Cobalt / Prime Study

In May 2016, the Knowledge Translation Unit (KTU) engaged with us to work on a data collection system for a set of questionnaires they needed to collect data with. They had already developed an Android mobile application with one of South Africa’s largest mobile development companies, which was used in the first round of data collection; however, the system was plagued with issues, both in terms of configuring the out-of-the-box tool to suit a questionnaire with complex validation and skip-logic, as well as the reliability of the data being received from the devices.

Often companies will build mobile apps when websites will work just as well, if not better. As the questionnaire had to be run on tablets in rural areas with limited or no connectivity, with the data being uploaded when a connection was available, this seems to favour development using an Android app. However, we took the approach of building a single-page app (SPA) with AngularJS, which was mobile-friendly (using Bootstrap), which would be downloaded using the cache manifest, and therefore would run offline. Data would be stored in the browser using PouchDb, a lite-weight Javascript based database that can also store data offline. When a connection became available, the researchers would upload their data via an Api to the centralised web-based database. A separate management tool let head office access and download the data in CSV format, which was them cleaned and imported into STATA, for statistical analysis.

It took under two months to develop and test the initial questionnaire, including numerous changes to the questionnaire itself, and the system was live by the end of July 2016. Four additional questionnaires were subsequently added, plus a tool for scheduling follow-up appointments.

The data collection for KTU was, to our knowledge, a breeze compared to their previous experience, and KTU have subsequently become a regular client.