By Bonner R. Cohen, Heartland Institute
“Medical knowledge is growing so rapidly that only 6 percent of what the average new physician is taught in medical school today will be relevant in ten years,” states a new National Bureau of Economic Research (NBER) report, “The Potential Impact of Artificial Intelligence on Health Care Costs.”
As AI transforms one sector after another, its “adoption within the next five years using today’s technologies could result in savings of 5 to 10 percent of health care spending,” the report finds. The projected savings would come from AI-enabled improvements in the management of hospitals, physician groups, and private payers.
AI in Practice
The authors, Nikhil R. Sahni, George Stein, Rodney Zemmel, and David Cutler define AI as “a machine or computing platform capable of making intelligent decisions.”
The report states there are two types of AI the health care industry is pursuing, “machine learning (ML) which involves computational techniques that learn from examples rather than operating from predefined rules.
The second AI type is natural language processing (NLP), “which is a computer’s ability to understand human language and transform unstructured text into machine-readable structured data.”
In practical terms, the report says, these technologies can be readily applied to health care. “ML examples include whether a patient is likely to be readmitted to a hospital, using remote patient monitoring to predict whether a patient’s condition is likely to deteriorate, optimizing clinician staffing levels in a hospital to match patient demand, and assisting in interpreting imaging and scans.”
NPL examples include, “extracting words from clinician notes to complete a chart or assign codes; translating a clinician’s spoken words into notes; filling the role of a virtual assistant to communicate with a patient, help them check their symptoms, and direct them to the right channel such as a telemedicine visit or a phone call; and analyzing calls to route members to the right resource and to identify the most common call inquires.”
Health Care: A Slow AI Adopter
Health care has lagged other industries in the adoption of AI, write the authors. AI, they write, follows an “S-curve,” which comprises “first developing solutions, then piloting, followed by scaling and adapting, and finally reaching maturity.”
Financial service companies are at the final stage of the “S-curve” as they are capable of deploying “sophisticated AI algorithms for fraud detection, credit assessments, and customer acquisitions,” write the authors.
Health care payment systems may not provide the incentives for such AI application, write the authors. “Another view is that management barriers, both at the organizational and industry level, are responsible for slower adoption in health care.”
One challenge for the health care industry is “data heterogeneity,” diversity and variability within a data set.
“In industries with greater AI adoption, most data are structured,” write the authors. “In health care, by contrast, large portions of key data are unstructured, existing in electronic health records. Clinical notes, the clinician’s recording of a patient’s response to a particular treatment, are one example. Further, these data exist in multiple sources, often with limited ways of connecting disparate pieces of information of an individual patient.”
Patient Confidence
Would patients trust diagnosis and treatment recommendations that are generated by AI, and would they be willing to submit private information to a machine?
“Patients may worry about how their data are being used and prevent the application of AI from being used for their medical needs,” write the authors.
As for treatment recommendations, there is a question of whether AI data can be trusted.
“There are many examples of bias in algorithms, and patients may not trust AI-generated information even if a clinician validates it,” write the authors. “There are also methodological concerns such as validation and communication of uncertainty, as well as reporting difficulties such as explanations of assumptions.”
AI has the potential to help doctors and patients and is being used successfully in one area, says Peter A. McCullough, M.D., MPH, president of the McCullough Foundation, who has published many studies.
“AI could transform medicine as it develops abilities to accomplish tasks such as retrieving medical records from disparate sources, assembling them, creating a timeline, and then presenting a synopsis to the doctor,” McCullough told Health Care News. “AI is far away from these functionalities, which will be dependent on systems integration, privacy assurances, etc. In the meantime, AI is performing well as an advanced search engine.”
Bonner Russell Cohen, Ph.D., (bonnercohen@comcast.net) is a senior policy analyst with the Committee for a Constructive Tomorrow (CFACT).
