RESUMEN
The increasing availability and capabilities of mobile phones make them a feasible means of data collection. Electronic Data Capture (EDC) systems have been used widely for public health monitoring and surveillance activities, but documentation of their use in complicated research studies requiring multiple systems is limited. This paper shares our experiences of designing and implementing a complex multi-component EDC system for a community-based four-armed cluster-Randomised Controlled Trial in the rural plains of Nepal, to help other researchers planning to use EDC for complex studies in low-income settings. We designed and implemented three interrelated mobile phone data collection systems to enrol and follow-up pregnant women (trial participants), and to support the implementation of trial interventions (women's groups, food and cash transfers). 720 field staff used basic phones to send simple coded text messages, 539 women's group facilitators used Android smartphones with Open Data Kit Collect, and 112 Interviewers, Coordinators and Supervisors used smartphones with CommCare. Barcoded photo ID cards encoded with participant information were generated for each enrolled woman. Automated systems were developed to download, recode and merge data for nearly real-time access by researchers. The systems were successfully rolled out and used by 1371 staff. A total of 25,089 pregnant women were enrolled, and 17,839 follow-up forms completed. Women's group facilitators recorded 5717 women's groups and the distribution of 14,647 food and 13,482 cash transfers. Using EDC sped up data collection and processing, although time needed for programming and set-up delayed the study inception. EDC using three interlinked mobile data management systems (FrontlineSMS, ODK and CommCare) was a feasible and effective method of data capture in a complex large-scale trial in the plains of Nepal. Despite challenges including prolonged set-up times, the systems met multiple data collection needs for users with varying levels of literacy and experience.
Asunto(s)
Teléfono Celular , Recolección de Datos/métodos , Envío de Mensajes de Texto , Adulto , Femenino , Humanos , Nepal , Embarazo , Asistencia Pública , Población Rural , Apoyo SocialRESUMEN
BACKGROUND: Low birth weight (LBW, < 2500 g) affects one third of newborn infants in rural south Asia and compromises child survival, infant growth, educational performance and economic prospects. We aimed to assess the impact on birth weight and weight-for-age Z-score in children aged 0-16 months of a nutrition Participatory Learning and Action behaviour change strategy (PLA) for pregnant women through women's groups, with or without unconditional transfers of food or cash to pregnant women in two districts of southern Nepal. METHODS: The study is a cluster randomised controlled trial (non-blinded). PLA comprises women's groups that discuss, and form strategies about, nutrition in pregnancy, low birth weight and hygiene. Women receive up to 7 monthly transfers per pregnancy: cash is NPR 750 (~US$7) and food is 10 kg of fortified sweetened wheat-soya Super Cereal per month. The unit of randomisation is a rural village development committee (VDC) cluster (population 4000-9200, mean 6150) in southern Dhanusha or Mahottari districts. 80 VDCs are randomised to four arms using a participatory 'tombola' method. Twenty clusters each receive: PLA; PLA plus food; PLA plus cash; and standard care (control). Participants are (mostly Maithili-speaking) pregnant women identified from 8 weeks' gestation onwards, and their infants (target sample size 8880 birth weights). After pregnancy verification, mothers may be followed up in early and late pregnancy, within 72 h, after 42 days and within 22 months of birth. Outcomes pertain to the individual level. Primary outcomes include birth weight within 72 h of birth and infant weight-for-age Z-score measured cross-sectionally on children born of the study. Secondary outcomes include prevalence of LBW, eating behaviour and weight during pregnancy, maternal and newborn illness, preterm delivery, miscarriage, stillbirth or neonatal mortality, infant Z-scores for length-for-age and weight-for-length, head circumference, and postnatal maternal BMI and mid-upper arm circumference. Exposure to women's groups, food or cash transfers, home visits, and group interventions are measured. DISCUSSION: Determining the relative importance to birth weight and early childhood nutrition of adding food or cash transfers to PLA women's groups will inform design of nutrition interventions in pregnancy. TRIAL REGISTRATION: ISRCTN75964374 , 12 Jul 2013.