IBM Research: Using data to fight cervical cancer in Kenya

Dr. Michal Rosen-Zvi

Editor’s note: This article is by by Dr. Michal Rosen-Zvi, senior manager of analytics at IBM Research – Haifa.

Cervical cancer continues to be the second largest threat to women’s health in Africa with mortality rising every year. The incidence of cervical cancer is 53 per 100,000 women in sub-Saharan Africa. This is more than seven times higher than incidence in Western countries. Much of this problem stems from poor awareness, scarce access to timely screening and treatment, and lack of strategic public health infrastructure. Cervical cancer however is largely preventable through vaccination and by implementing screening programs, as well as raising awareness of the disease.

IBM Research wants to make these prevention options available via data. Our scientists in Haifa, Israel and Nairobi, Kenya want to boost awareness, improve monitoring and decision making, and promote a proactive approach to public health with a new system that uses cloud and mobile technologies to gather, manage, analyze and visualize data on cervical cancer in Kenya.

From the point when the infection first begins, it takes years until an invasive cancer develops. Awareness and monitoring through screening programs are key to early detection since the early stages of the cancer can be completely asymptomatic. By identifying and removing pre-cancerous cells during this time frame, the cancer can be prevented.

Dynamic disease mode

The IBM team has developed a model for cervical cancer incidence, based on well-established epidemiological models incorporated into a state-of-the art Bayesian model. It combines factors such as living conditions – whether rural or urban – with HIV co-infection, intervention actions, and disease progression. As more real-time data collected in the field is fed into the model, the parameters are automatically refined to increase the accuracy of future predictions. In effect, the system 'learns' from data over time.

By merging our expertise in big data analytics with smarter healthcare, IBM can help transform information into something that will help reduce the unnecessary suffering caused by cervical cancer, and improve the quality of life for millions of women across Africa.

Our analytics engine uses the new disease model to sift through the vast amount of data collected. This allows researchers to create “what-if” scenarios to determine the best outcome of where to invest efforts and budget. For example, if all girls who attend primary-school are inoculated at age 12 from now on, what would the number of incidents of cervical cancer be 20 years from today? And what would this number be if no such action is taken?

Or what if we do screening and treatments for all the women in a particular region, according to recommended Kenya guidelines? How many of the disease incidents could be prevented? The cognitive system provides estimate of incidents for the next year, two years from now, and five years from now. The system includes the statistics for Kenya’s population, combined with information extracted from literature and model analysis modules.

IBM, in cooperation with global and local agencies such as the Bush Institute, PEPFAR, PinkRibbbon, RedRibbon, Future Group, and Strathmore University, has embarked on a journey to help monitor, treat, and prevent cervical cancer in Africa. This project is being conducted by IBMers in Kenya, Israel, China and other places around the world.