Design, deliver, and run enterprise blockchain workloads quickly and easily.
The robot will see you now
While everyone in the U.S. is focused on health insurance, the truth remains that there are not enough primary care doctors to tend to everyone. Regulation (or deregulation) alone can’t fix that shortage. In order to provide prevention care, illness and injury treatments, and life-saving measures for all, technology must fill in and assist human healthcare providers.
But this is not a situation unique to the U.S., as all countries around the globe struggle with a shortage of healthcare professionals and resources in varying degrees. Further, healthcare providers also struggle with work overloads, unacceptable error rates, and lost revenue from complexities and errors in payer claims.
Healthcare workers sometimes see technology as an impediment to patient care, but this attitude is likely to change in the future. Meet the soon-to-arrive ambient digital assistant for healthcare providers. You can think of it as a child of artificial intelligence (AI) and modern analytics and a cousin to the virtual assistants in the consumer space.
When will such a device be birthed onto the medical scene? Within five years, predicts Darren Schulte, a physician by training and currently CEO of Apixio, a data science company focused exclusively on developing an AI platform for healthcare.
Less typing, more treatment
For a machine to fill the role of physician assistant perfectly, it must always be present and alert, and often sterile as well. An ambient device fits that order perfectly. The capability to process vocalized natural language eliminates the need for human touch and thus the spread of germs.
Workflows are also simplified because the multiple workflow processes that are typical with mobile devices—such as opening software, physical task initiation, and manual text input—are no longer necessary. Even when software is integrated, using it on traditional mobile devices still requires multiple steps to get to the information needed or add patient information. It's important to minimize those steps and processes given perennial complaints that entering and working with patient data takes too much time away from working with the actual patient.
Further, next-generation ambient assistants will be able to ingest inputs from a variety of sensors, including their own visual (cameras) and audio-based sensors, and from IoT medical devices either on the patient or in the medical environment. The rapid acceptance of devices such as smart watches that run health monitoring software and fitness-specific wearables have prepared patients to interact with these IoT tools.
The monitoring capabilities of future health-focused personal devices with similar capabilities could enable ambient digital assistants to detect even subtle changes in the patient’s condition, facilitating everything from triage to diagnosis to treatment management.
“For example, keeping patients under sedation is automated by machines that are constantly monitoring a patient’s vital signs and titrating medications accordingly,” says Schulte. “Intervention is on auto-pilot.”
An ambient assistant could quickly and accurately inform physicians of the status of such automated care. It could also remind physicians of details ranging from past drug reactions to patient concerns and requests. These capabilities can significantly improve patient safety and potentially reduce malpractice claims and lawsuits.
Consumer virtual assistants such as Siri, Alexa, Cortana, and Google Assistant already use camera and audio input processing capabilities, largely so that the assistants can “listen” for commands and execute them immediately. Medical-grade ambient digital assistants will use data from similar sensors for a broader range of functionalities and data analyses.
The business case for digital assistants
Healthcare is costly. Providers must make a profit in order to stay in business and treat patients.
Ambient digital assistants can help solve a wide range of business problems, ranging from excessive overhead to costly errors in collecting payment for services rendered. For example, the wrong ICD-10 code on a prescription or treatment order requiring payer pre-approval or on a claim filed with the payer afterwards can result in a denial of payment or slow payment. Neither is good news for the healthcare provider’s cash flow.
Today’s healthcare analytics tend to focus on either medical or financial issues rather than unifying both aspects of patient care delivery. This creates yet more data silos and repetitive data inputs, further driving up costs and impeding patient-doctor interactions. Yet combining the two is challenging.
“Access to medical data for analysis is usually restricted by regulations worldwide," explains Dr. Irene Kopaliani, managing director of CxT Group, a U.S.-based software development and integration company. "Financial data within healthcare is easier to analyze because it doesn’t have many of these barriers. In fact, insurance companies, both large and small, already use big data tools to segregate and analyze data to highlight various trends and tendencies. It is also used to help locate fraud.”
Bottom line: Providing more sophisticated analytics for providers that continue to segregate financial data from patient data will not improve patient care or practice bottom lines.
The present and future of analytics
Many data analytics products on the market today consist of descriptive, prescriptive, or predictive analytics. Few package more than one, although some do.
Descriptive analytics describe what is (real time) or what has been (historical). Predictive analytics predict what will be if nothing about the current situation changes.
Prescriptive analytics prescribe an action. In retail, prescriptive analytics may automatically print coupons for a customer at checkout, for example. In streaming entertainment services like Netflix, prescriptive analytics will suggest movies the viewer is likely to enjoy. In healthcare, prescriptive analytics can not only predict hospital re-admissions, but also suggest actions likely to reduce the chances that the patient will need additional hospital care.
“The bottom line is that we have already started to reap the benefits of big data analytics for healthcare. The results are saving lives, making treatment more effective, and lowering healthcare costs,” says Alfred Poor, editor of Health Tech Insider.
Next-gen applications provide anticipatory analytics. This class of analytics anticipates what action is needed next, which goes beyond predictive analytics in that it not only predicts what event or need happens next, but also anticipates the actions that must be taken now or in the future due to that event.
In healthcare, anticipatory analytics boils down to the machine knowing what you need before you do. In hospitals, anticipatory analytics can compute the need for specific supplies and automatically order them from suppliers so that they arrive “just in time” to fill the hospital’s need for them. This helps control costs because it eliminates over-ordering or last-minute, higher cost purchases.
Anticipatory analytics can also be used in treatment planning so that patients receive appointments, prescriptions, home health services, or automated hospital, hospice, rehab, or nursing home bed reservations and admissions, as needed.
Ambient digital assistants can also anticipate workloads and patient flows by analyzing data and “reading the room”—for example, the doctor’s office waiting room. The machine can then make or change patient appointments as needed to leverage resources, prevent over-tiring the healthcare provider, and keep patient waiting times low.
In short, today’s analytics can handle many tasks in healthcare. The challenges aren’t with the machine, but with the humans using the machine.
“There is a great need for better analytics but a serious shortage of people who can do it,” says Bernard Munos, senior fellow at FasterCures, a research division of the Milken Institute. “Another problem is the need to connect a large number of databases that were never designed to be interfaced—for example, patient records from different systems, PubMed, genomic databases, and patents.”
An AI-driven, ambient digital assistant is the ultimate user-friendly interface for providers. It eliminates the need for data science skills by doing all that work itself and learning while it works. This technology seems likely to become a physician’s most trusted assistant, and an invaluable boost to effective and efficient healthcare.
“This will result in a total reset in how we view healthcare," Poor says. "We will be able to transition to a much more predictive system based on maintaining health, instead of our current reactive system of combating disease.”
The future of healthcare: Lessons for leaders
- AI and IoT will improve patient care and quality of life.
- Wearable devices will go beyond today's fitness bands to personal health monitoring.
- Big data analytics will allow for devices to recognize symptoms at an earlier point.
This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.