How Predictive Technology is Improving End-of-Life Care
Analytics help providers determine which patients need more care and those most at risk
by Dan Hogan

Predictive modeling is everywhere. Google uses predictive algorithms to guess what we are searching as we are typing in the search bar. Amazon uses these formulas to forecast what we might buy today based on millions of purchases made by people similar to us. These companies can do this by aggregating large swaths of data to identify commonalities between customers to predict what we might do. Today’s society is comfortable with using machines to ferret out data and tailor our information.

This practice of using predictive algorithms is being applied to health care too, as analytics companies gather thousands of records and pool patient data to better predict outcomes for patients, and help caregivers determine the best course of treatment.

In most industries today, data pooling is ubiquitous. Even within a company, data is regularly shared across departments to determine staffing needs, products to stock, contingency plans and many other things that allow a company to run more smoothly. Though the health care industry is behind some other industries in leveraging historical data, similar algorithms are now being used to predict outcomes that are life changing—or, more specifically, end-of-life changing. Care providers are using predictive analytics to determine everything from which patients might benefit from additional care to those at risk of passing away within 90 days.

Why Is This a Good Thing?

Each patient has the right to live and die with dignity. Analytics can help patients—and the doctors who treat them—do just that. Currently, even though 80 percent of patients would rather pass away at home, 60 percent of Americans die in acute care hospitals and 20 percent in nursing homes, according to the Stanford School of Medicine. This leaves only 20 percent who will die the way they had hoped. This sad trend is costing the health care system billions of dollars—without necessarily resulting in better health outcomes for patients or satisfaction with care. If a patient knows that he or she is nearing his or her final days due to the combination of concrete data and a caregiver’s experience, the patient can determine how to spend those days.

Statistically, patients who are dying—and their families—report better patient satisfaction ratings if they are transferred into hospice or palliative care programs earlier. It gives them extra time to put personal affairs in order.

Technology and health care work together to ensure our final days are comfortable and dignified. Predictive analytics can help caregivers know when additional or different care may be needed and when it’s time to discuss end-of-life care, homecare, hospice and other issues—allowing patients and families the opportunity to make decisions together.

The Future of Data Pooling

Predictive modeling is still being adopted within the health care industry. But it’s already going to the next level with prescriptive analytics, which takes patient data, analyzes it, compares it to other data, and then helps prescribe the statistical best course of action and treatment options for the patient. Predictive analytics provides objective data that caregivers can use in conjunction with their education, experience and instincts to create a care plan that is personalized—and potentially more effective.

To continue to achieve better outcomes for patients in the future, it would be tremendously transformative to pool our health care research data and create a “data commons,” a centralized, networked database where parties can share information. Medical professionals would be able to share data from medical studies reaching back generations. That data could be mined for correlations that lead to new treatments and personalized health care.

Whether increased data pooling comes as the result of private institutions coming together or a government project, having such a vast pool of information would allow us to predict outcomes even more quickly and more accurately. Caregivers could intervene at the right time to improve health in a way that’s never before been possible.

What Now?

No analytics tool will ever take the place of hands-on care or the expertise of your caregiver. Nor should predictive modeling ever be the final say. Caregivers can use predictive technology to improve the certainty of diagnoses and courses of care; however, by design, it must be used in conjunction with the knowledge caregivers have of individual patients and histories to have the greatest impact and ensure the best care for patients.

As patients and caregivers, we need to keep communication open and discuss the very things that might make us the most uncomfortable. For example, conversations around advance care planning should happen before entering the final days so that end-of-life preferences are understood, especially with regard to hospice. Fortunately, end-of-life conversations are being encouraged these days, and in 2016, Medicare’s advance care planning policy began covering up to $86 for an office-based counseling session. End-of-life decisions are not typically made after one conversation. It’s a lot to digest, and patients need to have time to absorb the information, discuss it with their families and make informed choices.

Eighty-nine percent of patients will rely on doctors to begin the conversation about end-of-life care. If doctors do not initiate the discussion, it’s important that patients encourage open dialogue early and often, so care options and preferences are clear. Caregivers are there to ensure decisions are carried out, and technology can be a huge help in making sure that decisions made in the final days are aligned with patient wishes and maximize quality of care.