What 26,000 stool samples can tell us

TUMIKIA field officers (Photo: William Oswald)
4 August 2016

As the TUMIKIA project reaches its halfway point, London School of Hygiene & Tropical Medicine Research Fellows, William Oswald and Katherine Halliday describe a recent midline survey conducted in households across Kwale county, coastal Kenya, as part of a large-scale cluster randomised trial. 

The TUMIKIA project is a two-year research study to determine whether combining school and community-based deworming is more effective at controlling and eliminating intestinal worms (soil-transmitted helminths [STH]) in Kenya than school-based deworming alone. The findings of this research will contribute to the evidence that governments and the global health community need to design and implement effective control programmes.

The Midline Household Survey

March 2016 marked a year since the TUMIKIA Project began. This meant it was time to launch our midline (12-month follow-up) household survey to measure prevalence and intensity of infection across the three arms of the study, namely: annual school-based deworming of children; annual school- and community-based deworming; and annual school- and twice-yearly community-based deworming. As the name suggests the midline survey comes half way between the baseline survey conducted at the start of the study and the endline (24-month follow-up) survey due in March 2017.

Survey Approach

The data collection, like the project itself, is a big operation. Approaching more than 30,000 households in 120 communities across the often difficult terrain of Kwale County requires a small army of 108 field officers, 39 laboratory technicians, 6 supervisors, 2 managers, and 5 London School of Hygiene & Tropical Medicine staff.

At each household that agreed to participate, the survey included a questionnaire and a household member was randomly selected and asked to provide a sample of their stool for analysis in the lab. Microscopic analysis of stool samples provides a relatively easy way to accurately measure parasitic worm infection. By counting the number of eggs in these faeces we can work out the levels of infection within individuals and across target communities. 

 Stella KephaThe questionnaires were designed to examine how at risk people are to STH infection. Questions about age, occupation, and even house construction, in conjunction with our knowledge of the different modes of transmission of each species of STH, helps paint a picture of infection likelihood. This questionnaire data, collected using smartphones, combined with the stool analysis data allows us to examine potential reasons for observed differences in the levels of infection. 


With so many stool samples it can be difficult keeping track of which sample belongs to which individual. Building on the team’s experience during the baseline survey, an innovation adopted for the midline survey has been the adoption of QR codes to link collected stool samples with their receipt at the laboratory and reading and recording of parasitological results. This greatly simplified the process of linking thousands of samples to their providers.

Blinding data

After two months of very early starts and long days, across thousands of miles covered on foot, by motorcycle, on boats, in vans, we conducted questionnaires with approximately 27,000 households and processed almost 26,000 stool samples. The midline survey was completed on the 12th May.

With all this hard work it is perhaps a little frustrating that we can’t actually look at the data for another year until we ‘unblind’ it. In rigorous evaluations like TUMIKIA, it is good practice not to analyse interim or midline data until the end of the study, in case the findings lead to unconscious biases in how the remaining work is conducted. We’ll have to let you know later what came out of all this poo...hopefully, evidence of fewer worms!