Although the terms health researcher and data scientist both appear in my job title, I am very much a researcher first. I came to Aviro through a master’s in public health, not a computer science degree; I haven’t actually written any code since using C++ in grade 11. Thankfully, most of my day-to-day duties are around the research and analysis of our platforms and projects, but occasionally, I swap hats and help build the product too.

Our decision making is informed by our research and analysis and to guarantee that this is accurate, we must ensure that the data coming through our platform is clean and reliable. To do this, when we are upgrading the product, through new partners or new features, I must make sure that we are properly defining, capturing and testing all the relevant new data points. This is exactly what has been keeping me busy for the last month or so.

‍

Since we launched the new version of the Pocket Clinic in February, we have helped over 10,000 users, identified over 900 HIV positive individuals and linked more than 95% of them to care. This high linkage to care rate is what prompted the demand for two of our new features; TB screening and treatment. TB screening must be done before any HIV positive individual can be started on lifesaving ARVs, and the Pocket Clinic now offers this tool as part of our linkage to care flow. If a user passes their TB screening and starts on ART, healthcare workers are now also able to capture which medication are being prescribed, as well as the option to collect additional medical information about the patient with the treatment module.

‍

‍

Our third new feature, which was the most requested upgrade over the last year, has been the ability to use the Pocket Clinic without internet connectivity. This offline feature has been added to the platform for our two newest partners, TB/HIV Care, which went live with four clinics in Eastern Cape a few weeks ago, followed by Medicines Sans Frontiers in Eswatini, which went live with their first two facilities last week. With these new features and partners, almost 100 new data points have been added to our database, which means that there were nearly 100 new rows that had to be defined in our data dictionary, and almost 100 new variables that had to be added to our daily dashboards for visualisation.

To be honest, adding, defining, and visualising all the new data points was the easy part, because this work was static. The hard part came once everything was built and we had to test everything and to make sure all of the new upgrades work. This is a strenuous process, which first involves user acceptance testing (UAT) on the front end, followed by rigorous quality assurance (QA) on the back end. For UAT, everyone on the team helps, so that we can get as many different eyeballs on the product as possible. We test user stories, where everyone runs through mock journeys on the Pocket Clinic, to ensure that the product allows them to complete the tasks that they were intended to. Once we know the front end works the way we want it, QA ensure that all the data is being captured correctly. After personally conducting hundreds of mock journeys through the Pocket Clinic and ensuring that each data point is coming through the back end and accurately populating our dashboards I can confidently say two things:

‍

1. The Pocket Clinic has passed QA, and is producing reliable data.

2. I am excited to put back on my research hat and start to analyse all these new data points!

‍

‍

‍

‍