Machine Learning in health insurance without operational disruption. Myth or reality?
October 3, 2019. Key health insurance stakeholders met at 5th MENA Health Insurance Congress to discuss the current state of health insurance in the Middle East and its trends. Netcetera Middle East was an official Silver Sponsor and a Machine Learning expert.
What is the role of Machine Learning for health insurance in the Middle East?
Kiril Milev, Managing Director Middle East, spoke about Machine Learning application in the health insurance fraud detection. While working on data analytics topics in this region, we have seen quite a divergent attitude of organizations towards data. But, mostly, organizations tend to underestimate the value of data and analytics.
What does RiSIC bring to health insurance?
“Over the last two years of making and using ML systems, we have developed our own implementation model”, – Kiril says. It was not a “one shot” job. For starters, we were able to put together the concept of the model only after we realized that detecting waste abuse and fraud (WAF) in claims processing was not a “one time job”, but an ongoing process. Secondly, it was equally important not to overestimate its potential and over promise the results. “When we expanded the model to other use cases, we saw that it would fit in the implementation of similar ML systems, that are not necessarily used for detecting WAF”, – Kiril added.
The model revolves around 3 steps: Assess, Automate and Augment.
Assess, Automate, Augment
When kicking off new implementation, the assessment is a critical step in establishing the base line. Once completed, we immediately identify the population behavior. We can distinguish outliers but we also detect and adjust positive and negative behavior. The second step is to automate patterns that the system accurately detects. In the case of claims processing, we identify immediate wins of claims whose decision can be automated. We were able to identify a set of claim items for denial, with accuracy higher than 90%. Once identified, we managed to successfully deny 60% of them. The 3rd phase, we call it augment, looks at new patterns which are not trivial and not easily detectable in the data.
As a result, for a larger insurer the system increases the denial rate by 3,5% that amounts to 14% savings.
About the event
The 5th MENA Health Insurance Congress, gathered key healthcare decision makers and insurance experts from the region to highlight the efforts made to develop the health insurance market in the MENA region. The Congress covered the impact of healthtech and insurtech on insurance, use of AI & analytics and innovative healthcare solutions for success, cyber security and examining health insurance fraud detection in the Middle East and abuse in the health insurance, exploring the latest trends in the Saudi health insurance market, developments of compulsory health insurance in Oman, corporate and employee wellness, health insurance claims processing in the Middle East.
Interested in exploring 3A model?