Contributed Commentary by Michael Byczkowski, SAP
February 3, 2022 | Is the healthcare industry ready for artificial intelligence and other digital approaches leveraging data? Before we try to answer this question, consider that in China, the AI-driven Ping An Good Doctor online platform already has in excess of 400 million registered users for services like online consultations, referrals, registrations, and online drug purchases, with the support of an in-house medical team of more than 1,800. Equipped with knowledge of 3,000 diseases and continuously being trained with data from hundreds of millions of consultations performed, the AI system “doubles the efficiency of doctor consultations, greatly reduces the possibility of misdiagnoses and missed diagnoses, and improves patients' experience with remote medical consultations,” the company stated.
In Europe, meanwhile, the consulting firm McKinsey portrayed and analyzed in detail 14 active applications for AI in healthcare, “from apps that help patients manage their care themselves, to online symptom checkers and e-triage AI tools, to virtual agents that can carry out tasks in hospitals, to a bionic pancreas to help patients with diabetes.”
Yes, the healthcare industry not only appears ready for AI, some of its most digitally progressive players are already actively embracing it. These early use cases have begun to hint at AI’s potential value to healthcare providers, for its ability to provide an elevated level of data intelligence that could significantly improve patient outcomes and the overall patient experience, and on the business side, for its ability to improve efficiency and innovation. Overall, there are four major areas in which AI looks particularly promising:
- In helping hospitals and healthcare systems manage and make sense of the increasing amount of data they are generating from patient treatments and particularly from connected Medical Internet of Things (MIoT) equipment, leveraging it as a basis for further research and process optimization;
- in patient risk assessments, data acquisitions and tests, diagnoses, therapy decisions, multimodal therapies, and follow-up; and
- in driving highly personalized treatment;
- in helping the clinics as commercial entities becoming intelligent enterprises by optimizing processes, safeguarding operations, and to alleviate healthcare supply-chain disruptions, for example.
To begin to explore these uses cases for AI, healthcare providers first need to ensure they have strong data capabilities in place. When it comes to AI, data is the ultimate asset and resource for a healthcare organization. So, the first step is to build a data infrastructure to manage and curate the vast volumes of data on which AI relies, because without high-quality data, AI can’t be trained properly. Connected to that infrastructure, organizations also will need channels through which data is readily accessible across departments and organizations. This lays the groundwork for eventually using AI to support doctors and nurses in making decisions, supplying then with more relevant background information than they would otherwise have access to.
At its core, healthcare is here to serve and support the patient. AI won’t replace doctors and nurses, but it certainly can augment their capabilities, revealing treatment pathways and other important insights that they can use to produce better outcomes for their patients. With its ability to rapidly and intelligently parse huge amounts of data, making them digestible for human minds and provide insights based on what it derives from this data, AI (and especially Explainable AI, which can outline its decision process beyond just showing statistical correlations used) can help medical teams make decisions based on a holistic, 360-degree view of the patient’s health data from multiple sources, along with data from other patients with similar conditions, symptoms, profiles, genetic predispositions, etc., and how they reacted to certain treatment paths. Expert systems, a specific variant of AI, on the other hand, are based on logical steps and decision trees and can help provide a direct digital implementation of medical guidelines for instant and automatic use at the point of care.
AI also can help the healthcare industry answer growing patient demand for highly personalized treatment and care, giving providers the means to develop and apply individual treatment pathways based on a patient’s unique situation, biomarkers, and additional insights it draws from its training data. In the case of a patient with cancer, for example, it can help script an optimal treatment plan around available options like surgery, radiation, and chemotherapies, based on treatment data from prior patients with similar profiles, while also potentially highlighting new drugs and treatment options by tapping into the latest research and clinical trial data from outside sources. Using AI, pharmaceutical companies, research institutions, hospitals, insurers, tech companies, and others can implement key aspects of the data economy by collaboratively leveraging their data to more efficiently and effectively develop pharmaceuticals, devices, and treatments, also designed to the patient’s individual needs as well as for rare diseases and conditions, right down to a “batch of one.”
What we’re really talking about here is the establishment of digitally connected, healthcare-focused business networks, or ecosystems, in which multiple companies align, use one another’s strengths (including AI capabilities), share data and cooperate to explore and develop new treatments, new pharmaceuticals and devices, and new ways of delivering their services, all with the aim of better serving the patient. China’s Ping An has already enlisted a broad healthcare ecosystem that includes government entities, patients, medical service providers, social and commercial health insurers, and technology. Among its primary goals is “to pursue vertical integration through serving the government, harnessing the resources of hospitals, doctors, and pharmaceutical companies, and empowering members of the ecosystem with technology.”
As valuable as AI can be in improving patient outcomes and experiences, it can also provide substantial benefits to the business side of a healthcare organization. For starters, it can identify ways for optimizing core processes (e.g., fraud detection, demand prediction), equipment deployment to maximize operational efficiency, as well as manage, support, and care for your employees through offering regular evaluations of their personal wellbeing. It can show organizations the best approaches for maintaining reliability, timeliness, and resilience in their supply chains. And through embracing an ecosystem, it can help organizations and their partners to develop additional revenue streams and business models around new medicines and treatments, as well as outcome-based services.
All these potential applications suggest there is a strong case for hospitals and healthcare organizations to integrate AI and other data-intelligence capabilities into their day-to-day operations. In doing so, they will need to ensure they have strong systems and processes in place for ensuring the value and quality of the data on which their AI-based systems rely, and for protecting and governing that data (their own, their patients’, and their business partners’), my colleague, Dr. Feiyu Xu, Senior Vice President and Global Head of AI, said during a recent webcast. They’ll need a clear, strictly followed, compliant and ethical strategy around the application of AI to patient data, according to Dr. Xu.
With those foundational elements in place, healthcare organizations can begin to tap the immense potential artificial intelligence holds for improving patient care and the business of delivering it, ultimately opening the door for all citizens, regardless of background or financial means, to reap the benefits of these important digital innovations as part of the established clinical routine.
Michael Byczkowski is Global Vice President and Head of SAP’s Healthcare Industry. He can be reached at Michael.Byczkowski@SAP.com.