We are excited to bring Transform 2022 back in-person July 19 and virtually July 20 - 28. Join AI and data leaders for insightful talks and exciting networking opportunities. Register today!


The means to predict mortality using artificial intelligence could be a transformative factor in the future of palliative health care. While this topic may seem a bit morbid, AI has the potential to help medical care providers and doctors significantly improve the delivery of patient care in hospice situations.

Getting the right kind of treatment at the end-of-life stage is more important than many assume. Not enough treatment — or even inaccurate treatment — can provide a painful experience for patients, and overcare may result in hundreds of thousands of dollars in unnecessary medical bills, even if the patient is covered by insurance. While it’s crucial to select the proper medical coverage that includes hospice care regardless of the situation — especially for people over 65 or older, because there are specific plans for specific purposes to help with these medical costs — AI advances may help patients and physicians determine illness sooner to prepare for end-of-life costs and treatments before it’s too late.

A recent study in the journal NPJ Digital Medicine shows that technology will soon allow physicians to improve the timing and delivery of patient care. Researchers used AI to scan electronic health records (EHR) and notes doctors left in patient records to detect potential clinical problems and health risks. The AI system predicted patient mortality rate and final diagnoses more accurately and quickly than physicians. So how does it work?

Using deep learning for patient insight

In the NPJ study, researchers fed almost 48 billion data points (including doctors’ patient notes, patient demographics, procedures, medications, lab results, and vital signs) into a deep learning model. This model analyzed the data and predicted, with 90 percent accuracy, medical issues like mortality rates, longer hospital stay lengths, unplanned readmissions, and patients’ final diagnoses. When compared to traditional predictive models, the deep learning model was more accurate and scalable.

For example, a woman in the final stages of breast cancer came to a city hospital with fluid already in her lungs. Two doctors reviewed her case, and she received a radiology scan. The hospital’s traditional predictive model reviewed her chart and estimated there was a 9.3 percent chance she would die in-hospital. A new type of algorithm (created by Google) reviewed the woman’s chart — about 175,639 data points — and estimated her death risk at an actual 19.9 percent. The patient passed away in a matter of days, proving the algorithm model to be more accurate.

Compared to the traditional method, the deep learning model was 10 percent more accurate. The system’s ability to sift through data that was previously unavailable helped it provide a more accurate mortality estimation. Rather than looking at a few risk factors, the model looks at the patient’s entire electronic health record (EHR), including notes buried deep in PDFs or scribbled on old charts. Using this process, in the future, may enable doctors to save lives and provide better patient care.

Saving lives and money

So what can we do with this information? With more accurate predictions of a patient’s mortality, hospitals and doctors can use better estimations to adjust treatment plans, prioritize patient care, and predict negative outcomes before they occur. In addition to this, health care workers wouldn’t have to spend as much time manipulating patient data into a standardized, legible format.

For example, a report by Futurism notes that Ultromics, an AI diagnostics system developed in England, can diagnose heart disease more accurately than doctors. The same report notes that a startup bot called Optellum is working on an AI system that can diagnose lung cancer by analyzing clumps of cells found in scans. This bot shows promise to diagnose 4,000 additional lung cancer cases per year and at an earlier rate than doctors are currently capable of diagnosing.

Not only can these AI diagnostics systems save lives, but they can also help hospitals save money. In an interview for Futurism‘s report, Timor Kadir, Optellum’s chief science and technology officer, stated that the AI system could cut health care industry costs by $13.5 billion. Sir John Bell, chair of the U.K.’s Office for Strategic Coordination of Health Research, added, “There is about $2.97 billion spent on pathology services in the National Health Service. You may be able to reduce that by 50 percent.”

Predicting death for better care

Current research shows that less than half of the eight percent of patients who need palliative care actually receive it. There are times when doctors make inaccurate or overly optimistic prognoses about a patient. Dr. Kenneth Jung, a research scientist at Stanford University School of Medicine, told NBC, “Doctors may not make the referral [for palliative care] simply because they’re so focused on managing their patients’ health issues that palliative care doesn’t cross their minds.”

Failing to identify patients who need palliative care can have devastating consequences. If the patient’s health suddenly declines, they may spend their final days receiving aggressive medical treatments in hopes of extending their lives by a few weeks. However, studies have shown that approximately 80 percent of Americans would prefer to die at home, rather than in a hospital. Sadly, the report also notes that 60 percent of these people die in acute care hospitals.

It’s in these cases that AI can help identify patients who are critically ill and might benefit from end-of-life care. Early identification of these patients can help them get the treatment they need sooner. And it may allow patients to remain at home, instead of in the hospital, during their final days.

While some may wonder about the future of AI in health care, the purpose of AI systems is to play a supporting role in the health care industry. These systems will serve as a powerful tool that will help physicians and other health care professionals provide higher quality care and offer palliative treatments in a timely manner.

Scott Bay is a writer who covers AI and Internet of Things for PC Mag, Wired, and Men’s Health.

VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn more about membership.

Author
Topics