Pilot testing of a parent record of child health

Diarrhoea is one of the two most significant causes of death in children under five (1,2), with inadequate water, sanitation, and hygiene (WASH) being a significant risk factor for diarrhoea. In several low and middle-income countries (LMICs), water and sanitation interventions are underway and planned to deal with inadequate water and sanitation provision. However, collecting relevant data on whether people are experiencing diarrhoea on a sufficient scale to monitor the effects of water and sanitation interventions has been proven difficult (3). The standard questions used, for example in the Demographic Health Surveys, enquire as to if a child has had diarrhoea over the past two weeks. These retrospective questions with a long recall period suffer from several biases, including recall bias, social desirability, and (in the case of longitudinal studies), the Hawthorne effect. We hypothesise that questioning if a child has had diarrhoea within the last 24 hours, covering the most immediate past period, would improve the accuracy of data collection. This would be combined with other measures, such as re-wording the question to find hidden cases of childhood diarrhoea.
Health questionnaires are increasingly being conducted through digital means, such as telecommunication. In this case, the respondent uses their mobile phone to answer the questionnaire. This method has two immediate advantages: it is not expensive in comparison to data collection in person, and it enables results to be collected from contrasting and moving populations. In this study, we wish to pilot the use of mobile phones to collect data from parents and carers covering child health, with particular emphasis on diarrhoea in one site in Mwanza, Tanzania. Our research questions include: 1) Does changing the wording to ask whether a child has had diarrhoea within the last 24 hours result in lower reported rates of diarrhoea (per day) than the two-week threshold?; 2) Does the use of a three-point scale including the Swahili word for ‘watery diarrhoea" result in lower reporting than the Swahili equivalent of ‘loose stools'?; 3) How does a three-level diarrhoea ‘scale' compare to the standard Demographic and Health Survey (DHS) two-level scale?; and 4) Does the existence of a small incentive influence survey response rates and bias?
In addition, we will harvest information on response rates over time and on peoples' attitude to the telephone questionnaire through qualitative work.