Reducing Readmissions with Predictive Analytics: Why Should Healthcare CIOs care?

By CIOReview | Thursday, August 11, 2016
607
923
206

Overview

With the increased rate of technology adoption (mainly because of the Centers for Medicare and Medicaid Services and Medicare penalizations), the U.S. healthcare sector in 2014 was undergoing a transition through the three wings of data management—data collection, data sharing, and data analytics. The healthcare organizations appreciated effective usage of analytical tools for this transition and therefore the scenario seemed appropriate to on board predictive analytics tools.

While clinical record digitization offered cosmic magnitudes of information, unstructured clinical notes were being managed by innovations in data management into compelling structures. The next step  in this transition phase was to make the extraction of facts from clinical and administrative repositories simple. First half of 2015 saw major healthcare providers and technology vendors joining hands to create predictive solutions and bringing them to market. Amplifying number of technologists and entrepreneurs started to understand the nuances of managing and organizing data to create and integrate predictions into healthcare workflow and keeping it simple at the same time. Their aim was straightforward—to reduce patient readmissions.

Extensive research and hard work paid off, and 2016 saw the health system's readmission rate dropping two percentage points to 13.5 percent. Today, Google uses unlikely data sources to predict in semi-real time the areas to be affected by flu and dengue. Providence Health Plan can determine the patient population that will be enrolled in a care management program by using a simple financial model. Cleveland Clinic also evaluates provider quality through a simple risk adjustment score. All these changes can be attributed to Predictive Analytics. But the core concern here is to identify of predictive analysis successfully running the Hospital Readmissions Reduction Program. If so, then what does the future behold for this technology? And why should a Healthcare CIO care? Let’s discuss.

The Current Scenario

 One of the biggest problems that healthcare organizations face today is identifying patients at risk of readmitting before they leave the hospital and let the care providers intervene Predictive Analysis can serve up the risk scores in interventions in a very clear and simple-to-use interface. The true value of your data is thus realized at the point of care with actionable insights which were acted upon immediately. Many healthcare facilities are benefitting by leveraging Predictive Analysis. So does this mean that the federal penalization has gone down?

The federal government penalized 758 hospitals with higher rates of patient safety related incidents, and more than half of those places had also been fined last year. Among the hospitals that were penalized for the first time, a significant number involved were well-known healthcare institutions. So where did they go

Electronic Health Record systems (EHR) are widely implemented over the last five years in the U.S. and as a result, vast amounts of data are now accessible to health care systems. But big data in healthcare still remains a buzzword, with health care systems still learning how to broadly apply such analytics to improve patient outcomes and reduce spending. So what must the healthcare CIOs do?

A word for the Healthcare CIOs

While Predictive Analytics have a lot of potential to offer for healthcare sector, implementation is the key. Healthcare CIOs need to identify steps that make Predictive Analytics and Algorithms an integrated part of routine patient care.

Determining the Clinical Decision

For nearly every potential clinical outcome, a plethora of data is now available through which a potential predictive algorithm can be sorted out. Healthcare CIOs need to form a process to develop these clinical algorithms while being sure that the algorithms are equally specific for the clinical decision they respond to.

Leveraging Data from EHRs

Healthcare CIOs must take note that algorithms are only as reliable as the data they are based on. Large amounts of data may not be required by algorithms for acute clinical issues (e.g., heart attack, septic shock) for risk prediction. Thus, predictive analytics process of any organization must be able to utilize amounts of clinical data with greater accuracy and potential clinical applications.

Focusing on low-value Decision Points

Physicians are often in a state of uncertainty when treating patients, leading to over-treatment or vice versa. Predictive analytics can solve this quandary as clinicians are allowed to steer high-cost interventions for the high-risk patients. Healthcare CIOs must incorporate predictive analysis to focus on these so-called “low-value decision points” and reduce medication costs and side-effects among patients.

Quit Forcing Analytics into the Existing Workflow

Physicians see hundreds of numbers (vital signs, laboratory values, etc.) each day. So there exist the chances   an algorithm’s output to be just another number that physicians ignore in case it does not fit well into their daily workflow. CIOs and the IT department must device a mechanism which selects only those analytical details which are iterative enough to be useful for a practicing physician. They must make sure that the employed mechanism doesn’t force every possible analytics into the existing workflow.

What the Future Has in Place!

Predictive analytics have started becoming more complex and sophisticated. However, if such analytical data cannot be applied by the health systems in improving value in everyday clinical care, then this sophistication has no value at all. Health systems must critically think about the clinical situations in the times to come where enhanced analytics proves to be useful and aid healthcare institutions in using them in patient care routinely. Healthcare CIOs must develop strategies to evaluate the clinical impact of analytics. Spending can be reduced and clinical outcomes can be improved if CIOs make their organizations in targeting interventions to patients who need them the most.