Interview

What are the most recent milestones from Comprehense?

We developed a mobile application that uses cutting-edge Artificial Intelligence technology to predict patients’ retention of HIV care after HIV diagnosis and before enrollment. The application uses newly diagnosed patients’ information to predict whether or not a patient will still be retained on HIV care 6 months, 1, 2, 3, 5, and 10 years after Antiretroviral Treatment (ART) initiation.

The models were trained, tested, and validated using historical data from adult patients enrolled in HIV care between 2015 and 2021 and the best models were selected based on accuracy, precision, and AUC.

This application was designed to provide support decision-making to health care providers at all levels (counselors, physicians, psychiatrists) and decision-makers, by identifying patients who will default HIV care before the event (default) happens, allowing them to provide appropriate support or interventions that are tailored to each patient’s needs as opposed the current one-size-fits-all approach.

This innovation will maximize the retention intervention’s effects and contribute to improving retention of care and subsequent improvement of health outcomes.

What are your key milestones to accomplish in the next 12 months? 

We are aiming at deploying the application to a selected number of health facilities for real-life validation.  We will add additional features such as instructions for follow-up depending on the retention status alerts, checking on patients’ status for counseling, patient satisfaction feedback, etc.

After validation for HIV, we will validate applicability to other chronic conditions such as High Blood Pressure (HBP) and Diabetes.

What frustrations did you experience before working with Omdena?

Limited skills and infrastructure for training, testing, and deployment of AI and ML algorithms. Also, there was limited knowledge about the benefits of AI and ML-based projects and benefits among potential adopters, leadership, and users. Funding limitations constitute a significant role as well.

How did working with Omdena’s dedicated AI teams provide better results?

The mobile application now returns predictions in real time. Omdena supported us with the deployment of trained models into the application. They also provided insight in terms of the horizons that can be achieved with new models configurations.

Want to work with us?

If you want to discuss a project or workshop, schedule a demo call with us by visiting: https://form.jotform.com/230053261341340