Anemia, diet, and cognitive development: Impact of health information on diet quality and child nutrition in rural India

https://doi.org/10.1016/j.jebo.2021.06.043Get rights and content

Abstract

Lack of information about health risks may limit the adoption of improved nutrition and other new healthy behaviors. This paper studies the effect of a nutrition-information intervention on household dietary behavior, hemoglobin levels, and cognitive outcomes of children in rural India. Using experimental data and a regression discontinuity design that exploits the exogenous cutoff of hemoglobin level for anemia, we find statistically insignificant treatment effects on dietary improvements, child health, and cognitive outcomes of children. Our findings suggest that light-touch nutrition information alone, even when parents are informed about the health risk of their children, may not induce adoption of healthy behaviors. Our findings also imply that factors other than information might constrain households in making nutritional investments for their children.

Introduction

Despite its importance as a determinant of human capital formation, micronutrient deficiencies - particularly anemia - are still widespread in many low- and middle-income countries. Globally, anemia affects one-third of the world's population and about 50% of the anemic cases are due to iron deficiency, but this share is substantially higher in India (Anand et al., 2014; WHO, 2016). Iron deficiency caused due to the consumption of a nutritionally deficient diet is a critically important cause of anemia among Indian children (Thankachan et al., 2008).1 Iron-deficient anemia (IDA) affects 42% of children under age five globally and its prevalence varies according to geographic regions (WHO, 2016). Countries in South Asia and Sub-Saharan Africa have the highest prevalence of IDA across all age groups. Despite substantial economic growth and several anemia-reduction programs, IDA is highly prevalent among children in India. Recent data shows that about 60% of the under-five children are anemic in India, making it an urgent public health priority (International Institute for Population Sciences (IIPS) and ICF 2017).

Iron deficiency is associated with increased physical and cognitive impairment in childhood, resulting in lower schooling, irreversible productivity loss in adulthood, and substantial economic burden in developing countries (Strauss and Thomas, 1998; Alderman et al., 2001; Alderman et al., 2006; Glewwe et al., 2001). The estimated annual costs of IDA in early childhood amount to intangible costs of 8 million disability-adjusted life years and productivity loss equal to 1.3% of the gross domestic product in India (Plessow et al., 2015).

Several effective and inexpensive nutrition technologies, such as iron supplements, iron-fortified products, or bio-fortified seeds, are available to prevent iron deficiency. But, diffusion and adoption of these technologies have been limited in developing countries (Banerjee et al., 2018). For example, while consuming a more diverse and nutritious diet is an effective strategy to reduce iron deficiency, getting individuals to adopt a change is challenging, despite significant potential health gains (Ogden et al., 2007; Oster, 2018). That the average iron intake among Indian children is 2.4% of the recommended daily intake (Onyeneho et al., 2019) indicates that increase in diet-based iron intake could reduce the high incidence of IDA.

The high prevalence of anemia and limited adoption of healthy behavior to prevent iron deficiency is puzzling as it can easily be tackled with nutrition-specific interventions which are simple and inexpensive. One reason for the limited adoption of health technology and health-promoting behavior could be that households are unaware of the health risks caused due to iron deficiency and the potential solutions to reduce these risks. Limited access to knowledge about illness or preventive strategies may play an important role in explaining limited adoption of healthy behavior or improved dietary choices (Dupas, 2011a; Luo et al., 2012; Madajewicz et al., 2007). In a review paper, Dupas (2011a) suggests that the discrepancy between the actual and the optimal health behavior is possibly due to information deficit and findings from several studies suggest that households are responsive to information to health risks. Thus, reducing information deficit about health could improve individual's health.

This study examines the causal effect of a nutrition-information intervention that informed parents of the anemic status of their children and preventive strategies to reduce anemia risk in their children. The intervention informed parents about the anemia status of their children and the harmful effects of anemia on children's health and cognitive development. Parents of anemic children were further advised to feed diverse and iron-rich food to their children to prevent and cure anemia. The dietary advice was only given to the parents of anemic children and parents of non-anemic children were only informed of the anemic status of their children. We first investigate how this information intervention affects dietary behavior and consumption of iron-rich foods. We subsequently examine whether the information intervention had any effects on children's health status—measured by the children's hemoglobin (Hb) levels—as well as its impacts on the cognitive ability and educational outcomes of the children.

Our study setting is the state of Bihar, India. Bihar is one of the poorest states and has one of the lowest levels of human capital. Bihar is an important setting to explore this question as anemia prevalence among 6-59 months old children was 63% in 2015-16 (NFHS, 2017). Furthermore, approximately 48% of all children under age five in Bihar are stunted and 44% are underweight (NFHS, 2017).2 Our study uses data from a randomized controlled trial conducted by the authors on the impacts of iron-fortified salt on anemia. In this experiment, grade-two children were tested for anemia in 104 primary schools in Bihar. We use this data to roll out another experiment related to nutrition information and anemia of the sampled children. Parents of anemic children (hemoglobin level ≤ 10.9 grams per decilitre (g/dL)) were informed of their children's anemia status and were also advised to feed their children iron-rich food items, namely green leafy vegetables and lean meats if the household consumes meat. Non-anemic children (hemoglobin level ≥ 11.0 g/dL) were not exposed to any information intervention. We exploit the discontinuity in the information intervention based on a hemoglobin cutoff of 11.0 g/dL and employ a regression discontinuity design (RDD) method to estimate the causal impacts of the information intervention on dietary behavior, hemoglobin, cognitive, and educational outcomes of the children. RDD addresses unobserved heterogeneity and omitted variable bias that might affect the knowledge of health risks and food dietary behavior.

Rigorous impact evaluation of health and nutrition information on health-related behavioral change is limited and has found mixed results. Several empirical studies have found positive effects of information treatment on health-promoting behaviors such as water purification (Jalan and Somnathan, 2008; Madajewicz et al., 2007), reduction in risky sexual behavior (Thornton, 2008; Dupas, 2011b), and consumption of fewer calories by diabetes patients (Oster, 2018).3 These studies suggest households in developing countries face information constraints and that health behavior is often responsive to information. Households do respond to learning about their health risks and health-specific information.

In contrast, the literature on nutrition information in the context of anemia is less encouraging. Wong et al. (2014) find modest effects of educating parents about nutrition and anemia on children's hemoglobin levels in north-western China. In an experimental study conducted by Childs et al. (1997), existing doctor-parent contacts were used to convey information about breastfeeding and the link between iron and diet to parents of newborns in the UK; however, no effects were found on anemia after 18 months. Using a relatively small sample of about 250 newborns in Brazil, Bortolini and Vitolo (2012) found that systematic dietary home counseling did not affect the prevalence of anemia, iron deficiency, or iron-deficient anemia. Resistance to dietary modification has also been noticed among both obese and obese-susceptible individuals in the USA (Ogden et al., 2007).

The effectiveness of information treatment depends on the type and intensity of information, who received the information, type of illness, and gender and education of the information recipient (Dupas, 2011a). When health information was shared with children rather than parents in Kenya, the information intervention had no effects on children's behavior (Kremer and Miguel, 2007). In contrast, parents are expected to be more responsive to information about child's health than children. Whether the message is delivered to the mother or father, or both also affects the effectiveness of information. For example, in a study about the use of bednets to prevent malaria in Kenya, Dupas (2009) found women to be more responsive than men to the information. The information campaign was more effective when a message was delivered to both parents instead of delivering it to either parent alone. The education level of the information recipient is another factor that modifies the effect of information intervention. More educated individuals have a higher propensity to change their behavior after receiving information.

While much of the literature relates to preventive healthcare, this study addresses the adoption of remedial behavior. In our treatment, the anemic status of a child is revealed by a diagnostic test. The hemoglobin testing makes explicit the disorder and the need for immediate action. In contrast, with preventive healthcare interventions, some individuals might believe that they will not be affected; hence, the need for preventive action might not be perceived as acute. Furthermore, individuals procrastinate preventive care if the required preventive actions are costly. There are only a few studies that link information on preventive healthy behavior with revealing an individual's health status in the context of HIV (Thornton, 2008), malaria (Cohen et al., 2015), and diabetes (Oster, 2018). In an RCT, Luo et al. (2012) informed the parents of Chinese elementary school children about the anemia status of their child and present strategies for addressing their child's nutritional deficiency (eating balanced meals, including iron-rich products, counseling a doctor, or taking iron supplements). The information was either conveyed by letter, by a single, or by multiple face-to-face information session(s). The different information interventions did not have any impacts on hemoglobin levels or anemia rates. We add to this literature by using a sample of rural Indian households and a very simple information intervention. An additional innovation of this study is not only the assessment of health and nutrition but also of human capital outcomes such as cognition and education. To our knowledge, this is also the first time that RDD was applied in the context of anemia and nutrition information.

Our main results confirm the findings in the previous literature that information alone is less effective in changing nutrition-related behavior, even when combined with the revelation of a nutritional disorder of a child (Luo et al., 2012). Our results show that parents who were informed of the anemic status of their children did not improve their dietary behavior and food consumption. In particular, they did not increase consumption of iron-rich food items. Consistent with insignificant effects on improvement in iron-rich diets, our information treatment effects on hemoglobin levels and cognitive outcomes are statistically insignificant and close to zero. Overall, our results indicate that informing parents about anemia and strategies to mitigate anemia risk for their school-age children appears to have no effect on children's health in rural Bihar, India.

One potential reason for the null findings in our study might be that our experiment has low power. The effect sizes found from related interventions in previous studies on hemoglobin ranged from 0.185 to 0.391 g/dL. Even with a bandwidth of 2.5, our study is only powered to detect an effect size of 0.551 g/dL. The power analysis indicates that our study is powered to detect large effects, and is not powered to detect the smaller effect sizes previously found in the literature from related nutrition interventions.

Our study contributes to several strands of literature. First, this study contributes to the limited and mixed evidence on the effects of an information intervention on household health and nutritional behavior in developing countries (Luo et al., 2012; Zhao et al., 2013; Shimokawa, 2013). Luo et al. (2012) and Shimokawa (2013) found null effects, i.e., being informed of anemia and obesity risk and a balanced diet to reduce these health risks, respectively, had no impacts on dietary behavior. In contrast, Zhao et al. (2013) found that information intervention related to hypertension reduced fat intake in China.4 Second, this paper contributes to the literature on the effects of nutrition information on children's health (Wong et al., 2014; Fitzsimons et al., 2016). While Wong et al. (2014) show that educating parents about nutrition and anemia had a modest impact on anemia in China, Fitzsimons et al. (2016) find that intensive information intervention had significant improvements in the diet and physical growth of infants in rural Malawi. Third, this study relates to the broader literature on the association between child health and the cognitive ability of children in developing countries (Almond et al., 2018). Our study departs from the previous literature by not only considering health and nutrition outcomes but also looking at productive outcomes such as cognition and education. Previous studies show that iron supplementation reduced anemia and led to an improvement in cognitive outcomes in China (Luo et al., 2012) and Peru (Chong et al., 2016).

The remainder of the paper is structured as follows: in Section 2, we describe the intervention, the dataset, and the sample. In Section 3, we present the empirical specification, and describe the results in Section 4. In Section 5, we present robustness checks, and discuss our findings in section 6.

Section snippets

Background

Anemia is a condition in which the level of hemoglobin in the blood is low. Though anemia can have different causes,5 iron deficiency is the most common cause of anemia in low-income countries (WHO, 2001). Anemia can be caused by a variety of

The regression discontinuity design (RDD)

This study uses a regression discontinuity design to estimate the causal impacts of the information intervention on dietary behavior, hemoglobin level, cognition, and educational outcomes of children by comparing children who are just below and above the 11.0 g/dL Hb cutoff.

The anemia status is a deterministic function of Hb level, as children are categorized as anemic if the hemoglobin level is less than 11.0 g/dL and children with a Hb level of at least 11.0 g/dL are non-anemic. Around this

Graphical illustration of the results

Fig. 3 illustrates the potential discontinuities by plotting the change in our outcome variables from 2014 to 2016, against the normalized hemoglobin values in 2014. Due to the normalization of the forcing variable, point 0 at the x-axis is equal to a hemoglobin value of 10.9 g/dL. Panel A of Fig. 3 shows discontinuity graphs for the feeding practice indicators (the dietary diversity score, the frequency of meat consumption, and the frequency of consuming green leafy vegetables), panel B for

Robustness checks

First, the model in Eq. (1) includes ΔY as the outcome variable to control for the baseline value of the outcome variable. However, as discussed before, this may suffer from low power. Therefore, to increase power, we reestimate Eq. (1) and use the endline value of Y as the outcome instead of ΔY. The smallest MDE that can be found without taking differences in outcomes are presented in Table A3. The MDE in Table A3 is slightly smaller than the MDE in Table 8. The estimated effects from this

Discussion

Using the RDD method, this study investigates the impact of a nutrition information intervention on feeding practices, hemoglobin, and cognitive and educational outcomes of young children in rural Bihar, India. After two years of intervention, we find little evidence of statistically significant effects on any of the measured outcomes, indicating that households are resistant to adjust their dietary behavior after acquiring new information about their children's health. Our findings are

Declaration of Competing Interest

The authors declare that they have no conflict of interest.

Acknowledgments

We are grateful for the comments from the editor Laura Schechter and two anonymous reviewers. We thank Aimee Chin, Darren Grant, and David Drukker for helpful suggestions. We also thank Abhijeet Kumar and Zaineb Makati for survey implementation and excellent research support, respectively. We thank the Foundation fiat panis for providing funding for this study, as well as the German Research Foundation (DFG), which provided funding within the scope of the RTG 1666.

References (67)

  • S. Shimokawa

    When does dietary knowledge matter to obesity and overweight prevention

    Food Policy

    (2013)
  • P Thankachan et al.

    Iron absorption in young Indian women: the interaction of iron status with the influence of tea and ascorbic acid.

    Am J Clin Nutr

    (2008)
  • M. Zhao et al.

    Does information on health status lead to a healthier lifestyle? Evidence from China on the effect of hypertension diagnosis on food consumption

    J. Health Econ.

    (2013)
  • D. Almond et al.

    Childhood circumstances and adult outcomes: act II

    J. Econ. Lit.

    (2018)
  • H. Alderman et al.

    Child health and school enrollment: a longitudinal analysis

    J. Hum. Resour.

    (2001)
  • H. Alderman et al.

    Long term consequences of early childhood malnutrition

    Oxf. Econ. Pap.

    (2006)
  • P. Anderson

    Assessment and development of executive function (EF) during childhood

    Child Neuropsychol.

    (2002)
  • ASER. 2014. “Aser tools & testing.”...
  • Banerjee and Duflo. 2012. Poor Economics,...
  • G.J. Bobonis et al.

    Anemia and school participation

    J. Hum. Resour.

    (2006)
  • G.A. Bortolini et al.

    The impact of systematic dietary counseling during the first year of life on prevalence rates of anemia and iron deficiency at 12–16 months

    J. Pediatr. (Rio J)

    (2012)
  • S. Calonico et al.

    Regression discontinuity designs using covariates

    Rev. Econ. Stat.

    (2019)
  • F. Childs et al.

    Dietary education and iron deficiency anaemia in the inner city

    Arch. Dis. Child.

    (1997)
  • A. Chong et al.

    Iron deficiency and schooling attainment in Peru

    Am. Econ. J. Appl. Econ

    (2016)
  • J. Cohen et al.

    Price subsidies, diagnostic tests, and targeting of malaria treatment: evidence from a randomized controlled trial

    Am. Econ. Rev.

    (2015)
  • P. Dupas

    What matters (and what does not) in households' decision to invest in malaria prevention?

    Am. Econ. Rev.

    (2009)
  • P. Dupas

    Health behavior in developing countries

    Annu. Rev. Econ.

    (2011)
  • P. Dupas

    Do teenagers respond to HIV risk information? Evidence from a field experiment in Kenya

    Am. Econ. J. Appl. Econ.

    (2011)
  • FAO. 2007. “Nutrition and consumer protection division. Guidelines for measuring household and individual dietary...
  • Human Vitamin and Mineral Requirements

    (2002)
  • GOI

    Horticultural Statistics at a Glance, Ministry of Agriculture and Farmer’s Welfare

    (2018)
  • J. Hahn et al.

    Identification and estimation of treatment effects with a regression-discontinuity design

    Econometrica

    (2001)
  • JS. Halterman et al.

    Iron deficiency and cognitive achievement among school-aged children and adolescents in the united states

    Pediatrics

    (2001)
  • View full text