The development of an Android platform to undertake a discrete choice experiment in a low resource setting

Discrete choice experiment (DCE) is a quantitative technique which helps determine preferences from a definite set of choices. DCEs have been widely used to inform health services in high-income country settings and is gradually being used in low and middle-income countries (LMICs). There are challenges in deploying this method in LMIC settings due to the contextual, cultural and language related barriers. Most DCEs are conducted using paper-based tools. With mobile technology readily accessible across LMICs, we developed an Android-based platform to conduct a DCE among community health workers (CHWs) in rural India.

This paper describes the development of a DCE for low-literacy community health workers (CHWs) in low-resourced setting on an Android platform. We illustrate the process of identifying realistic and locally relevant attributes, finalising the tool and cognitively testing it among respondents with an average of 10 years of education using 'think aloud' and 'verbal probing' techniques. The Android application was tested in two rounds, first by the research team and second, by the CHWs. The 'think aloud' and 'verbal probing' techniques were essential in assessing the comprehension of the CHWs to the DCE choices.

The CHWs did not take much time to familiarize themselves with the Android application. Compared to the paper based DCE, the time required for data collection using the Android application was reduced by 50%. We found the Android-based app to be more efficient and time saving as it reduced errors in data collection, eliminated the process of data entry and presented the data for analysis in real time.

Electronic administration of DCE on Android computer tablets to CHWs with basic education is more efficient, time-saving than paper-based survey designs once the application is provided. It is feasible to use technology to develop and implement DCEs among participants with basic education in resource poor settings.