This study was designed to investigate patterns of substance use among people who use drugs during the COVID-19 pandemic. Participants were recruited through community-based harm reduction organizations across Austin, El Paso, and San Antonio, Texas, and social media and Craigslist advertisements. Eligibility criteria were: 1) 18 years or older; 2) resident of Texas; 3) self-reported use of illicit drugs at least once in the previous month (excluding marijuana use only) and 4) ability to read and speak English.
Participants
A total of 214 participants were enrolled into the study (61% women, on average 33 years of age at enrollment) and completed three online surveys. Three-month follow-up surveys were completed by 188 participants (88% of sample) and 179 participants (84% of sample) completed six-month follow-up surveys. Across all participants and time points, a total of 572 participant surveys were completed for an average of 2.7 surveys per participant. To measure suicide ideation, at each time point participants were asked on a 5-point scale how much they had been distressed or bothered by thoughts of ending their life during the past two weeks (1=not at all; 2=a little bit; 3=moderately; 4=quite a bit; 5=extremely). Individuals were stratified into two groups: those who reported being distressed or bothered by thoughts of ending their life at any point during the six-month follow-up period (endorsing a 2 or higher at any point of assessment during the follow-up period), compared to those that did not report suicide ideation at any point during follow-up (endorsing a 1 at each assessment).
Quantitative Data Collection
Measurements
Demographics Characteristics
Twenty-one questions regarding the participants’ sociodemographic characteristics were created by the authors and collected at the baseline survey. Questions included, but were not limited to, age, gender, race/ethnicity, and sexual orientation.
COVID-Related Mental and Behavioral Health
Data regarding COVID-related health impacts were captured using the CoRonavIruS Health and Impact Survey [CRISIS; www.crisissurvey.org;40].
COVID-19 Exposure/impact Status (Past Two Weeks):
Participants were asked to report possible exposure to Coronavirus/COVID-19, possible symptoms of COVID-19, family member diagnosis of COVID-19, and whether there had been any impacts on family members such as hospitalization, quarantine, and job loss because of COVID-19 during the past two weeks.
COVID Worries (Past Two Weeks):
Participants reported on a 5-point Likert scale (ranging from 1=not worried at all to 8=extremely worried) how worried they have been during the past two weeks about personal infection, friends and family being infected, and possible impacts on physical and mental health, as well as time spent reading or talking about COVID-19, and hope that the pandemic will end soon.
Substance Use (Past Two Weeks):
Participants reported how frequently they engaged in using alcohol, tobacco, vaping, opiates, marijuana, and other drug use (cocaine, crack, amphetamine, methamphetamine, hallucinogens, or ecstasy). The survey had eight frequency options ranging from 1=not at all to 8=more than once a day.
DSHS data on COVID fatalities per day
The number of COVID-19 deaths per day for the state of Texas was obtained from the Texas Department of State Health Services (DSHS) website41. This data was pulled in the form of .csv documents and then cross-referenced to match the respective day of survey completion for each participant, across the longitudinal study.
Mood and Trauma Assessments
Depression
The Center for Epidemiologic Studies Depression Scale (CES-D42) is a 20-item self-report instrument with high internal consistency (α=0.85) that measures depression symptoms such as feeling lonely, loss of appetite, and crying. Participants reported on how often they experienced symptoms on a 4-point Likert scale (0=rarely or none of the time/less than 1 day to 3=most or all of the time/5-7 days). Total scores range from 0 to 60 and scores of 16 or higher typically indicate risk for clinical depression43.
Anxiety
The Generalized Anxiety Disorder (GAD-744) questionnaire is a seven-item, self-report instrument designed to measure symptoms of anxiety such as nervousness, feeling edgy, and worrying during the past two weeks. On a scale from 0=not at all to 3=nearly every day, participants rated their frequency of anxiety symptoms. Total scores range from 0 to 21. Cut-off point scores for mild, moderate, and severe anxiety are 5, 10, and 15 respectively. The GAD-7 has high internal consistency (α=0.92).
Posttraumatic Stress Disorder
The Primary Care PTSD Screen for DSM-5 (PC-PTSD-545) is a five-item instrument with excellent diagnostic accuracy (Area Under the Curve/AUC=0.94). Participants are asked to respond yes (=1) or no (=0) to experiences of PTSD symptoms such as having nightmares, being on guard, and feeling numb or detached from people. Total scores range from 0 to 5 with a cut-off point score of 3 indicating probable PTSD diagnosis. Internal consistency is incalculable due to the binary nature of the response options.
Perceived Childhood Trauma
The Childhood Traumatic Events Scale (CTES46) was used to measure perceived childhood trauma. It has six options for trauma: “death of a very close friend or family member,” “major upheaval between your parents,” “traumatic sexual experience,” “victim of violence,” “extreme illness or injury,” and “other major upheaval.” Participants are instructed to 1) select the types of traumas they experienced before the age of 17; 2) indicate their age at the time of occurrence; 3) rate the severity of the trauma on a 7-point Likert scale (range from 1=not at all traumatic to 7=extremely traumatic); 4) rate the degree of confiding to others at the time on a 7-point Likert scale (range from 1=not at all to 7=confided a great deal). Internal consistency is incalculable due to the binary nature of the response options, but the scale is sensitive to clinical symptoms of early life trauma47.
Statistical Approach
Demographics
Differences in age, sex, race/ethnicity, and sexual orientation between individuals who presented with suicide ideation at any point during the follow-up period and individuals who never presented with suicide ideation during follow-up were investigated with student t-test or chi-squared test as appropriate.
Between Group Differences: Individuals Reporting Suicide Ideation (Over Course of 6 Month Follow-up) compared to Individuals who did not Report Suicide Ideation Over Follow-up
Mixed models were used to model main effects of sex, group (presenting with suicide ideation at some point vs. not endorsing suicide ideation at any point during study assessments), and sex by group interactions, with age included as a covariate and time of survey (baseline, three-month, six-month follow-up) as a within subject variable. Dependent variables investigated include prior 2-week substance use (including alcohol, vaping, tobacco, marijuana, opiates, and other substances), depression, anxiety, and PTSD symptoms, and perceived severity of childhood trauma. Following a significant group by sex interaction, models were stratified by sex and repeated. Following no significant group by sex interactions, the interaction term was dropped from the model and main effects of group or sex investigated. A parallel model was ran investigating between group differences in COVID worries and COVID exposure/impact, but additionally controlling for number of fatalities that occurred in Texas the day the CRISIS survey was filled out by a respective participant. Findings were considered significant at p<.0042. (Bonferroni correction for 12 dependent variables being investigated).
Predictors of Fluctuations in Suicide Ideation in Individuals who Report Varying Levels of Suicide Ideation over Follow-up
Variables that significantly differed between individuals with suicide ideation over follow-up, compared to those without, were further investigated to see if fluctuations in these variables relate to fluctuations in suicide ideation over the follow-up period. This analysis was restricted to n=86 individuals in the group that showed suicide ideation over follow-up and had more than one survey completed with different levels of ideation reported across surveys (e.g., no ideation at baseline and ideation at a different time point or an individual that rated ideation as a 5 at one timepoint and ideation as a 3 at another). On average, each of these 86 participants completed 2.8 surveys providing 244 surveys included in this analysis. Age, sex, prior two-week alcohol, tobacco, vaping, opiate, other substance use, CRISIS worry, COVID exposure/impact, number of fatalities on the day the survey was completed, and PTSD symptoms were included in the model with time of survey (baseline, three-month, six-month follow-up) as a within subject variable, and suicide ideation score as a continuous dependent variable. Suicide ideation severity score was not normally distributed and was log transformed. Depression and anxiety symptom scores were not included in this model as these two variables highly correlated with PTSD symptom scores. We prioritized PTSD symptoms in these models since we observed a sex by suicide ideation group interaction on PTSD symptoms and investigating sex differences is a primary aim of this study. This model was repeated stratified by sex to explore for sex differences in risk factors contributing to fluctuation in suicide ideation. For this exploratory analysis, findings were considered significant at p<.05.