Data system helps address complex questions, improve medicine
The federal government has targeted billions of dollars of savings for the Medicare program by reducing the number of preventable readmissions.
Financial penalties to hospitals with excessive readmission rates begin Oct. 1.
Much of the information reported on attempts to lower readmission rates, including that on the CMS website Hospital Compare, indicates little documentable progress in this regard. Why is this so? What is the anatomy of a readmission that makes the clinically appropriate reduction of readmissions apparently so difficult?
NorthShore University HealthSystem fully implemented a comprehensive electronic healthrecord system across all of its hospitals and ambulatory-care centers nearly a decade ago. Data from the system feeds the data warehouse and creates the opportunity for the application of data analytics to address complex clinical questions. Applying such data analytics at NorthShore to the question of readmissions led to insights into the number and type of readmissions and the opportunity for improvement.
Beginning in the spring of 2011, a team at NorthShore applied data analytics tools to focus on readmissions—particularly conges- tive heart failure (CHF). Data analytics created a unique and specific predictive modeling tool that identified NorthShore patients at high, medium and low likelihood of readmission. By successfully deploying people, processes and technology toward the high-risk population, NorthShore reduced its readmission rate from 21% to 17% over a period of six months in initial units. It became apparent that caregivers also needed to rethink their beliefs about readmitted patients. Rather than considering a readmission as usual or a “frequent flyer” (repeat offender), the care team embraced a culture shift for positive transitions to home or skilled facilities.
What did NorthShore find as to the anatomy of readmission patients—their characteristics? The CHF readmission profile of more than 2,000 Medicare CHF patients reveals an extraordinarily complex and extremely ill population requiring the management not only of the original reason for admission, but also the effects of multiple chronic disease conditions. In summary, the typical patient’s profile included: Average age: 84 Multiple simultaneous diagnoses: CHF, arrhythmia, pneumonia, diabetes
Average of being on 10 medications
Little/no functional family support
The implication of this severe anatomic profile is that so-called preventable readmissions will be nearly impossible to eradicate, let alone significantly reduce in numbers. Therefore, a CMS policy that financially penalizes and excoriates providers as “bad actors” on the stage of healthcare “reform” may, in extremis, unnecessarily lead such health systems in the future to hesitate in taking on such high-risk patients. The implications for such a state are substantial not only to the CMS and providers, but to Medicare participants and their families.
The application of data analytics in the healthcare field is an exciting opportunity to better understand and improve the practice of medicine. No attempt is made in this brief summary to claim statistical significance on readmission reductions from this NorthShore study, only a glimpse into better understanding the anatomy of a readmission and the complexity providers face each day.
The typical readmitted patient was elderly and suffered from multiple medical conditions.