As if the US healthcare system weren’t complex enough these days, fraud is also on the rise. The National Health Care Anti-Fraud Association (NHCAA) estimates that the financial losses due to healthcare fraud are in the tens of billions yearly. This may account for up to ten percent of US healthcare spending, says HealthLeaders Media. That’s why it is crucial that healthcare providers take steps to prevent fraud and malpractice. In an earlier post, we discussed how artificial intelligence (AI) will help healthcare institutions provide value-based care. In this post, we will discuss AI solutions in healthcare and tackle the question of fraud protection by leveraging AI technology.
The cost of fraud
Fraud and counterfeits critically impact healthcare companies’ costs of operation, liability, and brand image. These extra financial and quality-of-life costs detract time and attention from more critical aspects of healthcare. This is even more so when related to products and services; it could negatively impact patients’ physical health.
Common examples of fraud and abuse in healthcare include the following:
- Treatment is paid for but not provided
- Falsified claims
- Multiple claims by different providers for the same patient/treatment
- Stolen patient identities are used for reimbursing medical services never provided
- Multiple treatments and diagnostics requested but not required
How can machine learning help?
The classical way of dealing with these behaviors requires highly trained people with acute audit senses. They need to “catch” patterns in distributed, incomplete data. Predictive analytics and big data systems can help prevent fraud. These technologies allow us to identify inaccurate claims, unwarranted procedures, occluded duplications, etc.. Using machine learning, we can begin to leverage these professionals’ expertise and scale up. Analyzing historical claims, especially when some prove fraudulent, can teach us. We learn what a normal claim looks like, when abnormal claims happen, what conditions they have in common, and plausibly how early we can tell the likelihood of a fraudulent claim.
Furthermore, we can assign claims and providers a “trustworthiness score” to avoid costly manual checks. We can also flag low-scoring claims or providers, anomalies, and suspicious billing patterns for detailed revision. Daily, field agents are constantly making repeated decisions with limited information. Machine learning strikes gold precisely in those two areas. It leaves “long tail” decisions to the algorithm, including previously impossible-to-process sources like social media, news articles, etc.
Some AI solutions in healthcare involving machine learning —to prevent fraud and reduce costs—are:
- Indicating whether a given claim is a fake
- Predicting the total cost of following through with a fake claim
- Finding common factors missing in fraud cases in order to devise a new policy
- Predictive filtering of the claim’s validity upon filing to avoid wasting time on suspicious ones
- Including an interactive helper when filing a claim to minimize misfiling
- Revealing many large-scale trends that are otherwise invisible to local teams because, on the individual level, the signals are soft or undetectable
An example of what AI can do
Initiatives involving AI, big data analytics, and smart packaging can be game changers for protecting against healthcare fraud. In terms of consumer items, smart packaging is a great option. It allows field agents to check the authenticity of products by scanning certain details and markings on the package. In this way, they can ensure the authenticity of the products and minimize the chances of product tampering. Also, this technology allows for tracking their origin. This helps with compliance, product recall, information for end users, and quality management.
The digital journey ahead
As far as artificial intelligence technology and healthcare are concerned, many areas remain left to explore. Leveraging AI and big data, healthcare providers can take fraud protection to a new level. And if done successfully, this reduces fraud losses and improves the brand image and overall user experience.
Check out Globant’s Data & AI Studio to learn how we can help you mitigate fraud and implement AI solutions in healthcare.
This post is the third in a healthcare series we are working on. Look here for the second post.