoi HH, 2006. Efficacy of combined iron and zinc supplementation on micronutrient status and growth in Vietnamese infants. Eur J Clin Nutr 60: 443454.

    • Search Google Scholar
    • Export Citation
  • 20.

    Dijkhuizen MA, Winichagoon P, Wieringa FT, Wasantwisut E, Utomo B, Ninh NX, Hidayat A, Berger J, 2008. Zinc supplementation improved length growth only in anemic infants in a multi-country trial of iron and zinc supplementation in south-east Asia. J Nutr 138: 19691975.

    • Search Google Scholar
    • Export Citation
  • 21.

    Fahmida U, Rumawas JS, Utomo B, Patmonodewo S, Schultink W, 2007. Zinc-iron, but not zinc-alone supplementation, increased linear growth of stunted infants with low haemoglobin. Asia Pac J Clin Nutr 16: 301309.

    • Search Google Scholar
    • Export Citation
  • 22.

    Yang YX, Han JH, Shao XP, He M, Bian LH, Wang Z, Wang GD, Men JH, 2002. Effect of micronutrient supplementation on the growth of preschool children in China. Biomed Environ Sci 15: 196202.

    • Search Google Scholar
    • Export Citation
  • 23.

    Jumbe NL, Murray JC, Kern S, 2016. Data sharing and inductive learning—toward healthy birth, growth, and development. N Engl J Med 374: 24152417.

    • Search Google Scholar
    • Export Citation
  • 24.

    HBGDki, 2017. Data Science Resources Tools & Models. Available at: http://hbgdki.org/tools-models. Accessed November 15, 2017.

Author Notes

Address correspondence to Emily L. Deichsel, Department of Epidemiology, University of Washington, Box 359931, 325 9th Avenue, Seattle, WA 98104. E-mail: deichsel@uw.edu

Financial support: START is a collaborative effort with, and is funded by, the Bill & Melinda Gates Foundation. The funder commissioned the study but did not have exclusive control over design, data collection and analysis, decision to pursue publication, conclusions, or preparation of the manuscript.

Authors’ addresses: Emily L. Deichsel and Jessica E. Long, Department of Epidemiology, University of Washington School of Public Health, Seattle, WA, E-mails: deichsel@uw.edu and jesslong@uw.edu. Kirkby D. Tickell, Department of Global Health, University of Washington School of Public Health, Seattle, WA, and Childhood Acute Illness & Nutrition Network (CHAIN), Nairobi, Kenya, E-mail: kirkbt@uw.edu. Nelson L. Jumbe, Pharmactuarials LLC, Mountain View, CA, E-mail: drshasha@gmail.com. Ali Rowhani-Rahbar, Department of Epidemiology, University of Washington School of Public Health, Seattle, WA, and Department of Pediatrics, University of Washington, Seattle, WA, E-mail: rowhani@uw.edu. Judd L. Walson, Department of Global Health, University of Washington School of Public Health, Seattle, WA, Department of Epidemiology, University of Washington School of Public Health, Seattle, WA, Department of Pediatrics, University of Washington, Seattle, WA, and Childhood Acute Illness & Nutrition Network (CHAIN), Nairobi, Kenya, E-mail: walson@uw.edu.

Save