A Risk Algorithm for the Persistence of Suicidal Thoughts and Behaviors During College

Abstract

Objective: The primary aims of this study are to (a) identify patterns of suicidal thoughts and behaviors (STB) during college among students with lifetime pre-matriculation STB and (b) develop a risk-screening algorithm for persistence of pre-matriculation STB during college.

Methods: Data come from the Leuven College Surveys, a series of prospective cohort studies of all incoming KU Leuven University freshmen. In the academic year 2012-2013, 4,889 incoming freshmen (73.2% response rate) provided baseline data on sociodemographic variables, childhood-adolescent traumatic experiences, 12-month stressful experiences, 12-month mental disorders, 12-month STB, and severity markers of pre-matriculation STB. A total of 2,566 students (69.3% conditional response rate) participated in 12- and 24-month follow-up surveys during the first 2 college years.

Results: Thirteen percent (weighted n = 535) of incoming freshmen reported lifetime pre-matriculation STB. Of those, 28.0% reported 12-month STB in 1 follow-up assessment, and another 27.7%, in both follow-up assessments. High persistence of STB (ie, 12-month STB in 2 follow-up assessments) was most strongly associated with severity markers of pre-matriculation STB, with odds ratios in the 2.4-10.3 range and population attributable risk proportions between 9.2% and 50.8%. When the aim was for less than 50% of false-positive cases (positive predictive value = 54.4%), a multivariate predictive risk algorithm (cross-validated area under the curve = 0.79) situated 59.9% of highly persistent cases among the 30% respondents with highest baseline predicted risk.

Conclusions: An individualized web-based screening approach is a promising strategy to identify students at the time of university entrance who may be at high risk for STB persistence during their academic career.

Figure 1

Course of STB during college among students with a prematriculation lifetime history of STB (weighted n=535). Note: proportions indicate the number of participants that report (no) 12-month STB within the group where the arrow originates from. Groups with (no) 12-month STB in follow-up 1 were grouped together to calculate proportions in follow-up 2 (as indicated by the vertical bars). ^a Reporting one time 12-month STB during two-year follow-up ^b Reporting two times 12-month STB during two year follow-up Abbreviations: 12-m = 12-month.