For many students, the stress of school goes beyond academic achievement or financial pressure. Increasingly, students are also dealing with mental health issues as they pursue their post-secondary education. A recent survey of 19 colleges across eight countries published by the American Psychological Association found that one third of students identified as having at least one mental health challenge, including depression or generalized anxiety.

In the U.S., on-campus mental health services report that anxiety is their top concern. While schools are increasing funding for care, resources frequently remain insufficient to meet the growing need. The result is that students often go undiagnosed or unsupported and then drop out of school.

At the same time, schools are also focused on keeping their retention rates up because they receive financial support based on performance. In many parts of the U.S., if a college has an 80 percent or higher graduation rate, they receive significantly more funding than if they’re below this level. Preventing students from dropping out isn’t just good for students, it’s also financially beneficial for institutions.

Mental health plays a pivotal role in a student’s success or failure, which is why colleges have been looking at ways to identify problems in order to offer support. But, despite their efforts, they have found that identifying students at risk has been easier said than done.

Why attendance and grade-based solutions fail

In the last decade, several enterprises have developed solutions that would flag students who are at risk of not finishing their studies due to mental health concerns. These tools allow colleges to track students in relation to their academic achievement with the aim of identifying potential issues. The assumption is that, if students stop attending classes, do poorly on a series of quizzes, or fail to submit work, they may be having a problem.

But, even if we assume that changes in grades or attendance are clear indicators of potential mental health issues, these approaches include considerable time lag. To get meaningful data and recognize a red flag, you have to see changes in a pattern. Both attendance and grade-based patterns take weeks, if not an entire semester, to establish. If a student is having trouble in the first month, there’s a good chance that the system will not have accumulated enough data to recognize it.

Schools need an easier and more time-sensitive way to identify changes. And they need to be looking at more than just attendance and grades.

A better way to identify at-risk students

In order to be able to pinpoint at-risk students quickly and easily, solution providers are giving colleges the tools they need to analyze the data already being collected by their access control systems (ACS). They want to use this data to understand individual student behavior and then quickly identify anomalies or changes. The ability to do this lies with machine learning and predictive analytics.

Consider the first-year student at a medium-sized college. He lives in a residence hall, attends all his classes and gets good grades. According to the traditional methods of tracking students, this young man is not at risk in any way. But the ACS data that this school has been collecting for more than 10 years tells a different story.

The school’s ACS indicates that, from Monday to Friday, the student leaves his dorm room on time for class and works out. But it also shows that, from Friday afternoon to Monday morning, he never leaves his room at all.

At this college, the system is set to send automatic notifications if it detects an anomaly. In this case, the student’s resident assistant (RA) receives a notification to check in on him. He finds that the student hasn’t yet made any friends at school and is lonely. This is leading him to isolate himself all weekend, every weekend. With the RA’s help, the young man gets some counseling, joins on-campus clubs and starts participating in activities in residence. Within weeks, he’s making new friends and thriving at college.

Using ACS data to foster better mental health

Of course, it could have gone the other way. The student could have had a good reason for not leaving his room. What is important is that the school’s ACS was able to use analytics on the data it was aggregating and notify the appropriate person as soon as that data suggested something might be amiss. This allowed the RA to check in and provide support to a student who needed it.

The reality is that colleges are already watching every student, but no one is paying attention unless we flag concerns. By keeping track of student activity on campus, activity that is structured and predictable at an individual level, we can identify anomalies and changes. This gives campus community members the opportunity to make sure students are okay.