Accelerating leprosy elimination
Novartis and the Novartis Foundation have been working with international partners for over three decades to help eliminate leprosy. Since 2000, we have donated more than 65 million blister packs of multidrug therapy (MDT) through the WHO, helping to treat more than 7.3 million patients worldwide.
Beyond treatment, we have pioneered innovative approaches to fight leprosy. One of these is the Leprosy Post-Exposure Prophylaxis (LPEP) program, which aims to decrease leprosy transmission by providing preventive treatment to close contacts of newly diagnosed patients. The evidence generated by LPEP in the seven countries where the program was operational from 2014-2018 led to the inclusion of this strategy in the WHO Guidelines for the Diagnosis, Treatment and Prevention of Leprosy. Since 2014, thanks to this approach, more than 170 000 contacts of patients have been traced, and more than 150 000 have received preventive treatment. Our epidemiological modeling showed that large-scale implementation of this strategy could massively reduce the global burden of leprosy. The results of the five-year LPEP program were published in The Lancet Global Health in October.
We continue to leverage artificial intelligence (AI), data and digital technologies to accelerate leprosy elimination efforts. In January 2020, Microsoft selected the Novartis Foundation as one of its four initial partners for its new AI for Health initiative, a five-year program to scale up global health initiatives using the power of AI. This will provide the Novartis Foundation with access to the latest technology, resources and technical expertise to help further embed AI in our leprosy work.
Microsoft and the Novartis Foundation are also collaborating with local investigators from the Oswaldo Cruz Foundation (Fiocruz) in Brazil. The group developed a protocol to collect, examine and process anonymized leprosy skin lesion images paired with information about the patients’ leprosy symptoms. This data was used to train an AI model – called AI4Leprosy – to accelerate the diagnosis of leprosy through early screening. The protocol and the results of the first model will be published in 2021.
While the algorithm is currently being improved with additional skin images and data from India, a global protocol has been developed together with the University of Basel in Switzerland, and endorsed by international leprosy experts to help further improve the AI model. This should enable data collection to be extended to new geographies, minimizing bias and improving the accuracy of the prediction. AI4 Leprosy has the potential to be easily integrated at no cost in various tools supporting healthcare practitioners in the field or in clinics.