Troubled by student retention rates that remained unchanged over the past 20 years, Middle Tennessee State University (MTSU) recently turned to a patient-care model from the healthcare industry as a possible solution. And early indications are that the problem is responding to treatment: In just one year, the 23,000-student institution saw the number of students returning for the spring semester rise by 400.
The healthcare model, known as Population Health Management, has been tailored for the needs of MTSU and other schools by Education Advisory Board, a consulting and research company that provides a predictive analytics solution called Student Success Collaborative to about 200 institutions.
In the healthcare setting, PHM aims to make the most efficient use of scarce resources by providing differing levels of support to three patient groups: low risk, rising risk, and high risk. At one end of the spectrum, low-risk patients require little care; at the other, high-risk patients need a broad array of expensive medical services. Interestingly, the middle group is where medical providers see the best chance to impact long-term health outcomes. “These rising-risk patients may have an identifiable condition, such as diabetes, but perhaps it’s not a problem right now,” explained Ed Venit, a senior director at EAB. “If healthcare providers catch the problem through some really aggressive monitoring, they can prevent them from moving into the high-risk category. ”
MTSU and EAB have been working together since late 2014 on how the PHM approach can be molded to tackle the issue of student retention, which presents many of the same challenges. Central to any initiative of this sort, though, is a robust analytics capability. Predictive analytics have been front and center in higher education for the past five years at least, so Venit feels most colleges and universities are now in position to benefit from a PHM-style approach. “You need to have analytics for this to work,” he said. “The good news is that most schools now have a tool of some kind.”
While the data set used by MTSU is unique (as it is for most schools), it shares commonalities with information sources available at most U.S. colleges, usually through the SIS. “We want data sources that have at least 10 years’ worth of information; include every student on campus; have high predictive value; and, maybe most important, sources that are relatively easy to obtain,” said Venit about EAB’s approach to working with schools. “If you have to go around campus gathering data, it becomes a very weighty endeavor for the IT office. It can extend implementation time or even kill the project altogether.”