JMIR Publications

ABSTRACT

Background:

In Uganda, cervical cancer (CaCx) is the leading cancer, accounting for 35.7% of all cancer cases in females and 20.5% of all cancer cases overall. While CaCx is preventable, human papillomavirus vaccination rates in Uganda are <25% and CaCx screening is <5%. Digital health tools have been implemented in several African countries to support and enhance CaCx screening, but such implementations in Uganda are lacking.

Objective:

To describe the implementation and experiences of a telemedicine system for improving CaCx screening in Uganda.

Methods:

The study was conducted between 2019 and 2021 at the Uganda Cancer Institute in Kampala and at its four satellite clinics in Arua, Mayuge, Mbarara, and Jinja districts. We implemented a mobile phone-based telemedicine system to support data capture and sharing for remote consultation, monitoring, and development of artificial intelligence (AI) models for automatic analysis of cervicographs. We conducted focus group discussions with staff at the screening clinics (n=27) and women undergoing screening (n=15) about their experiences and perceptions of the system, as well as taking field observations of the screening process and utilization of the telemedicine system.

Results:

The system consisted of the Gynocular portable colposcope coupled with a Samsung smartphone with the open-source data capture app Open Data Kit (ODK), and a centralized ODK Aggregate server which the consultant gynecologists would log in, view the data dashboard, and give a second opinion when requested. We also made a custom mobile app into which the AI model was deployed for testing. We screened 2682 women, with VIA positivity of 12%. Qualitative findings show that the patients and health workers were positive about the system, highlighting the following benefits: better visualization on the smartphone screen, ability to get a second opinion, and ease of communication among the nurses to discuss cases, and with patients to explain the findings or treatment by showing them the pictures of their cervix. Other perceived benefits included improved clinical data management, supporting performance monitoring and feedback, and “modernization” of screening and thus increased trust in the otherwise simple VIA procedure. We also observed that the time needed to train the nurses on the system was short, about 2-3 days, but ongoing supervision was needed to ensure the VIA assessments were correct. Only a few patients expressed concern about the privacy of their information, but these concerns were easily allayed with explanations of the data collection and handling.

Conclusions:

This study demonstrates the feasibility and value of digital health tools in cervical cancer screening in Uganda, especially with regards to improving patient experience, and the quality of screening services.