Women's Work in Agriculture and Infant Nutrition

Overview

Background Over the last 10 years there has been an increase in the female agriculture labour force, in Pakistan, resulting in a feminisation of agriculture; which could have either a positive or negative impact on maternal and young child nutrition. It could have a positive impact through increased female wage earnings that improve her bargaining/decision-making power within the household. Women are more likely than men to make pro-nutrition choices with regards to household expenditure. Conversely, women's involvement in agricultural work may have a negative impact on infant or maternal nutrition by reducing time available for child care, through increased expenditure of physical energy without compensatory increases in food consumption or exposure to harmful toxins present in pesticides and other chemicals used in farming. Understanding the dynamics of these pathways, in a specific context, is important to ensure agriculture programmes and policies do not disadvantage women or their children. Overall aim To provide insights into positive and negative pathways between women's work in agriculture and maternal and child nutritional status, in different agriculture workload contexts, to inform agriculture interventions and policies in Pakistan. Specific Objectives 1. To determine whether the number of hours a mother participates in agriculture work is associated with maternal body mass index and infant nutrition. 2. To identify factors that modify the influence of maternal participation in agriculture work on maternal and infant nutritional status. Study Design A cohort study was conducted from September 2015 in irrigated rural areas of Pakistan. Infant-mother dyads were recruited when the infant was between 2 and 12 weeks of age inclusive. Anthropometric measurements (maternal and infant height / recumbent length and weight), interviewer administered questionnaires and spot observations were collected at recruitment (Time 1) and again when the infants were between 9-15 months of age (Time 2). The interviewer administered questionnaires were collected from each infant's mother (or the household head if the father was not present). A one page questionnaire was also completed at recruitment to record the numbers of women who agreed to participate in the study, the number who were approached but were not recruited into the study and the reasons they were not eligible to participate or their reasons for refusal.

Study Design This study is an longitudinal observational study which was used to generate hypotheses rather than testing them. Mother-infant dyads were recruited between December 2015 and February 2016; and were followed up between November 2016 and January 2017. Sampling The sample size (n=1000) was calculated to detect an increase/decrease in maternal BMI of around 0.18 for every additional hour worked with 80% power at a 5% level of significance. To estimate it, simulations were run to explore the power we would have to detect the relationship between maternal BMI and number of hours worked. It was based on a mean number of working hours of 6 and a standard deviation of 4 and a within cluster variation of 0.2 . This is a small effect size and suggests that the proposed sample size will provide adequate power to allow us to explore proposed relationships and generate hypotheses. Participants were selected via systematic random cluster sampling. Initially, administrative villages with perennial canal irrigation were selected; villages with a population\<10% or \>90% of average village size were excluded, and random sampling was used to select villages from the eligible villages (n=2,909). All dyads, in the selected villages (n=62) were invited to participate in the study if: (i) the infant was a singleton birth ≥2 weeks and ≤ 12 weeks of age on the day of the first interview; healthy without congenital deformations that would impact on their ability to eat and (ii) the primary caregiver (i.e. the biological mother) intended to reside in the study area over the next 10 months. To recruit these dyads, all recent births in the identified village were listed through a systematic multi-stage community profiling procedure using local key informants/resource persons including: health workers, midwives, doctors and paramedics, and locally well-informed individuals. In the first stage, several key informants per village were asked to exhaustively list all kinship groups or castes and localities within the village. Then they were asked to list all births from within those castes/localities within a given time period. Probes were used at both stages to counter exclusions. Afterwards a team of recruiters visited each of the listed mothers to confirm eligibility (i.e. correct age of the new born). Recruiters also probed for other births within the locality, to identify new cases through this snowballing technique. All dyads in the village who fulfilled this condition and met the inclusion criteria were recruited and actual date of birth was recorded. Procedures and Methodology At each of the two data collection periods, interviewer administered questionnaires were collected from the mothers; and included questions related to socio-demographic status, health, dietary intakes, maternal agency and nutrition knowledge. These questionnaires were pilot tested and modified to ensure each questionnaire took less than 60 minutes per respondent to complete. For mothers, questions related to infant feeding, immunisation and pregnancy varied between the first and second data collection periods to reflect differences in infant age and maternal status. All data were entered directly into an android tablet. Anthropometric measurements of maternal height and weight and infant weight and recumbent length were done in duplicate by trained anthropometrists using high quality equipment. If the first and second measurements did not agree to within a specified limit (e.g., 0.5 kg or 0.5 cm) a third measurement was taken. These data were entered directly into an android tablet to allow consistency checks for quality assurance. Spot observations were also done to determine housing materials and the hygienic conditions of the environment, which were entered directly into an android tablet.