Equity, Diversity and Inclusion: Mental health in medical and biomedical doctoral students during the 2020 COVID-19 pandemic and racial protests

Office of Medical Education, University of North Carolina at Chapel Hill School of Medicine, United States
Department of Biochemistry and Biophysics, University of North Carolina at Chapel Hill School of Medicine, United States

Sep 6, 2022

https://doi.org/10.7554/eLife.69960

Open access
Copyright information

Download
Cite
CommentOpen annotations (there are currently 0 annotations on this page).
Share

3 figures, 1 table and 3 additional files

Figures

Figure 1

Download asset Open asset

Logistic regression summary tables and correlation scatterplot matrix.

Logistic regression tables show main effects and interactions for each of the primary factors (Program, Year, HE-Gender, HE-Race/Ethnicity) across four measures of mental health (Depression, Anxiety, Hazardous Alcohol Use, Problems with Drug Use) and three measures of suicidal ideation (Ever, While Enrolled, In The Last 12 Months); the correlation matrix scatter plots illustrate the relationships between the four mental health measures. All percentages were calculated out of total valid responses; see Figure 2 for percentages. Mental health outcomes were coded into a bivariate 0/1 indicating the absence or presence of the respective symptoms for depression and anxiety (see Measures). For suicidal ideation, response options included ‘Yes’ (indicating suicidal ideations) or ‘No’ (indicating no suicidal ideations) for each of the three categories (see Methods for exact question wording). (**A & B**) Logistic Regression. A significant odds ratio greater than 1 indicates that the target category is more likely than the comparison group to indicate a higher score on that variable, whereas a significant odds ratio of less than 1 indicates that the target category is less likely than the comparison group to indicate a higher score on that variable. Nonsignificant effects suggest odds ratios do not show a difference greater than expected by chance (~1). If the 95% confidence interval includes 1, then the odds ratio does not differ more than expected by chance (e.g., OR = 0.98, CI95%=0.83–1.14 means there is not a significant effect of Year for Suicidal Ideation Ever); if the confidence interval does not include 1, then it differs more than expected by chance. (C) Scatterplots of the relationship between variables of interest (Depression, Anxiety, Problems with Drug Use and Hazardous Alcohol Use) displayed include graphical representations in a matrix format.

Figure 2

Download asset Open asset

Levels of depression, anxiety, problems with drug use, hazardous alcohol use and suicidal ideation based on type of training and historically excluded status linked to race/ethnicity.

(A) Percentages of medical (MD) and biomedical doctoral (PhD) students reporting the presence (dark grey) or absence (light grey) of symptoms/problematic use pertaining to depression, anxiety, problems with drug use and hazardous alcohol use. (B) Percentages of medical (MD) and biomedical doctoral (PhD) students in 2019 and 2020 reporting having had (dark grey) or not having had (light grey) suicidal thoughts ever, in the last 12 months or while enrolled. (C) Percentages of historically excluded and non-historically excluded students based on race and ethnicity (HE-RE vs HE-NRE) reporting the presence (dark grey) or absence (light grey) of symptoms and problematic use pertaining to depression, anxiety, problems with drug use and hazardous alcohol use. HE-RE students showed higher rates of depression in both 2019 (60% of HE-RE, 51% of NHE-RE) and 2020 (48% of HE-RE, 35% of NHE-RE). This pattern was also seen for anxiety (in 2019: 60% of HE-RE, 53% of NHE-RE; in 2020: 51% of HE-RE, 38% of NHE-RE). (D) Percentages of historically excluded and non-excluded students based on race and ethnicity (HE-RE vs HE-NRE) reporting, in 2019 and 2020, having had (dark grey) or not having had (light grey) suicidal thoughts ever, in the last 12 months or while enrolled. Historically excluded students exhibit higher rates of suicidal ideation, in particular ‘while enrolled’ and ‘during the last 12 months’ (2019, positive responses for ‘while enrolled’: 23% of HE-RE, 15% of NHE-RE; positive responses for ‘during the last 12 months’: 19% HE-RE, 11% NHE-RE; 2020, positive responses for ‘while enrolled’: 18% of HE-RE, 10% of NHE-RE; positive responses for ‘during the last 12 months’ 17% HE-RE, 12% NHE-RE). All percentages were calculated out of total valid responses. HE-RE students were coded as such if they indicated that they belonged to historically excluded racial or ethnic categories (e.g., African American/Black, Hispanic/Latinx; see Methods for details); NHE-RE included students who did not indicate a marginalized racial/ethnic identity.

Figure 3

Download asset Open asset

Levels of depression, anxiety, problems with drug use, hazardous alcohol use, and suicidal ideation for historically excluded/non-excluded populations based on gender and sexual orientation.

Graphical representation displaying the percentages for each of the two primary factors in this figure (Gender and Sexual Orientation) across the four measures of mental health and three measures of suicidal ideation for populations historically excluded on the basis of sexual orientation or gender (HE-SO and HE-G). All percentages were calculated out of total valid responses. Historically excluded versus non-historically excluded students based on sexual orientation (HE-SO vs. NHE-SO) were compared. Members of the HE-SO group were classified as such if they indicated an excluded identity (e.g., Lesbian, Bisexual, Queer – LGBQ+ sexual orientation; see Methods for details); NHE-SO included any students who did not indicate a historically excluded identity. Historically excluded on the basis of gender included women; non-historically excluded students by gender included men (HE-G vs. NHE-G). (A) Percentages of LGBQ+ (HE-SO) and non-LGBQ+ (NHE-SO) men and women (NHE-G vs. HE-G) reporting the presence (dark grey) or absence (light grey) of symptoms and problematic use pertaining to depression, anxiety, problems with drug use and hazardous alcohol use. (B) Percentages of LGBQ+ (HE-SO) and non-LGBQ+ (NHE-SO) men and women (NHE-G vs. HE-G) reporting having had (dark grey) or not having had (light grey) suicidal thoughts ever, in the last 12 months or while enrolled.

Tables

Table 1

Coding of the four measures of mental health used in the analyses.

Each measure of mental health (depression, anxiety, hazardous alcohol use, and problems related to substance use) was examined and coded according to its respective validated scale (PHQ-9, GAD-7, AUDIT, DAST-10). The scores were recoded as interim measures based on symptom severity, from 0 to 4. These were then transferred into a clinically meaningful bivariate category (no symptoms, 0; any symptoms, 1) to facilitate the planned analysis (bivariate logistic regression).

Measures	No Symptoms(Coded as 0)	Any Symptoms(Coded as 1)
	Interim Recoded as 0	Interim Recoded as 1	Interim Recoded as 2	Interim Recoded as 3	Interim Recoded as 4
PHQ-9	1–4: Minimal or No Depression	5–9: Mild Depression	10–14: Moderate Depression	15–19: Moderately Severe Depression	20–27: Severe Depression
GAD-7	0–4: Minimal Anxiety	5–9: Mild Anxiety	10–14: Moderate Anxiety	15–21: Severe Anxiety
AUDIT	0–7: Low Risk	8–15: Hazardous	16–19: Harmful	20+: Possible Dependence
DAST-10	0: No problems	1–2: Low Problems	3–5: Moderate Problems	6–8: Substantial Problems	9–10: Severe Problems