Skip to main content

Table 3 Lessons learnt and accompanying checklist for planning the design and implementation of community-based maternal, newborn and infant health studies

From: Implementing effective community-based surveillance in research studies of maternal, newborn and infant outcomes in low resource settings

  Checklist of items to consider at the planning stage: Solutions
1. Understand your population What are the cultural and social norms on disclosing a pregnancy in the population?
What are the cultural and social norms and behaviours relating to delivery in the population?
• Conduct formative research to understand the norms around disclosing a pregnancy in your population
• Consider fieldworker characteristics, location, training, and supervision, and how these might affect a woman’s likelihood to discuss their pregnancy with them
• Conduct formative research to understand behaviours around delivery (in particular, whether women move shortly before and after delivery and how they make plans for delivery)
• Collect data from women on their delivery plans (including planned place of residence and place of delivery), so that data collection and intervention delivery issues can be mitigated
2. Plan your data collection cycle How frequent should study visits be?
Does the frequency of visits need to change as participants progress through the study?
Do you need other methods of data collection to improve detection rates in between visits?
Do you need to recruit the mothers before they give birth?
When will study visits end?
• Plan for study visits that are as frequent as resources allow
• Consider intervention delivery requirements in addition to data collection requirements
• Consider the issues that might lead to changes in visit frequency as participants progress through the study (e.g., instigate more frequent visits once a woman is pregnant, or in the later stages of a pregnancy if births need to be detected quickly)
• Consider alternative methods of data collection (e.g., contact with key informants, mobile phone use)
• Consider whether women need to be enrolled in the surveillance system before the pregnancy outcome (e.g., this may be essential for intervention delivery in some cases, but not in others)
• Plan the last visit at a reasonable time after the end of the period of interest (e.g., to capture maternal outcomes at the end of the 6-week postpartum period, plan the visit for the 8–12th week or later if the data collection is not time-sensitive)
3. Enhance routine surveillance with additional data collection methods Can you use data from multiple sources to identify pregnancies and outcomes, and for triangulation?
Have you developed a strategy for dealing with inconsistencies between different sources?
Is additional confirmation needed for some data?
• Consider which additional data sources are useful, including:
-Data collection at hospitals or clinics (e.g., for data on morbidities; also good for identifying pregnancies not reported in the field and for confirming dates)
-Verbal autopsies (for cause of death information; also good for identifying pregnancies not reported in the field and for confirming dates)
• Include resources to employ, train and supervise senior staff to collect data form these additional sources
• Incorporate strategies for dealing with inconsistencies between sources a priori into the data cleaning plan
• Consider whether additional confirmation is needed for some data (e.g., is a woman’s self-report of a pregnancy adequate or is formal pregnancy testing needed?)
4. Design a field and data management system that is adaptable in real-time Which events should trigger changes in data collection or intervention procedures?
Have you developed a mechanism for reporting and dealing with data errors, in real-time if necessary?
• Consider whether data collection will be different at different points in the study, and how this can be changed in real-time (e.g., do you want to collect more data from women once they are pregnant?)
• Consider whether there needs to be regular changes to work listings (e.g., do you want to collect data on mothers and infants after birth, and should the infant appear on a fieldworker’s work listings?)
• Allow for appropriate intervals in the production of work listings (e.g., allowing for mourning periods before collecting verbal autopsy data)
• Consider the specific errors that may occur with data collection on mothers and infants and develop a plan that will allow for the correction of these in real-time