Artificial Intelligence in Healthcare: 2019 Industry Trends
Artificial intelligence (AI) is now woven into many facets of our everyday lives. In many instances, it’s so firmly ingrained that you’ve likely forgotten your initial sense of wonder at its ability to give you what you need, or want, quickly, accurately, and inexpensively. For example, via Siri, Alexa, Google Maps, and Amazon shopping.
But in healthcare, most of us are still in the “amazed phase” of how AI is impacting our industry. AI is revolutionizing how healthcare is practiced, delivered, billed, and paid for, and the transformation is happening quickly. When I reviewed healthcare industry trends in 2019, artificial intelligence was front and center. Here are my top five picks for how AI impacted our industry last year.
Artificial Intelligence in Healthcare
#1: Medical Research
At the beginning of 2019, my friends conducting academic medical research were a bit skeptical about AI. But during the year, we made a lot of progress applying it to medical research. AI is igniting and fueling research into the early detection of disease, boosting diagnostic capabilities to enable improved treatment and care outcomes. This is generating a lot of excitement as advancements are being made in areas where our human efforts have been thwarted or stalled; for example, in the early detection of Alzheimer’s disease (Artificial Intelligence Can Detect Alzheimer’s Disease in Brain Scans Six Years Before Diagnosis), and in bringing parity to the early diagnosis of breast cancer (MIT AI Tool Can Predict Breast Cancer Up to 5 Years Early; Works Equally Well for White and Black Patients).
These are but two of the many AI-related research announcements of 2019, and early detection is just one area benefiting from AI’s capabilities. We’ll continue to see breakthroughs and “leaps-and-bounds” progress in research as AI is applied both in the laboratory and to human studies.
Of course, AI is still just a tool. However, if the adoption of this tool immediately gives us a double-digit increase in accuracy over previously available methods, then this is a tool that cannot be ignored any longer.
#2: Natural Language Processing
During 2019, the greatest AI progress was made in the area of NLP. Natural language processing (NLP) is a form of AI that helps computers understand humans’ natural language. Needless to say, when computers can “interpret” human language, translate/convert it, and accurately apply it to the project at hand—the speed and efficiency of a process can be increased exponentially. Some of the latest improvements in NLP came from horizontal technology vendors. For example, Google developed the popular BERT model to increase NLP accuracy. At the same time, AWS launched Amazon Comprehend Medical to extract medical information from unstructured text.
NLP solutions specific to healthcare are beginning to have a significant impact on many processes, such as clinical documentation, automated coding of claims, prior authorization, clinical decision support, risk adjustment, compliance, and clinical-trial matching.
#3: AI Experimentation
Many healthcare providers began experimenting with weaving AI into their workflows last year. Common examples included using AI to predict length of stay and determine probability of readmission. Most providers find that it is possible to develop AI models with pretty high accuracy, but there is a nontrivial need to improve the way data is collected and processed across the organization in order to build a model that will prove valuable. This experience is often invaluable to the provider organization. However, AI’s accuracy is driven by the depth and breadth of data used to train the model; that’s why many providers purchase the solutions of healthcare IT vendors that have access to massive volumes of data.
#4: Government Interest in Applying AI
The U.S. government got very serious about AI during 2019. President Donald Trump issued an Executive Order launching the American AI Initiative on February 11, indicating the government will follow the lead of other industries in leveraging AI to detect fraud, waste, and abuse in healthcare. Improper payments accounted for $31 billion of Medicare’s net cost in 2018, and CMS plans to pursue a proactive, preventive approach to address the problem. (The traditional approach of relying on humans to sift through massive volumes of payments, identify irregularities, and rescind improper payments has been inefficient and ineffective.) AI’s ability to identify patterns of fraud, waste, and abuse via machine learning could exponentially improve detection, saving the government billions of dollars annually. Many commercial payers are following suit, embracing AI’s ability to facilitate payment accuracy and integrity to drive cost savings.
#5 Provider Revenue Cycle
Payers weren’t the only ones to discover how AI could streamline administrative processes and improve financial outcomes. In 2019, providers also began to incorporate AI technology to boost efficiency and the bottom line. From optimizing workflows, to proactive and enhanced charge capture, to flagging claims at risk for denial before they are filed, AI-infused solutions drove a lot of excitement in the provider arena.
New Year, New Opportunities
As we move into 2020, artificial intelligence will continue to enable us to make amazing strides in improving the overall efficiency of the U.S. healthcare system—including reducing administrative waste—and that will benefit patients as much as it will providers and payers. We can eagerly look forward to what AI surprises are unveiled this year.