Seeing the Future with Predictive Analytics and Clinical Data | #PutData2Work

Predictive models can be found in every corner of our lives today, from the targeted advertisements that pop up on our screens to the weather forecasts we see every day.  When predictive models are used to improve patient care and outcomes, the benefits can be profound.  Predictive analytics can be a crystal ball to improving outcomes and managing costs in three areas:

  1. Improving Care Delivery,
  2. Cost Containment, and
  3. Reducing Preventable Readmission

Since healthcare organizations are rich in clinical data, use of predictive analytics provides the science and methodology to transform this wealth of observations and results into actionable forecasts that can accurately identify patients at the highest risk for adverse outcomes. 

Sometimes, predictive analytics tells us what we already know; but robust algorithms can lead us to variables, and—more importantly—combinations of variables that may be the most optimal predictors with the highest performance.  These predictors have the potential to serve as an early warning system that contributes to a change in clinical practice, and can result in reduction in mortality and morbidity.

In the U.S., the Centers for Medicare and Medicaid Services (CMS) funds most of the health care costs for the end stage renal disease (ESRD) patient population, including non-Medicare populations.  CMS oversees dialysis services in a highly regulated and closely measured environment.  There are over 600,000 ESRD patients in America, and to sustain life, patients with ESRD require either a constant regimen of dialysis that generally occurs daily to three times per week, or a renal transplant.  The estimated average cost of caring for an ESRD patient is nearly $90,000 per year.  ESRD patients are typically very ill with many comorbidities.  The interaction of ESRD patients with providers at dialysis facilities is unique, with both very frequent and long-term ongoing episodes of care, unlike any other area in the healthcare ecosystem.

In my work at Fresenius Medical Care, I have been focusing on an innovation to leverage the data from more than a million patients and 250 million dialysis treatments to develop, test, and implement models to predict which patients are most and least likely to be hospitalized, miss scheduled treatments, or have a decline in their functional status. We have used these predictions to effectively target interventions that reduce both suffering and cost of care for our patients.  Our success is iterative: We learn and improve our models over time.

We tested a model that predicts which patients will have more than five hospital admissions in the subsequent year based upon a set of variables. By providing these predictive risk factors to the clinicians, we were able to facilitate early intervention and reduce the hospital admission rates by 40% in the high risk patients.

We are using the power of data and analytics to positively influence the clinical and social outcomes for our patient populations.  Millions of ESRD patients face a constant regime of treatments, costing on average more than $80,000 per year per patent. Learn how one organization leverages data to develop and implement statistical predictive models that improve patient outcomes through timely intervention at the HIMSS Clinical & Business Intelligence (C&BI) Community May event “End Stage Renal Disease (ESRD) and Use of Advanced Analytics in Improving Outcomes.”

What is your wealth of data? How could you use it to predict the future?

Join the HIMSS C&BI Community today to learn more about predictive analytics and other critical topics that will help you on your journey to Turn Data to Action.

Predictive AnalyticsClinical DataESRDPopulation HealthClinical & Business Intelligence