Little is known about the long-term effects of maternal multi-micronutrient supplementation on the growth of children during early childhood. In this follow-up study, the effects of maternal supplementation with multi-micronutrients in pregnancy on postnatal child growth are examined.
A longitudinal follow-up of a subset of newborns (n=1388) whose mothers were randomly assigned to receive the supplements of folic acid, iron–folic acid or multi-micronutrients daily during pregnancy in the original trial was conducted. Children’s weight and length were measured and assessed during monthly home visits from birth to 30 months of age.
The pooled prevalence rate of stunting over different time points during the first 30 months was 13.5, 14.9 and 12.1% for the folic acid group, iron–folic acid group and multi-micronutrient group, respectively. However, there were no significant differences in the pooled odds of stunting in children between the multi-micronutrient group and the folic acid (odds ratio (OR) 0.97, 95% confidence interval (CI): 0.74–1.26), and between the multi-micronutrient group and the iron–folic acid group (OR 0.82, 95% CI: 0.63–1.07). Similar results for the three groups were found for the occurrences of underweight and wasting in children. Furthermore, no significant differences were observed in length, weight, length-for-age, weight-for-age and weight-for-length Z-scores among the three treatment groups.
Currently, available evidence is insufficient to support a greater advantage of the effect of maternal multi-micronutrient supplementation on child growth over iron–folic acid or folic acid only supplementation during the first 30 months.
Pregnant women in developing countries are particularly vulnerable to multiple micronutrient deficiencies (Pathak et al., 2004; Jiang et al., 2005; Kontic-Vucinic et al., 2006; Molloy et al., 2008). This, combined with poor dietary quality, undernourishment and anemia (Seshadri, 2001; Ma et al., 2004; Zeng et al., 2006; Pathak et al., 2007), increases the risk of intrauterine growth restriction, preterm delivery and low birth weight (Lindblad et al., 2005; Molloy et al., 2008), thereby conferring the risk of long-term adverse health outcomes for the child. A formulation for a multiple micronutrient supplement for pregnant women in developing countries was proposed by UNICEF and WHO (UNICEF/UNU/WHO, 1999). It is expected that providing pregnant women with a full complement of micronutrients will have health benefits, both for the pregnant woman and the child.
Supplementation with iron and folic acid is routinely recommended during pregnancy to avoid the harmful effect of anemia on birth weight (Allen, 2000; Juarez-Vazquez et al., 2002; Schümanna et al., 2007). It has been reported that iron plus folic acid supplementation is more effective than folic acid supplementation alone in the reduction of the risk of low birth weight (Christian et al., 2003), but a trial in China did not show a significant difference in birth weight between iron–folic acid supplementation and folic acid supplementation alone (Zeng et al., 2008). Recently, a meta analysis (Kawai et al., 2011) of 15 randomized controlled trials in developing countries reported that multi-micronutrient supplementation during pregnancy was more effective than iron and folic acid supplementation at increase birth weight (42 g, 95% confidence interval (CI):28–60) and reducing the risk of low birth weight (relative risk (RR): 0.86, 95% CI: 0.79–0.93) and of small size for gestational age (RR: 0.85; 95% CI: 0.78–0.93). A study in Nepal found that the children of prenatal multi-micronutrient supplementation were 204 g heavier at 30 months of age than those of the iron+folic acid control. A similar result was reported by a non-randomized, non-blinded trial in Vietnam, suggesting that the children aged of 2 years old were taller, and the stunting rates were lower in multiple micronutrients than in iron–folic acid (Huy et al., 2009). However, a study of school-age children in rural Nepal randomly exposed to prenatal micronutrient supplements showed no impact of multiple micronutrient supplementation on linear growth, blood pressure, cholesterol and measures of insulin resistance (Stewart et al., 2009a, 2009b).
With our knowledge, the effect of antenatal micronutrients supplementation on infant growth is still unclear. To investigate whether micronutrient supplementation can improve the growth of infants and children, we followed up a subgroup of children whose mothers were enrolled in a randomized trial of micronutrient supplementation in pregnancy from birth to the age of 30 months.
Subjects and methods
The study was a longitudinal follow-up of children whose mothers had been enrolled in a double-blind cluster, randomized controlled trial conducted in two poor rural counties in Shaanxi Province of northwest China. The details about this study have been described elsewhere (Zeng et al., 2008).
In the original trial, villages were randomly assigned to three supplementation groups (daily folic acid, iron–folic acid or multi-micronutrient supplements). Pregnant women resident in the counties between August 2002 and January 2006 were enrolled after signing the consent form and received one of the three daily antenatal micronutrient supplements until delivery. The three supplements were folate (400 μg), iron (60 mg) plus folic acid (400 μg), and multi-micronutrients of 15 minerals and vitamins (30 mg iron, 400 μg folate, 15.0 mg zinc, 2.0 mg copper, 65.0 μg selenium, 150.0 μg iodine, 800.0 μg vitamin A, 1.4 mg vitamin B1, 1.4 mg vitamin B2, 1.9 mg vitamin B6, 2.6 μg vitamin B12, 5.0 μg vitamin D, 70.0 mg vitamin C, 10.0 mg vitamin E and 18.0 mg niacin) (UNICEF/UNU/WHO, 1999). The three supplement types were of identical appearance and packaged in blister packs of 15 tablets each. The group assignment was randomized and double-blinded. The Ministry of Health in China and the Ethical Committees of the Medical College of Xi’an Jiaotong University approved the follow-up study.
Study population and sample size
The research teams recruited the cohort of pregnant women in Jan 2004 and obtained their consent to enroll in the sub-study at the third trimester antenatal care visit. Live-birth singleton infants were invited to participate in the sub-study; babies with major congenital abnormalities were excluded.
We estimated that a minimum of 1173 infants (391 infants per group) were needed to detect a 33% reduction in the prevalence rate of stunting of 24% (based on several studies conducted from 1999 to 2003 in 40 counties in western China (Lingxia et al., 2003)) between either multiple micronutrient and folic acid, with or without iron groups, with 80% power and α=0.05.
A subgroup of singleton neonates (born in 2004, the middle 1 year of the total 3.5 years of recruitment until January 2006) was included in the postnatal surveillance and followed up until 30 months of age.
Primary carers were interviewed about feeding practice, time spent outdoors per day, and morbidity of their infants during home visits, which were at 3-month intervals from birth up to 12 months, and at 6-month intervals thereafter. At all visits, anthropometric measurements and physical examinations were conducted by field teams of trained research assistants and nurses.
Infant feeding was categorized into the following groups: (1) non-breastfeeding; (2) partly breastfeeding (receiving some breast milk plus liquid or solid foods); (3) exclusive breastfeeding (receiving only breast milk without any milk or solid foods with the exception of water and routine vitamins).
At each home visit, the caretaker was asked to recall the infant’s symptoms of morbidity and the duration the infant suffered from that symptom in the 30 days preceding the visit. The symptoms were recalled as the following: diarrhea, defined as ⩾3 loose or liquid motions per 24 h period; fever (hot to the touch); runny nose; cough; difficulty breathing, breathing with severe noise or wheezing or difficulty inhaling.
The weight and length of infants and children were measured using standardized methods (WHO, 1995). Weight was measured to the nearest 10 g on an electronic scale, excluding the weight of children’s clothes. Scales were checked by calibration with a standard weight (10 kg) at regular intervals. Recumbent length was measured to the nearest 1 mm on a length board. All field assistants and nurses received training and were supervised regularly and remained blinded to the treatment group until the end of the follow-up period.
Data entry and management was done with Microsoft Access 2000. Anthropometric indices weight-for-age Z-score, length-for-age Z-score and weight-for-length Z-score were calculated with WHO Anthro 2005 (WHO, 2006) and WHO Child Growth Standards 2006 (de Onis et al., 2006). Stunting, underweight and wasting were defined from length-for-age, weight-for-age, and weight-for-length Z-scores, respectively, as less than −2. As the results of birth measurements had been reported previously (Zeng et al., 2008), the birth measurements were considered as baseline data and the analysis was restricted to the ages between 1 to 30 months.
A household wealth index was constructed from an inventory of 16 household assets or facilities with a principal component analysis method (Filmer and Pritchett, 2001).
The primary outcome was the occurrence of stunting at each visit during the first 30 months. The occurrences of underweight and wasting were secondary endpoints. Other secondary endpoints included length, weight, length-for-age, weight-for-age and weight-for-length Z-scores at each visit during the first 30 months.
The primary endpoint was analyzed using a mixed-effects logistic regression model. In this model, treatment, age of child and some potential confounding factors were treated as fixed effects, whereas village and individual subject were treated as random effects. The above model yields the pooled prevalence rates of stunting and odds ratios (ORs) for the occurrence of stunting over different age points during the first 30 months, compared with different treatment groups. Furthermore, an interaction term between treatment and age was added to the above model to derive the ORs at each age point if the interaction term was the statistical significance at 5%. The occurrences of underweight and wasting were analyzed in a similar way.
For the analysis of a continuous outcome of child growth (length, weight, length-for-age, weight-for-age and weight-for-length Z-scores), we estimated mixed linear regression models that included the baseline measurement of a given outcome, treatment, age of child, together with the potential confounding factors as fixed effects, whereas the village and individual subject were treated as random effects. In addition, interaction effects between age and treatment were also tested.
The potential confounding factors included gender, birth weight, preterm, parity, feed methods, time of stay at outdoor, illness or health in last month before the interview, mother's height, education level, occupation, number of supplement tablets consumed, and family socio-economic status at baseline.
Treatment differences in various outcomes together with their 95% CI were reported. A two-tailed P-value of <0.05 was considered to be statistically significant. All statistical analyses were carried out using the Statistical Analysis System (SAS) version 9.2 (SAS Institute, Inc., Cary, NC, USA).
Figure 1 shows the trial profile. Of the 5828 women enrolled in the original trial, 1400 singleton live births between January 2004 and December 2004 were eligible for the follow-up study. Among these children, 12 infants were excluded from analysis at the final stage due to congenital diseases (n=7), death (n=3) and parent's problems (n=2). The study included 1388 children who were followed from 1 month up to 30 months of age or death. Retention rates in the three supplement groups from 1 to 30 months of age after birth were 80.6–85.7%. The lost cases were mainly due to the children’s parents having moved outside the study region. Measurements collected before children were lost to follow-up were still included in the analysis.
The characteristics of the pregnant women and their infants in the three treatment groups were comparable at baseline (Table 1). The mean gestation and birth weight, mean age of mother at delivery, mean height and years of education of mother at enrollment, mean number of supplements consumed and mean value of household wealth index were similar. Among the three treatment groups, there was no significant difference in the proportions of male infants, low birth weight (<2500 g), preterm (gestation <37 weeks) and primipara. There is no statistically significant difference in the baseline characteristics between follow-up and lost to follow-up children by three treatment groups.
Growth pattern of the children
The pooled mean length and weight (s.e.) over different age points for the groups supplemented with folic acid, iron–folic acid and multi-micronutrients were estimated as 72.25±0.09, 72.06±0.09 and 72.15±0.09 cm, and 8.99±0.03, 8.90±0.04 and 8.95±0.04 kg, respectively, from the mixed model analysis adjusting for the given outcome measurement at baseline and selected confounding factors (Table 2). Compared with children whose mothers received folic acid or iron–folic acid supplements, maternal supplementation with multi-micronutrients did not result in significant difference of length and weight of children from 1 month to 30 months of age. The P-values for the tests of interaction between age and treatment were all greater than 5%, so age-specific differences in weight and height were presented.
Compared with the folic acid or the iron–folic acid groups, the multi-micronutrient group showed no significant difference in length-for-age, weight-for-age and weight-for-length Z-scores from 1 month to 30 month of age (Table 2).
Prevalence and incidence of malnutrition
The pooled prevalence rate of stunting over different age points during the first 30 months was 13.5, 14.9 and 12.1% for folic acid group, iron–folic acid group and multi-micronutrient group, respectively. Although the pooled prevalence rate of stunting in the multi-micronutrient group was lowest (12.1%) among the three treatment groups, there were no significant differences in the pooled odds of stunting of children between the multi-micronutrient group and the folic acid (OR 0.97, 95% CI: 0.74–1.26), and between the multi-micronutrient group and the iron–folic acid group (OR 0.82, 95% CI: 0.63–1.07; Table 3). The differences in the prevalence rates of underweight and wasting between the multi-micronutrient group and the folic acid group or the iron–folic acid group were not significant, although the values of OR were under 1.0 in favor of multi-micronutrient group. For these outcomes, the tests for the interactions between treatment and age were not significant at 5%, suggesting no age-specific treatment differences.
We examined the effects of prenatal maternal supplementation with multi-micronutrients on the growth of children during the first 30 months of life in rural western China. Our findings did not provide evidence that maternal supplementation with multi-micronutrients daily during pregnancy was more effective than only folic acid or iron+folic acid supplementations at improving the growth of infants and children during the first 30 months of age. It is worth mentioning that this study was designed as a follow-up study of an efficacy randomized, controlled trial. The potential confounders at baseline were controlled in the multivariable regression analyses. During the follow-up, double blinding was effective and the retention rate was high. All anthropometric indicators were measured carefully and were checked with regular quality control to ensure the accuracy of measurements.
The lack of significant effect cannot be explained by selection bias, as significant differences in mental development scores of infants at 1 year of age were observed between multi-micronutrient supplementation and folic acid or iron–folic acid supplementation (Li et al., 2009). The current sub-study sample and the original sample (Zeng et al., 2008) are comparable in terms of birth weight, birth length and low-birth weight rate in the different supplementation groups, respectively. Although the incidence rate of preterm is higher in original trial than in the subset, the difference is not statistically significant. Even though cluster randomization was performed by village with approximately equal distribution across treatment groups, we included a large number of villages (N=531), and the average number of subjects in each village was small. All statistical analysis models used were to accommodate the cluster sampling method and possible correction due to repeated measurements. However, the final sample size calculated for the study had inadequate power (<80%) to detect a reduction of 33% in stunting, because the observed prevalence rate of the primary outcomes was lower than expected.
Our study was the first to assess the long-term effect of prenatal multi-micronutrient supplements on the continuous growth from birth to infants and children, a topic rarely addressed in existing research. A randomized controlled trial of antenatal micronutrient supplements among school-age children in rural Nepal reported that there was no significant difference in linear growth among prenatal supplementations with multiple micronutrients, folic acid, iron plus folic acid and the control (Stewart et al., 2009a). This finding is similar to ours, although postnatal follow-up times are different between two studies. It was also reported by Stewart et al. (2009a) that antenatal supplementation with folic acid+iron+zinc improves linear growth in 6- to 8-year-old children in rural Nepal, meanwhile supplementation with multi-micronutrient had no such effect. The findings suggested the possibility of nutrient interactions in the multi-micronutrient supplement that may have inhibited the effects of zinc on linear growth (Stewart et al., 2009a).
In a previous report, the significant effect of maternal supplementation with multi-micronutrients on birth weight in our original trial (Zeng et al., 2008) was 42 g greater than the folic acid group, but our current report suggests that this effect did not last beyond 1 month after their birth. In contrast with our findings, a closely related study conducted in Nepal (Vaidya et al., 2008) showed that children whose mothers had antenatal multi-micronutrients supplementation were an average of 77 g heavier at birth and 204 g heavier at 30 months of age than those of the iron and folic acid supplements. One possible explanation could be due to the differences in the two study areas. For example, the nutritional status of the subjects (the mean birth weight) was nearly 500 g lower in Nepal than in China. The prevalence rates at 30 months of stunting and underweight children were over 80% and 70% in Nepal (much higher than in China). The extent of micronutrient deficiency of the subjects’ mothers and disease epidemiology may also be different between two study populations. The average of mother’s height was 8–9 cm lower in Nepal than in China. Another possible explanation was that during the period of follow-up of this study, a new co-operative medical service was introduced in the two study counties (Mu et al., 2005). Free provision of high-quality antenatal care, perinatal medical support and regular physical examination for children, perhaps may have contributed to lower incidences of stunting and underweight children. The third possible explanation for the lack of effects may be that the follow-up time was not long enough to reveal the true effect of maternal multi-micronutrient supplementation on children. In a Tanzania trial, the effect of a daily high dose multi-vitamin supplementation of HIV-infected women on child growth (except birth weight) was not significant until the age of 2 years old (Villamor et al., 2005). It was suggested that the effect of antenatal multi-vitamin supplementation is limited in early childhood. It is also reasonable to expect the long-term effect of maternal multi-micronutrient supplementation increased with the age of children, as the risk of stunting, underweight and wasting of children in maternal multi-micronutrient supplementation was less than that of supplementation with folic acid or iron–folic acid, although these findings did not achieve statistical significance.
This study could not provide sufficient evidence to support a greater advantage of the effect of maternal multi-micronutrient supplementation on child growth over folic plus iron, or folic acid only supplementation during the first 30 months. Further follow-up investigation is needed to assess the long-term effect.
Allen LH (2000). Anemia and iron deficiency: effects on pregnancy outcome. Am J Clin Nutr 71 (Suppl), 1280S–1284S.
Christian P, Khatry SK, Katz J, Pradhan EK, LeClerq S, Shrestha SR et al. (2003). Effects of alternative maternal micronutrient supplements on low birth weight in rural Nepal: double blind randomised community trial. BMJ 326, 571–576.
de Onis M, Garza C, Onyango AW, Martorell R (eds) (2006). WHO child growth standards. Acta Paediatrics 95 (Suppl 450), S1–S101.
Filmer D, Pritchett L (2001). Estimating wealth effects without expenditure data-or tears: an application to educational enrollments in states of India. Demography 38, 115–132.
Huy ND, Le Hop T, Shrimpton R, Hoa CV (2009). An effectiveness trial of multiple micronutrient supplementation during pregnancy in Vietnam: impact on birthweight and on stunting in children at around 2 years of age. Food Nutr Bull 30 (4 Suppl), S506–S516.
Jiang T, Christian P, Khatry SK, Wu L, West KP (2005). Micronutrient deficiencies in early pregnancy are common, concurrent, and vary by season among rural Nepali pregnant women. J Nutr 135, 1106–1112.
Juarez-Vazquez J, Bonizzoni E, Scotti A (2002). Iron plus folate is more effective than iron alone in the treatment of iron deficiency anemia in pregnancy: a randomized, double blind clinical trial. BJOG 109, 1009–1014.
Kawai K, Spiegelman D, Shankar AH, Fawzi WW (2011). Maternal multiple micronutrient supplemen pregnancy outcomes in developing countrie and meta-regression. Bull World Health Organ 89, 402–411.
Kontic-Vucinic O, Sulovic N, Radunovic N (2006). Micronutrients in women's reproductive health: I. Vitamins. Int J Fertil Womens Med 51, 106–115.
Li Q, Yan H, Zeng L, Cheng Y, Liang W, Dang S et al. (2009). Effects of maternal multimicronutrient supplementation on the mental development of infants in rural western China: follow-up evaluation of a double-blind, randomised, controlled trial. Pediatrics 123, 685–692.
Lindblad B, Zaman S, Malik A, Martin H, Ekström AM, Amu S et al. (2005). Folate, vitamin B12, and homocysteine levels in South Asian women with growth-retarded fetuses. Acta Obstet Gynecol Scand 84, 1055–1061.
Lingxia Z, Hong Y, Xiong G, Shaonong D, Hong X (2003). Analysis on malnutrition of children under 3 years old in 40 poor counties in the western areas of China. Chin J Public Health 19, 55–58.
Ma A, Chen X, Wang Y, Xu R, Zheng M, Li J (2004). The multiple vitamin status of Chinese pregnant women with anemia and nonanemia in the last trimester. J Nutr Sci Vitaminol (Tokyo) 50, 87–92.
Molloy AM, Kirke PN, Brody LC, Scott JM, Mills JL (2008). Effects of folate and vitamin B12 deficiencies during pregnancy on fetal, infant, and child development. Food Nutr Bull 29 (2 Suppl), S101–S111.
Mu L, Lingxia Z, Jinping W, Lihua C, Yanming L, Haili L et al. (2005). Analysis of prevalence study on maternity and child care in the experimental units of new-type rural co-medical treatment of Shaanxi province. J Xi’an Jiaotong Univ Med Sci 26, 392–394.
Pathak P, Kapil U, Kapoor SK, Saxena R, Kumar A, Gupta N et al. (2004). Prevalence of multiple micronutrient deficiencies amongst pregnant women in a rural area of Haryana. Indian J Pediatr 71, 1007–1014.
Pathak P, Kapil U, Yajnik CS, Kapoor SK, Dwivedi SN, Singh R (2007). Iron, folate, and vitamin B12 stores among pregnant women in a rural area of Haryana State, India. Food Nutr Bull 28, 435–438.
Schümanna K, Ettle T, Szegner B, Elsenhans B, Solomons NW (2007). On risks and benefits of iron supplementation recommendations for iron intake revisited. J Trace Elements Med Biol 21, 147–168.
Seshadri S (2001). Prevalence of micronutrient deficiency particularly of iron, zinc and folic acid in pregnant women in South East Asia. Br J Nutr 85 (Suppl 2), S87–S92.
Stewart CP, Christian P, Le Clerq SC, West KP, Khatry SK (2009a). Antenatal supplementation with folic acid + iron + zinc improves linear growth and reduces peripheral adiposity in school-age children in rural Nepal. Am J Clin Nutr 90, 132–140.
Stewart CP, Christian P, Schulze KJ, Le Clerq SC, West KP, Khatry SK (2009b). Antenatal micronutrient supplementation reduces metabolic syndrome in 6- to 8-year-old children in rural Nepal. J Nutr 139, 1575–1581.
UNICEF/UNU/WHO (1999). Composition of a Multi-Micronutrient Supplement to be Used in Pilot Programmes Among Pregnant Women in Developing Countries. UNICEF: New York, NY.
Vaidya A, Saville N, Shrestha BP, Costello AM, Manandhar DS, Osrin D (2008). Effects of antenatal multiple micronutrient supplementation on children's weight and size at 2 years of age in Nepal: follow-up of a double-blind randomised controlled trial. Lancet 371, 492–499.
Villamor E, Saathoff E, Bosch RJ, Hertzmark E, Baylin A, Manji K et al. (2005). Vitamin supplementation of HIV-infected women improves postnatal child growth. Am J Clin Nutr 81, 880–888.
WHO (1995). Physical Status: The Use and Interpretation of Anthropometry. Expert Committee Report. WHO Technical Report Series No. 854. World Health Organization: Geneva (http://whqlibdoc.who.int/trs/WHO_TRS_854.pdf).
WHO (2006). Anthro 2005 for mobile devices version 1.00 SfagadotwscGW (http://www.who.int/childgrowth/software/en/).
Zeng L, Cheng Y, Dang S, Yan H, Dibley M, Chang S et al. (2008). Impact of micronutrient supplementation during pregnancy on birth weight, duration of gestation, and perinatal mortality in rural western China: double blind cluster randomised controlled trial. BMJ 337, a2001–a2011.
Zeng F, Song YY, Li Z (2006). A research on relationship between fetal growth restriction and vitamin nutritional status in late pregnancy. Chin J Matern Child Health Res 17, 4–5.
This study was supported by United Nations Children's Fund (Grant Number YH101-H12/03) through a cooperative agreement between UNICEF and the Centers for Disease Control and Prevention, Atlanta, US, and the National Natural Science of Foundation of China (Grant Number 30271131), Beijing, China. We are grateful to all the participants and their families who participated in the trial; the field-team members; Gong Aijing and Pan Yonghe who entered data. We thank the field supervisors Li Yanqing, Kang Yijun, Xing Yuan, Fang Bo, Liang Wei-feng, Wang Bei, Duan Sheng-gang and Shen Yuan who helped in the field procedures. This study has been registered as an International Standard Randomized Controlled Trial (Number ISRCTN08850194).
This study has been registered as an International Standard Randomized Controlled Trial (No. ISRCTN08850194).
The authors declare no conflict of interest.
Contributors: All authors contributed to the design and implementation of the study. WW coordinated the study, trained field stuff, supervised the quality control, collected data, cleaned, analyzed and interpreted the data, and drafted the manuscript. HY designed the study and supervised the field operation, data analysis, supervised the manuscript preparation and revised the paper. LZ developed the protocol and the field procedures, made substantial contributions to the execution and supervision of the field investigation. YC contributed to the supervision of the field investigation and data management. DW analysed the data independently and reviewed and revised the paper. QL helped the data analysis. All authors contributed to the manuscript and approved the final version.
About this article
Effects of Maternal Prenatal Multi-Micronutrient Supplementation on Growth and Development until 3 Years of Age
International Journal of Environmental Research and Public Health (2019)
Cochrane Database of Systematic Reviews (2019)
The impact of folic acid supplementation on gestational and long term health: Critical temporal windows, benefits and risks
Porto Biomedical Journal (2017)
Effect of low-dose versus higher-dose antenatal iron supplementation on child health outcomes at 36 months of age in Viet Nam: longitudinal follow-up of a cluster randomised controlled trial
BMJ Global Health (2017)
Cochrane Database of Systematic Reviews (2017)