Appendix D

The Rate of Student Death from Suicide from the Big Ten Counseling Centers: 2009-2018

Counseling Center, University of Illinois at Urbana-Champaign
Raquel Mendizábal Martell, Ph.D. & Matthew King, Ph.D.

Prepared for the Committee on Mental Health, Substance Use, and Wellbeing
of the National Academies of Sciences, Engineering, and Medicine

July 24, 2020

METHODOLOGY

Study Design

Silverman et al. (1997) conducted a comprehensive 10-year study at 13 Big Ten university campuses to get a more accurate understanding of deaths by suicide in college campuses. Silverman et al. (1997) collected data from 1980 to 1990, at which point the study stopped. In an effort to continue to understand suicide rates and trends in college campuses, the Big Ten university counseling centers decided to continue to collect data on deaths by suicides in their campuses. Data collection resumed in 2003 and it continues to be collected, with some data entered retrospectively and some data entered prospectively.

Representatives at the Big Ten universities counseling centers (e.g., directors, data analysts, designated staff) were asked to report all student deaths by

suicide to a database. Each participating Big Ten universities counseling center sought research approval from their universities to participate in the study and contribute data. All schools have different methodologies by which they gather and contribute data to the database. Generally, representatives at the counseling centers received information about student deaths by suicide from other departments, such as the Office of Dean of Students, and then they entered relevant information into the database.

Time Frame

For the purpose of this report, data from the most recent 10 years were analyzed given that various universities could not verify the data on deaths by suicide for the early 2000s. Thus, the time frame for this report is September 1, 2009, to August 31, 2019. The current report uses an academic year time frame, which is defined as the 12-month period from September 1 to August 31. This reporting time frame is consistent to the time frame used by Silverman et al. (1997), which more closely aligns to academic calendars in university campuses, which is relevant as the population studied is college students.

While most universities entered data prospectively, some universities entered data for some years retrospectively. A few universities entered the study after its start, and thus entered some data retrospectively and some data prospectively as well. The Big Ten universities counseling centers consider the Suicide Registry an ongoing project and data continue to be collected.

Period of Exposure to Risk

The calculation of period of exposure to risk for the current analyses and report is consistent to Silverman et al. (1997). Specifically, for each academic year, the fall semester enrollment numbers for each university were added. Enrollment numbers were gathered using the public information available through each university’s website. For universities for which only some of these data were available, general reported rates or percentages for each school were used to determine an approximate number of students for demographic information.

Definitions: Student Status and Suicide

The definitions of “student” and “student suicide” in this report were kept consistent to the definitions used by Silverman et al. (1997). A “student” is defined as an individual who was registered for credits as a full- or part-time student during the academic year in undergraduate or graduate programs in the main campuses at the Big Ten universities. A “student suicide” is a death by

suicide that occurs by an individual who was a “student” as previously described within 6 months from being registered as an active student. The “6-month rule” accounts for student absences in which students are not actively registered for a short period of time such as summer months, academic leave, and medical leave (Silverman et al., 1997).

Consistent with Silverman et al. (1997), “suicide” is “a self-inflicted injury resulting in death” (p. 289). Representatives at each Big Ten counseling center identified a student death in their university as a “suicide” internally based on information they received of the incident. Student deaths for which cause was inconclusive based on the information representatives received (e.g., cause of death was considered unknown, accidental, or natural), were not included in the database or this study as they are not considered a death by suicide.

Confidentiality

The current study ensured adherence to proper research procedures and the confidentiality of personal information. All participating counseling center representatives sought the proper permissions in their respective universities (e.g., IRB approval) to contribute data.

To ensure that data were contributed in a secure way that would protect the identity and privacy of the students, all Big Ten universities were assigned a numerical “University Code” that they used to enter their data to the database rather than using the institution name. The data is stored in a secured database at the University of Illinois Counseling Center, which serves as the repository for this data. The raw database is only accessed by the study team. Data entered is not identifiable given that student names or initials are not entered. As such, the raw database does not have any first or last names, and universities have random numerical codes that have only been shared among the Big Ten counseling center representatives. All data are reported in aggregate form, and data were pooled when appropriate to ensure that data cannot be identified by school or to an individual.

Participating Universities

Thirteen Big Ten universities participated in and contributed data for the study. The 13 participating universities were located in the Midwest and Northeast in the United States in the following states: Illinois (Northwestern University, University of Illinois at Urbana – Champaign), Indiana (Indiana University, Purdue), Iowa (University of Iowa), Michigan (Michigan State, University of Michigan), Minnesota (University of Minnesota), Nebraska (University of Nebraska – Lincoln), New Jersey (Rutgers), Ohio (The Ohio State University), Pennsylvania (Pennsylvania State), Wisconsin (University of Wisconsin – Madison).

Data Sources

Enrollment and student population demographic data for participating Big Ten universities were obtained from public records and information published by each university through its website (e.g., Office of the Registrar website). Data from fall 2009 through fall 2018 were obtained. Enrollment and demographic data for most universities were readily available. Demographic and gender identity breakdowns for universities for which these data were not publicly and readily available were estimated using the information publicly available for each of those schools (e.g., percentages available).

Table D-1 summarizes enrollment numbers and student population demographic data. Academic class was used to classify academic parameters. There are two classifications: undergraduate (i.e., freshmen, sophomores, juniors, seniors, non-declared, or other undergraduate) and graduate (i.e., masters, doctorates, professionals, post-graduates). Gender identity is classified as male, female, and given the small numbers of reported students who identified as transgender or another gender, these numbers were pooled in the demographic information and in student suicide data with those whose gender was unknown under the category “Other.” Racial identity is classified as Asian, Black, Hawaiian/Pacific Islander, Hispanic, Native American, White, Multiracial, and Unknown. Given the small number of student suicides and students who identified as other than White, Asian, and Black, student suicides by a student who was Hawaiian/Pacific Islander, Hispanic, Native American, or Multiracial were classified as “Other.” Given that some universities reported International/Foreign students as a racial group and other universities reported International/Foreign students separately and in addition to a racial category, analyses were run based on nationality with two classifications: International and Domestic.

Results

The total number of students across all Big Ten universities for the years (2009-2018) for which data were reported was about 4.8 million students (Table 1). The total number of deaths by suicide reported by the participating universities was 268, and the total number of deaths by suicide entered into the database by the participating universities was 231.

The overall average suicide rate per 100,000 students is 5.60 for the years 2009 to 2018 in the Big Ten college campuses. This is lower than the rate calculated by Silverman et al. (1997), which was 7.5/100,000. Specifically, there was a 25.3% decrease in the suicide rate in Big Ten universities in 30 years from 1980-1990 to 2009-2018.

The rate of 5.60 is also lower than the U.S. population national average suicide rate of 14.2/100,000 based on 2018 data from the National Center for Health Statistics (NCHS) at the Centers for Disease Control and Prevention (CDC; Hedegaard, Curtin, and Warner, 2020). The 5.60/100,000 rate in college

TABLE D-1 Demographic Profile of Campus Population at Big Ten Universities: 2009-2018

TABLE D-2 Student Deaths by Suicide at 13 Big Ten University Campuses (2009-2018)

*per 100,000 students.

campuses is also lower than the average U.S. population suicide rate for age groups 15-24, 25-34, 35-44, and 45-54 for males and females which was about 17/100,000 based on National Institute of Mental Health (NIMH) data from 2017 (National Institute of Mental Health, n.d.). A weighted average of the NIMH suicide rates by age group was calculated using age percentage distribution data published by the U.S. Census Bureau (Census Bureau, 2019), finding the weighted average suicide rate for age groups 15-24, 25-34, 35-44, 45-54, and 55+ was 14.72/100,000, which is still higher than 5.60/100,000.¹

In other words, the overall average suicide rate at Big Ten college campuses was 39.4% of the U.S. population national average. Additionally, the overall average suicide rate at Big Ten college campuses is 38.0% of the weighted average U.S. population suicide rate for college-aged individuals. Taken from a different perspective, the U.S. population national average suicide rate is about 254% higher than the average suicide rate in Big Ten universities. The weighted average U.S. population suicide rate for college-aged individuals is 263% higher than the suicide rate at Big Ten universities.

The NCHS identified a 35% increase in suicide rates in 19 years from 1999 to 2018 for the U.S. population. Conversely, the suicide rate in Big Ten college campuses has decreased by 25.3% in 30 years.

Comparison of Percentages

Percentages of deaths by suicide by gender, race, nationality, and class standing were calculated (Table D-2). When comparing percentages of deaths by suicide by gender identity, the majority of deaths by suicide occurred in males (67.53%) and it is higher relative to male representation on campus (51.18%). The majority of deaths by suicide by race occurred in White students (43.29%), which is lower relative to White students’ representation on campus (i.e., 74.70%). Asian students made up 10.39% of deaths by suicide on campus, making them the second largest racial group represented in deaths by suicide; this is somewhat similar to Asian students’ representation on campus (8.99%).

In comparing international students with domestic students, there was a higher percentage of deaths by suicide in domestic students (46.75%). Finally, a comparison by class standing indicates that undergraduate students make up the majority of deaths by suicides in college campuses (57.58%), which is lower relative to their representation on campus (i.e., 73.95%).

___________________

¹ These data from the U.S. Census Bureau were obtained from the “Current Population Survey” from October 2018 (Census Bureau, 2019). Data were provided by respondents from more than 56,000 families across the United States who completed follow-up interviews providing information, among other topics, about student status of individuals in the family. These data were compared to enrollment figures across Big Ten Universities, confirming that the number of students in each age group nation-wide is consistent with the number of students in these age groups across the Big Ten universities.

Comparison of Rates

Suicide rates and Poisson 95% Confidence Intervals (CIs) for gender, race, nationality, and class standing were also calculated (Table 2) using the Epidemiology Tools package R (Aragon, 2020; R Core Team, 2020). Poisson 95% CIs that are not overlapping signify that the rates are significantly different. For gender, the suicide rates for female (1.72/100,000) and male (6.37/100,000) students are significantly different, indicating that male university students have a higher suicide rate than female students. Regarding race, the suicide rates for Asian (6.33/100,000) and Black (8.34/100,000) students are significantly different from the rates for White students (3.18/100,000), but not from each other. As such, the data indicate that while most deaths by suicide in college campuses occur in White students (i.e., 43.29%), the suicide rate of White students is significantly lower than for Asian and Black students. The suicide rates for Black and Asian students were the highest suicide rates for race.

Suicide rates for which CIs overlapped were further explored using the Exact Rate Ratio Test R package using Python² (Fay, 2014; PSF, 2020). The results confirmed that there were no significant differences in the suicide rates for undergraduate (3.76/100,000) and graduate students (3.13/100,000). There were also no significant differences in the suicide rates for international (3.22/100,000) and domestic students (2.61/100,000). Regarding race, the results confirmed that the suicide rate for students in the “Other” race category (3.05/100,000) was not significantly different from the suicide rate of White students and Asian students, but the suicide rate was significantly different (p < 0.05) from Black students.

Limitations

One of the most noteworthy limitations of the current report is the potential inaccuracy of the data reported on suicide incidents in college campuses. As expressed by Silverman at al. (1997), the accuracy and the level to which the data reported is complete can be of concern in suicide-related studies. Steps were taken to ensure that data were accurately entered into the database, such as constant communication between the counseling centers at the participating Big Ten universities and the strong commitment to the study by all institutions. Each university had its own internal system to ensure that the data contributed to the study and the database were correct and complete. Given that a standardized audit from the research team for each university was not performed, it is possible that some data are missing.

___________________

² The Exact Rate Ratio Test package was unavailable for R version 4.0.0, so a Python script for this package stored in GitHub repository was used and can be accessed through github.com/kad-ecoli/rateratio_test.

Another limitation to note is that some universities were unable to enter all their data into the database and only reported the number of incidents by year in their campuses. This resulted in some missing data for the analyses.

Additionally, a limitation of this study is its scope. The purpose and scope of this report is to update the previous suicide rate for 100,000 in college campuses in the United States. Analyses examining the gender, race, nationality, and class standing suicide rates and percentages were also performed to understand the demographic composition of Big Ten universities and of students who have died by suicide. However, this report does not seek to prove causality or draw any conclusions about motive.

Conclusion

The growing discussion of mental health in recent years has led to closer attention to the mental health issues and struggles of college students, including deaths by suicide. The current report seeks to update data on deaths by suicide in Big Ten universities. Silverman et al. (1997) published, arguably, the most comprehensive study on deaths by suicide in college campuses. Silverman et al. (1997) collected data from 12 Big Ten universities from 1980 to 1990, finding that the suicide rate in college campuses was 7.5/100,000. This was lower than the U.S. population average suicide rate at the time, which was 15/100,000 (Silverman et al., 1997). The Big Ten university counseling centers decided to continue data collection on deaths by suicide. The current study found that the updated suicide rate in college campuses using data from 2009 to 2018 was 5.6/100,000. This rate is lower than the rate reported by Silverman et al. (1997). Therefore, in about 30 years, the suicide rate in Big Ten universities has decreased 25.3%. The suicide rate found in this report is also 39.4% of the U.S. population average suicide rate in 2018, which was 14.2/100,000 (Hedegaard, Curtin, and Warner, 2020). Notably, while the suicide rates in the U.S. population increased by 35% in 19 years from 1999 to 2018 (National Institute of Mental Health, n.d), the suicide rate in Big Ten college campuses decreased by 25.3% in 30 years.

Although the goal of this report does not include an analyses of suicide ideation, motives for suicide, or general mental health in college campuses, this report highlighted important trends on deaths by suicide in college students at Big Ten universities. The updated suicide rate of 5.6/100,000 in college students and the trends on deaths by suicide identified should aide in guiding conversations about mental health trends in college campuses with more accurate information. The 25.3% decrease in the suicide rate in Big Ten universities can be evidence of the success of suicide prevention efforts taken by college campuses and counseling centers at Big Ten universities. The higher suicide rates found in this report for certain populations, such as male students and Asian and Black students, should be further explored and should be taken into consideration as

universities and counseling centers continue their active efforts in suicide prevention education.

REFERENCES