University of North Carolina at Chapel Hill, United States
Project: The HeHealth app is an artificial intelligence (AI)-driven tool to screen for symptomatic sexually transmitted infections (STIs) using a smartphone camera. App users can use their own smartphone cameras to take pictures of their own penises to screen for symptomatic STIs (see Figure 1).
Issue: Monkeypox was declared a public health emergency of international concern (PHEIC) on 23rd July 2022. AI applications have shown promise in the management of pandemics and have been used to assist the identification, classification, and diagnosis of medical images, such as in the case of the last PHEIC, COVID-19. In response to the global outbreak of monkeypox, our team needed to develop a digital screening test for symptomatic monkeypox through AI approaches.
Results: The AI model was initially developed using 5000 cases and use a modified convolutional neural network (CNN) to output prediction scores across visually diagnosable penis pathologies. Of all the STIs, our tool could diagnose Syphilis, Herpes Simplex Virus, Human Papilloma Virus and Genital Viral Warts with accuracy of 86%, 93%, and 96%, respectively. From June 2022 to October 2022, a total of about 22,000 users had downloaded the HeHealth app, and about 21,000 images have been analysed. We then engaged in formative research, stakeholder engagement, rapid consolidation of monkeypox images, a validation study, and implementation from July 2022 to develop the monkeypox module. Since July 2022 to October 2022, we had a total of 1000 number of monkeypox-related images that have been used to train the monkeypox screening tool. Our digital diagnostic tool shows accuracy of 87% to rule in monkeypox and 90% accuracy to rule out the symptomatic infection.
Lessons Learned: Several hurdles identified, which were subsequently mitigated, included issues of data privacy and security for app users, initial lack of data to train the AI tool, and the potential generalizability of input data. We offer several suggestions to help others get started on similar projects in emergency situations, including engaging a wide range of stakeholders, having a multidisciplinary team, prioritizing pragmatism over research elegance, as well as the concept that ‘big data’ in fact is made up of ‘small data’.
