The following is a guest blog post by Jennifer Bergeron, Learning and Development Manager at The Breakaway Group (A Xerox Company). Check out all of the blog posts in the Breakaway Thinking series.
As 2016 approaches, individuals and organizations are beginning to consider their New Year’s resolutions. In order to make a plan for change, we imagine ways we might reach our goals: “If I eat more vegetables, I’ll lower my cholesterol and have more energy. But if I eat more vegetables and skip the donuts, I will see the same improvements faster!” What if someone could tell us exactly what action or combination of actions would produce which results over a specific timeframe?
In healthcare, predictive analytics is doing just that – providing potential outcomes based on specific factors. The process involves more than gathering statistics that define group results, but research of patient outcomes that allows predictions for individuals. Both technology and statistics are used to sift through these results and turn them into meaningful insights. Considering big data and a patient’s own health information, diagnoses can be more accurate, patient outcomes improved, and readmission rates reduced.
Predictive analytics is being used to help improve patient safety, predict crises in the ICU, uncover hereditary diseases, and reveal correlations between diseases. Researchers at the University of California Davis are using electronic health record (EHR) data to create an algorithm to warn providers about sepsis. Genomic tests, an example of precision medicine, are now available for at-home DNA testing, which allows individuals to discover hereditary traits through genetic sequencing. Correlations can be found between illnesses using EHR data. Thirty thousand Type 2 diabetic patients were studied to predict the risk of dementia.
BMC Medical Informatics & Decision Making reported on the use of EHRs as a prediction tool for readmission or death among adult patients. The model was built using specific criteria: candidate risk factors had to be available in the EHR system at each hospital, were routinely collected and available within 24 hours, and were predictors of adverse outcomes.
But predictive analytics can only be as good as the data it uses. Accurate, relevant data is necessary in order to receive valuable information from the algorithms. But the information can be hard to find, considering that healthcare data is expected to grow from 500 to 25,000 petabytes between 2012 and 2020 (A petabyte is a million billion bytes). In an effort to solve this challenge, more than $1.9 billion of capital has been raised since 2011 to fund companies that can gather, process, and interpret the increasing amount of information.
There are four principles to follow in order to optimize how information is captured, stored, and managed in the EHR system:
- Ensure that leadership delivers the message to the organization about the importance and future impacts of the EHR system
- Quickly bring staff up to speed
- Measure and track the results of the staff’s learning
- Continue to support and invest in EHR adoption.
The EHR stands as the first point of collection of much of this data. Given the importance of accuracy and consistency, it is critical that EHR education and use is made a priority in healthcare.
Xerox is a sponsor of the Breakaway Thinking series of blog posts. The Breakaway Group is a leader in EHR and Health IT training.