Child Malnutrition

Samoa is an upper-middle income country located in the Western Pacific region with a population of ~ 197,000 and 6100 births each year [18 ]. Polynesia as a whole, and Samoa specifically, are recognized for their extremely high burden of chronic disease (93% of women aged 25–64 are overweight or obese) [19 (link)] and concurrent issues of maternal and child malnutrition (one-third of women of reproductive age are anemic, 29% of children aged 2–5 years are stunted, and 34% anemic) [20 (link)], both of which may be positively impacted by improving breastfeeding practices. While uptake of breastfeeding is high (94% of all infants are breastfed) and median duration of breastfeeding is 21 months, exclusive breastfeeding declines after 2 months of age from 75 to 55% at 4–5 months [18 ].

For the inter-observer study, the required sample size was determined using the sample size design for reliability studies [15 (link),16 (link)]. This method calculates sample sizes for studies where the intra class correlation (ICC) is used as a measure of reliability.
For the repeated measures study, we calculated the sample size for paired studies using the formula described by Fleiss L. J et al [17 (link)], which is dependent on the difference between means and the within-group variability of individual measurements. To establish mean and standard deviation values for this population, we used historical MUAC data of hospitalised children from birth to 5 years collected between the years 2000 to 2005. We assumed that an arbitrary difference of 0.5 cm for MUAC would be clinically relevant.
Calculated sample sizes were n = 53 for the inter-observer study, and n = 285 for the repeated measures study each giving a 90% power of rejecting the null hypothesis of no difference at the 0.05 level of significance. In order to allow for drop out of children from the study, we rounded up these numbers to n = 60 and n = 300 children for the inter-observer and repeated measures study respectively.
Anthropometric Z score calculations for weight-for-length (WFLz), weight-for-age (WFAz), height-for-age (HFAz) and MUAC-for-age (MUACz) were computed using the 2006 WHO standards [2 ]. Statistical analysis was carried out using STATA version 11.0 (STATA Corp, College Station, Texas).
Rehydration solutions (ORS and ReSoMal), and the starter milk formula (F-75) used for stabilization of severely malnourished children provide hardly any energy and little protein, making deposition of lean or fat tissue negligible and unlikely to influence measures. Thus, we assumed that changes in weight and MUAC among acutely ill hospitalized children with signs of dehydration over 48 hours were predominantly due to changes in body water rather than tissue deposition. Absolute, percentage and Z score changes in MUAC and WFLz of the children between the first and the second measurements were calculated. Linear regression, adjusted for age and sex was used to quantify changes associated with rehydration.
Four categories of rehydration were defined: less than 0%, 0-4.9%, 5-9.9%, and 10% or more change in body weight. Changes in MUAC and WFLz for children in these categories of dehydration were calculated. Three categories of wasting were defined as severe, moderate and none using MUAC cut offs of < 115 mm, 115 to 124 mm, and 125 mm or more; and using WFLz by <-3, -3 to -2.01, and -2.01 or more. The proportion of children whose anthropometric nutritional classification changed after 48 hours was calculated with 95% confidence intervals (CI).

All children aged 3–59 months at the time of sampling which took place 1–2 weeks before the first cycle of SPAQ administration and residents in the intervention and control districts were eligible for inclusion. Children with all levels of malnutrition including the severe forms were enrolled and received SPAQ if they were stable and able to take medicines orally; and not on other contraindicated medicines such as those containing sulpha. Unstable children with danger signs were referred to the health facilities for stabilization and further management. Once stabilized, the trained health workers at the facilities would determine an appropriate time to offer SPAQ for protection against malaria.

Primary data were collected using a structured questionnaire and anthropometric measurements of children aged 5 to 17 years. For anthropometric measurements, the children’s height and weight were measured. Weight was measured using a standard digital weight scale, Seca (Model no. 874, Germany), and height was measured using a wall-mountable Bioplus Stature Meter (Model no. 26 M/1,013,522, India). The measurement of weight was recorded to the nearest 0.1 kg and height to the nearest 0.1 cm. Other information was collected by administering a questionnaire to the parents or caretakers of the children or staff at that institution.
Face-to-face or telephone interviews were conducted in the Nepali language to collect the data from parents/caretakers. The questionnaire to the institutional staff included information about the types of disability, overnight stay of CWDs, and socio-demographic information. The parents/caretakers’ included feeding practices, dietary patterns, and maternal and child health-related information. The questionnaire was drafted in this study based on Nepali translation from the existing English questionnaire used to assess malnutrition among children with disabilities in Kenya [8 (link)]. To ensure its validity, the questionnaire was modified to fit the local context. For example, the food names in the 24-hour recall questions were reviewed so that parents/caretakers could understand them. The first author did all interviews and anthropometric measurements.