Medicare fraud is quite, unfortunately, an ongoing epidemic according to a recent study. Machine learning has recently become a very useful tool in rooting out a variety of Medicare fraud that has been occurring, however.
In 2018, the total Medicare program cost $583 billion,, about 14 percent of total federal government spending. It’s estimated that Medicare fraud is responsible for almost $65 billion in losses each year. With AI going through a wide range of cases it could be possible to prevent some of these effects from happening. According to researchers at Florida Atlantic University, it may be possible to use machine learning to identify instances of fraud effectively.
The university tested six different types of machine learning algorithms against balance data sets, and it was possible for each one of these algorithms to determine possible instances of fraud for further investigation. What this could mean is the chance for researchers and adjudicators to have an ally on their side when it comes to going through case files.
There are so many various intricacies in determining what constitutes fraud, especially in clerical errors. Ph.D. holders and students are regularly responsible for tracking medical services and billing which can be commonly neglected when it comes to the continuous Medicare part B data.
AI can track a vast number of variables including instances of data sets for fraudulent providers, fraudulent codes and more. Organizing fraudulent providers in a separate database and regularly logging case files that could be seen as fraudulent can be very important to manage data sets appropriately.
As most studies would suggest, a small number of fraudulent providers tend to continue racking up large numbers of fraud bills. The investigative power that is often overlooked when it comes to rooting out some of these mistakes. Having machine learning on the side of physicians could help to make sure that it could be possible to easily root out the cause of possible fraud and continue further investigation is required.
Fraud in medical data doesn’t have to be wholly negligent or even easy to spot either. Most providers that were studied throughout the machine learning process discovered a sweet spot where only 10% of the data was fraudulent and enough to add up an extra amount of Medicare bills. With around 10% of fraudulent data, most practitioners would simply void the mistake for its low cost. When AI learning is introduced, however, it is possible that some of the smaller mistakes can be identified. Identifying some of the smaller fraud cases can help to prevent an ongoing snowball effect within the Medicare system.
Machine learning detection tools are not currently in place for checking into fraud cases throughout the Medicare system. A Dean at the college of engineering, however, has said that machine learning to Texan tools could quickly become a game changer for fraud detection within Medicare for the future.
Asif Razzaq is an AI Journalist and Cofounder of Marktechpost, LLC. He is a visionary, entrepreneur and engineer who aspires to use the power of Artificial Intelligence for good.
Asif's latest venture is the development of an Artificial Intelligence Media Platform (Marktechpost) that will revolutionize how people can find relevant news related to Artificial Intelligence, Data Science and Machine Learning.
Asif was featured by Onalytica in it’s ‘Who’s Who in AI? (Influential Voices & Brands)’ as one of the 'Influential Journalists in AI' (https://onalytica.com/wp-content/uploads/2021/09/Whos-Who-In-AI.pdf). His interview was also featured by Onalytica (https://onalytica.com/blog/posts/interview-with-asif-razzaq/).