How deep learning helps match the right patient with the right healthcare | Artificial intelligence
Most healthcare systems and insurance providers know who the high-risk patients are: Usually people living with chronic conditions—diabetes, asthma, heart disease—who aren’t managing their health well enough. These individuals are at a high risk of ending up in the ER. There are several problems with this high-risk list approach to caring for patients.
First, these risk scores are often compiled using insurance claims. This is a limited data source, and it is all about the past, not the present. This is not the kind of decision support that prevents readmissions or gets the right care to the right person at the right time.
Second, insurers and physician practices often use phone calls to check in with these individuals. This method is as inefficient as it sounds.
Finally, this approach doesn’t provide any clue as to what is preventing an individual from being healthy. The healthcare providers know who the sickest people are, but they don’t always know what the person needs to stay healthy. The missing element could be transportation to doctor visits, medication reconciliation, financial support to buy food or prescriptions, or even more frequent doctor visits.
The promise of artificial intelligence in healthcare is to help doctors and nurses—and even insurance companies—match the right care at the right time to the right person. That change—mass personalization in healthcare—is the promise of the specialized version of AI called deep learning.
SEE: IT leader’s guide to deep learning (Tech Pro Research)
Deep learning is branch of machine learning. Deep learning systems are modeled on the human brain. These artificial neural networks learn by passing data through layers of algorithms. Training data is fed into the bottom layer. Each node in the layer assigns a value to a data point. If the value passes a certain threshold, the data moves on to the next layer, until it arrives at the output layer. During training, these thresholds are adjusted until analysis of similar data sets yield similar outputs.
MIT is using this technology to power to simulate a clinical trial to determine the lowest possible dose of chemo for people with brain cancer. They are also working on a model that could suggest treatments for sepsis.
Health tech companies are using deep learning to, for instance, predict which person will develop pressure sores during a hospital stay or which heart attack patient will be back in the hospital within a week. Doctors need deep learning tools to compile data from multiple sources, look for patterns, and rate risk at the patient—not the population—level.
SEE: Turning big data into business insights (ZDNet special report) | Download the report as a PDF (TechRepublic)
Jvion and Cyft are two companies developing technology platforms to help doctors make treatment decisions informed by a much wider data set and ultimately help patients get the most appropriate care.
How Jvion is making hospitals healthier
Jvion describes its technology as a “cognitive clinical success machine” built with an Eigen Spheres engine. The engine is an n-dimensional space upon which millions of patients are mapped against tens of thousands of Eigen Spheres. This is a mathematical way of talking about how complex health is and how many factors influence it. A doctor could do everything right in the hospital setting only to have other factors outside the hospital sabotage a person’s health.
What does Jvion do?
Jvion’s uses its Eigen Sphere engine to combine many sources of data about a patient—including clinical, socioeconomic, and behavioral—and to consider many possible outcomes. This analysis creates an individual risk score based on data from the individual patient, as opposed to a more general score based on broad demographics like age and medical condition. This analysis considers whether a person is moving toward or away from health problems.
For example, older people can develop new health problems after a hospital stay—issues unrelated to the original illness that sent them to the hospital in the first place. Jvion’s predictive technology could help identify people at the highest risk for this “post-hospital syndrome” and prevent the associated downward spiral for older patients. Jvion also works with hospitals to prevent healthcare-associated infections in America, a problem that affects 5 to 10% of hospitalized patients in the US per year. These infections have resulted in about 99,000 deaths and an estimated $20 billion in healthcare costs in America.
How does Jvion help?
Healthcare systems now have a financial incentive to reduce hospitalizations due to provisions in the Affordable Care Act, a law passed in 2010 in America that improved access to healthcare. Hospitals face a financial penalty if Medicare patients with certain conditions return to the hospital within 30 days. A study published in 2017 predicted that these Medicare penalties for American hospitals would be $528 million in 2017, $108 million more than in 2016. The benefit to patients is obvious: Personalized care that fits an individual’s particular situation.
Who is Jvion’s target customer?
Hospitals that serve a lot of Medicare patients or that are working in value-based contracts, instead of the traditional fee-for-service model. These contracts link payments with better outcomes for patients instead of the volume of care provided. Medicare is leading the way with the transition, but some healthcare systems are moving to this system as well. Healthcare leaders need new tools and decision-making processes to make this huge shift away from the traditional healthcare business model.
SEE: AI and the NHS: How artificial intelligence will change everything for patients and doctors (ZDNet)
How Cyft uses data analytics to personalize patient treatment plans
While Jvion is working inside the hospital, Cyft is focused on the world outside the doctor’s office, which is where most of our healthcare challenges exist. Research on the American population suggests that direct medical care represents only about 20% of the factors that influence a person’s health. Socioeconomic factors such as education, income, and family support have the biggest impact at 40%; healthy behaviors are next at 30%; and the environment is the smallest impact at 10%.
This focus on home life, economic status, and a person’s overall environment is a significant shift for healthcare leaders. Many population health startups have sprung up over the last few years to help health systems address problems such as food insecurity, substance abuse, and homelessness. As more health systems move to pay-for-performance payment models instead of fee-for-service, healthcare leaders have to figure out how to address these “outside the doctor’s office” influences.
Leonard D’Avolio, Ph.D., the co-founder and CEO of Cyft, thinks that healthcare systems have all the data they need to do this, but they just don’t know how to use the information effectively. The other key is to analyze the most relevant data. Cyft’s promise is to make sense—and predictions—from small sets of messy data. Cyft’s technology analyzes information from multiple sources to determine what kind of care a patient needs as well as the likelihood that a particular solution will work.
What does Cyft do?
Cyft (as in sifting through stacks of information) helps a healthcare provider figure out which patients would most benefit from a particular intervention; this could be a phone call, an office visit, a change in medication, fall prevention, or mental health care. The Cyft software pulls in data from separate sources—patient surveys, health assessments, EHRs, doctors’ notes, call center transcripts—and identifies the most relevant risk factors. This assessment provides a priority list as well as personalized recommendations for patients. Two other examples of Cyft’s work include identifying patients who are likely to have longer hospital stays after surgery and analyzing therapy notes for 300 American veterans with PTSD to determine whether they were receiving “best practice” care.
How does Cyft help?
Many times, a care provider has to guess as to what a patient needs or which person needs immediate attention. Cyft’s analysis can recommend treatment plans that fit the individual. Cyft can make predictions for all patients in a health system, including healthy people with no prior hospitalizations.
Who is Cyft’s target customer?
Health systems that are “at-risk” or ” value-based.” When healthcare providers take on “risk,” their payment for services can lowered if a patient picks up a hospital-acquired infection or if her conditions worsens. The goal is to help a person become—and stay—healthy. Cyft also works with insurance companies to improve member retention and with government-sponsored health plans to improve reimbursement rates