Jan
12
2015
Predictive analytics reduces readmissions and advances the quality of care.
Imagine a standard of care where before the patient leaves the hospital, the institution already knows not just the patient’s critical data, but has also taken actions in response to the data that further reduce the risk of readmission within the next thirty days. The potential impact goes beyond delivering a more advanced quality of care. For institutions facing the radical industry shift to patient-based outcome models, predictive analytics presents an opportunity to make significant strides to decrease patient re-admission rates — and gain a stronger foothold in meeting compliance mandates the industry faces as a result of the Affordable Care Act.
In October 2012, the first round of readmission fines began to be levied against the nation’s hospitals. Reporting on recently-released federal data, the Wall Street Journal estimates that some $227 million in fines are being doled out.1 Hospitals are in the midst of a readmission crisis and existing solutions fall short in taking into account the complexity of patient disease as well as finding ways to get to the root cause of the patient’s health condition.
The Health Research & Educational Trust (HRET) recommends that hospitals risk stratify patients as an approach to the readmission problem. Using risk stratification, the industry has learned the importance of predicting hospital readmission as a very useful way to deploy appropriate, targeted interventions during the patient’s initial stay. LACE is a commonly used risk assessment model. A patient’s LACE score combines: length of stay (L), acuity of admission (A), Charlson index (modified) (C), and number of emergency department visits in the last 6 months (E) to assign a score to patients at risk of readmission.
But the LACE model has several inherent limitations. First off, the LACE attributes are only indirect markers of risk. They aren’t necessarily the actual factors indicative in the underlying causes of a particular patient’s risk. In addition, the “L” component of LACE (length of stay) cannot be determined until after the patient is discharged. Limiting practitioners to this “after the fact” knowledge does little to advise hospitals on the immediate actions they can take, in real-time to prevent readmission in the first place. In addition, knowing a patient’s LACE score does not offer insights on ways to directly improve or attend to the underlying causes of a particular health situation.
Not Just a Risk Score
Health Information for Professionals reports that Gartner forecasts that by 2016, 70% of the world’s most profitable companies will manage business processes by using predictive analytics.2 A few healthcare organizations are carrying over these innovations into their own operations. Predictive models offer a view into the relationships among many factors and thus allow more accurate risk assessment from a particular set of conditions. Predictive analytics findings give insights that guide immediate, actionable care decisions and treatment methods. Drawing on techniques from statistics, modeling, machine learning, and data mining that analyze current and historical facts to make predictions about future events, predictive analytics uses patterns found in historical and concurrent data to pinpoint patient risks and opportunities.
At Carolinas Healthcare System, Chief Innovation Officer Jean Wright set out to build a reliable model of readmissions. The test was to discover if predictive analytics could be used in a live environment to make decisions at the point of care. She was also testing whether, once the data was there, there was a solution that could be designed and used by clinicians themselves. The goal was to identify patients at risk of readmission before they left the hospital and enable care providers to intervene.
The Carolinas system deployed a predictive analytics solution in just 6 months across 7 hospitals. To date, it has been used by 208 case managers and facilitated 123,000 interventions on 100,000 patients. In addition to accurately predicting patients at risk of readmission, predictive analytics facilitated fundamental productivity improvements in the way nurses and case managers cared for patients. Instead of making patient rounds based on room number or scheduled time of discharge, rounds shifted to patients with more complex and urgent care needs. This also translated to the organization’s ability to better manage and respond to workloads across floors and units. Predictive analytics driven interventions also helped the hospitals give patients more comprehensive materials for their aftercare once they were discharged, and ensured a greater level of consistency in care.
Predictive analytics is one way organizations can advance the quality of their patient care and avoid the financial consequences that come with higher readmission rates. Predictive analytics and the use of cloud-based solutions put actionable information into the hands of healthcare providers who can use it instantly within a hospital’s IT system or via a mobile device. The result is a win/win for hospitals looking to mitigate risk and improve outcomes for the patient.
About the Author
Nish Hartman, healthcare director at Predixion Software, is an experienced leader at building and executing strategies for innovative healthcare solutions. She has been an active participant in professional and community focused organizations including The American College of Healthcare Executives, a former Board member of North Carolina HiMMS and Fellow for the Batten Leadership Institute.
References