Posttraumatic stress disorder (PTSD) is a mental health condition that may develop in response to past trauma (American Psychiatric Association, 2022). Military veterans are at high risk of developing PTSD with 7.6–13.1% of regular force Canadian Armed Forces (CAF) veterans meeting PTSD criteria compared to only 1.3–5.2% of the general Canadian population after adjusting for age and sex differences (Gauvin et al., 2021; Thompson et al., 2015). This increased risk of PTSD is likely the result of military-related trauma, with 97% of Canadian veterans with PTSD attributing their diagnosis to prior military service (Thompson et al., 2011). Higher exposure to certain traumatic events may cause PTSD presentation in veterans to differ from civilians (Lehavot et al., 2018). For example, combat trauma may produce more negative alterations in cognition and mood than other forms of trauma (Graham et al., 2016). PTSD in veterans is also associated with a range of negative outcomes such as chronic pain, respiratory issues, sleep disturbance, lower life satisfaction, poorer relationship functioning, occupational impairment, and other forms of psychological disorder (Benedict et al., 2020; El-Gabalawy et al., 2018; Schnurr et al., 2009; Smith et al., 2016). For example, United States (US) veterans with PTSD were significantly more likely to have an alcohol use disorder (55% vs. 33%), substance use disorder (SUD; 27% vs. 8%), mood disorder (62% vs. 16%), and/or anxiety disorder (49% vs. 11%) compared to veterans without PTSD (Smith et al., 2016). Recently, Canadian veterans have expressed interest in using cannabis as a form of treatment for PTSD (Atlas Institute, 2022). However, cannabis is not currently mentioned or recommended in clinical practice guidelines for the treatment of PTSD (American Psychological Association, 2017; US Department of Veterans Affairs & Department of Defense, 2017).
Cannabis use is common among Canadian veterans, with 47% of veterans using cannabis consistently for 1 or more years (Sterniczuk & Whelan, 2016). Currently, Veterans Affairs Canada (VAC) will reimburse Canadian veterans for up to 3 g of medicinal cannabis per day and up to 10 g per day if approved by a medical specialist (VAC, 2019). Since Canada legalized recreational cannabis use in 2018 (Cannabis Act, 2018), funding for medical cannabis reimbursement by VAC has increased substantially, growing from $51 million in the 2017–2018 fiscal year to $168 million in the 2022–2023 fiscal year (VAC, 2023). However, evidence for the effectiveness of cannabis in treating PTSD is equivocal (Bonn-Miller et al., 2021; Nacasch et al., 2023; Wilkinson et al., 2015).
The pharmacological effects of cannabis are complex, with cannabis containing hundreds of potentially psychoactive compounds (Klumpers & Thacker, 2019). One of the most abundant of these compounds is the cannabinoid delta-9-tetrahydrocannabinol (THC) which appears to be responsible for cannabis’s intoxicating effects (Drennan et al., 2021). THC may produce undesired effects such as temporary cognitive impairment, anxiety, psychomimetic behaviour, and addiction (Arkell et al., 2019; Morgan et al., 2018; Zehra et al., 2018). However, certain forms of THC may have medicinal use for conditions such as insomnia and neuropathic pain (Jetly et al., 2015; Sainsbury et al., 2021). Another cannabinoid of interest is cannabidiol (CBD), a non-intoxicating anxiolytic that produces little to no adverse reactions (Han et al., 2024; Lo et al., 2024). The effects of THC and CBD may differ with dosage, with one study finding a parabolic relationship between THC levels and acute pain relief (Wallace et al., 2007). Cannabinoids may also impact one another’s effects. For example, low and high doses of CBD may enhance and ameliorate the intoxicating effects of THC, respectively (Solowij et al., 2019). Conversely, CBD may not impact THC’s other effects, such as cognitive impairment and psychomimetic behaviour (Arkell et al., 2019; Morgan et al, 2018). The onset and duration of cannabis’s effects may differ based on the route of administration. When cannabis is inhaled, THC quickly reaches the brain producing a high within minutes (Fabritius et al., 2013). Conversely, ingestion produces a delayed high that tends to last longer than inhalation (Schlienz et al., 2020). Ingested and inhaled cannabis may also produce differential psychoactive effects, as ingested THC is more readily metabolized into 11-hydroxy-THC, a stronger form of THC (Klumpers & Thacker, 2019). Another route of administration is absorption, such as in topical creams. These products are often taken medicinally, though their efficacy is unconfirmed (Mahmood et al., 2022). As opposed to inhaled and ingested cannabis, there is no evidence that topically absorbed cannabis produces any psychoactive effects.
Canadians with PTSD are twice as likely to use cannabis compared to those without PTSD (28.2% vs. 11.2% respectively; Lake et al., 2020). Some theorists suggest that those with PTSD may engage in cannabis use as a form of self-medication to temporarily alleviate PTSD symptoms (Khantzian, 1985, 1997). Some evidence supports this theory, with one study showing a reduction in self-reported PTSD symptoms by up to 50% immediately after the use of medicinal cannabis (LaFrance et al., 2020). However, other evidence suggests that those who use cannabis may be more likely to develop PTSD or have increased PTSD severity. For example, longitudinal studies on US veterans have found that cannabis use predicted greater PTSD severity, violent behaviour, and suicidality while cannabis cessation was associated with a reduction in PTSD severity, violent behaviour, and suicidality (Manhapra et al., 2015; Metrik et al., 2022; Wilkinson et al., 2015). However, some of these findings could be explained by those with higher severity PTSD seeking out cannabis rather than cannabis worsening PTSD symptoms. Medicinal cannabis use may also put veterans at risk of developing a cannabis use disorder, as veterans with PTSD are at risk of developing SUDs (Smith et al., 2016).
Conversely, other research suggests that cannabis use may help alleviate PTSD symptoms. Multiple studies have reported that cannabis and nabilone, a synthetic form of THC, can improve sleep quality among those with PTSD (Fraser, 2009; Jetly et al., 2015; Nacasch et al., 2023). Additionally, longitudinal studies have observed that medicinal cannabis use is associated with improved quality of life, as well as reduced depression, anxiety, and PTSD severity at 4 and 10-month follow-ups (Chan et al., 2017; Drost et al., 2017). A retrospective chart review of Canadian veterans found that medicinal cannabis use was associated with a reduction in PTSD severity, suicidal ideation, and pain (Smith et al., 2017). Both medicinal THC and CBD for the treatment of PTSD are well tolerated and have resulted in reduced PTSD severity when paired with other treatments (Elms et al., 2019; Roitman et al., 2014). Finally, cannabis may offer a more tolerable replacement for medications with more unwanted side effects such as antidepressants (Smith et al., 2017; Storey et al., 2023).
What is causing these equivocal findings? One possibility is that the relationship between cannabis and PTSD is not just based on whether cannabis is used, but how cannabis is used. For example, much of the reviewed research on recreational cannabis found a positive correlation between cannabis use and PTSD severity (e.g., Manhapra et al., 2015; Metrik et al., 2022; Wilkinson et al., 2015), while research on medicinal cannabis found a negative correlation (e.g., Chan et al., 2017; Drost et al., 2017; Smith et al., 2017). Thus, medicinal and recreational use may have differential relationships with PTSD. It is also possible that other cannabis use characteristics such as route of administration, THC to CBD ratio (TBD:CBD), and grams of cannabis consumed per day (grams/day) could modify the relationship between cannabis use and PTSD.
To our knowledge, no studies have yet investigated whether differences in cannabis use characteristics are associated with differences in PTSD outcomes. To address this research gap, we performed a secondary analysis using data from a large cross-sectional study of medicinal cannabis use among Canadian veterans living with chronic pain. Chronic pain (i.e., pain that lasts 3 or more months) and PTSD are often comorbid with one another (Kind & Otis, 2019). Currently, medicinal cannabis is a conditionally recommended treatment for certain types of chronic pain (Allan et al., 2018). Due to its potential to treat both PTSD and chronic pain, some Canadian veterans have expressed interest in using cannabis to treat comorbid PTSD and chronic pain simultaneously (Storey et al., 2023). Due to the common occurrence of PTSD and cannabis use among those with chronic pain, we believed this sample would prove adequate for our investigation.
The objectives of the current study were: to determine if differences in cannabis use and cannabis use characteristics are associated with differences in Canadian veterans’ PTSD severity; and to determine if differences in cannabis use and cannabis use characteristics are associated with differences in PTSD sub-dimensions and PTSD-related conditions such as depression, anxiety, and insomnia (Benedict et al., 2020; Smith et al., 2016). Specifically, the present study will address the following research questions:
Is cannabis use status (never, past, or current user) associated with PTSD severity, psychological distress, and/or insomnia?
Is cannabis use status associated with specific outcomes among the Diagnostic and statistical manual of mental disorders (DSM-5; American Psychiatric Association, 2022) PTSD symptom clusters?
Among cannabis users, are different cannabis use characteristics (i.e., medicinal vs. recreational use, route of administration, THC:CBD ratio, and grams/day) associated with PTSD severity, psychological distress, and/or sleep insomnia?
Among cannabis users, are different cannabis use characteristics associated with specific outcomes among the DSM-5 PTSD symptom clusters?
Methods
Participants
Participants were recruited through: the Chronic Pain Center of Excellence for Canadian Veterans mailing list, comprised of individuals who self-identified as Canadian veterans and had agreed to participate in research; and targeted advertisements dispersed on Facebook via the Sussex Strategy Group. All recruitment ads and emails were provided in English and French and included a direct link to the study’s Qualtrics XM page. Participants received either a $50 gift card for Amazon or Tim Hortons after completing the survey as compensation. Participants were self-identified Canadian veterans (i.e., former members of the CAF who completed basic training and were discharged) with chronic pain who were 18 years of age or older and were able to complete a questionnaire over the Internet. Veterans without access to the Internet or who could not read and write in English or French were not eligible for participation.
Measures
The survey used in this study consisted of a range of self-report measures covering participants’ demographic characteristics, military service, medical history, and cannabis use characteristics. Several validated psychometric measures of pain, depression, anxiety, PTSD, and sleep disturbance were included in the survey. The psychometric measures relevant to the current analysis were the Primary Care PTSD Screen for DSM-5 (PC-PTSD-5), the PTSD Checklist for DSM-5 (PCL-5), the Insomnia Severity Index (ISI), the Patient Health Questionnaire (PHQ-9), and the General Anxiety Disorder Scale (GAD-7). All survey items were optional.
Cannabis Use Characteristics
Several characteristics of cannabis use were identified as factors that could potentially mediate the relationship between PTSD and cannabis use. The first was whether a veteran used cannabis medicinally or recreationally, as cannabis’s effects are often different between studies that focused on recreational use (e.g., Wilkinson et al., 2015) versus medicinal use (e.g., Drost et al., 2017; Smith et al., 2017). However, cannabis use is often stigmatized (Storey et al., 2023) which may motivate those who use cannabis recreationally to justify their use by claiming it as medicinal. To control for this possibility, recreational and medicinal use was determined based on whether participants reported having a prescription for cannabis. The second variable of interest was route of administration, as this can impact the timing of drug action, the potential for intoxication, and the way cannabinoids are metabolized (Fabritius et al., 2013; Klumpers & Thacker, 2019; Schlienz et al., 2020). The third variable of interest was the THC:CBD ratio that a veteran primarily uses, as THC and CBD have differential effects, and augment the effects of one another (Solowij et al., 2019). The final variable studied was grams/day of cannabis consumption, as cannabis’s effects may change based on dosage (Wallace et al., 2007).
PC-PTSD-5
The screening question from the PC-PTSD-5 was used to determine if participants met criterion A for a diagnosis of PTSD as defined by the DSM-5 (Prins et al., 2016). Specifically, this question asked: “Sometimes things happen to people that are unusually or especially frightening, horrible, or traumatic. For example: a serious accident or fire, a physical or sexual assault or abuse, an earthquake or flood, a war, seeing someone be killed or seriously injured, having a loved one die through homicide or suicide. Have you ever experienced this kind of event? Yes or No”. A response of ‘Yes’ indicated that a participant met criterion A for a diagnosis of PTSD.
PCL-5
The PCL-5 is a 20-item, self-report measure designed to screen, provisionally diagnose, and monitor the severity of PTSD symptoms (Blevins et al., 2015). Each item on the PCL-5 corresponds to a symptom of PTSD listed in the DSM-5, with items 1–5 corresponding to intrusive symptoms (criterion B); items 6–7 corresponding to avoidance symptoms (criterion C); items 8–14 corresponding to negative alterations in cognition and mood (criterion D); and items 15–20 corresponding to alterations in arousal and reactivity (criterion E). Respondents rate how much they were bothered by each symptom in the past month on a scale of 0–4, with higher scores indicating greater symptom severity. PCL-5 scores range from 0–80, with scores of 31–33 having been recommended as PTSD cutoff scores for veteran populations (Bovin et al., 2016). The PCL-5 has been shown to have good internal consistency (α = .94), test-retest reliability (r = .82), convergent validity, and divergent validity in veteran populations (Blevins et al., 2015; Bovin et al., 2016).
PHQ-ADS
The Patient Health Questionnaire Anxiety and Depression Scales (PHQ-ADS) is a 17-item composite questionnaire comprised of the PHQ-9 and the GAD-7 (Kroenke et al., 2016). Combining the PHQ-9 and GAD-7 total scores into a single composite score is theoretically sound and provides several advantages. For example, depression and anxiety are frequently comorbid, correlate highly, are effectively treated by many of the same interventions, and are trans-diagnostically encompassed by an overarching construct of psychological distress (Kroenke et al., 2016). From a methodological standpoint utilizing the PHQ-ADS also allowed us to increase the power of the current study’s analyses by lowering the number of outcome variables. PHQ-ADS scores range from 0–48, with scores of 10, 20, and 30 corresponding to mild, moderate, and severe symptoms of psychological distress respectively. The PHQ-ADS has been shown to have good internal consistency (α = .88) and convergent validity with other measures of depression and anxiety such as the Short Form Health Survey (Kroenke et al., 2016).
The PHQ-9 is used to assess depressive symptoms (Kroenke et al., 2001), while the GAD-7 is designed to detect generalized anxiety disorder (GAD) symptoms (Spitzer et al., 2006). The PHQ-9 and GAD-7 are self-administered questionnaires that ask respondents how often they have experienced a certain depressive or anxious symptom during the last 2 weeks, respectively. Each question is answered on a scale of 0–3 with higher scores representing more severe symptoms. PHQ-9 and GAD-7 scores range from 0–27, and 0–21, respectively. PHQ-9 and GAD-7 scores of 5, 10, and 15 correspond to mild, moderate, and severe symptoms of depression and anxiety, respectively. Both tests have good internal consistency (PHQ-9, α = .89; GAD-7, α = .92), test-retest reliability (PHQ-9, r = .84; GAD-7,ICC = .83), and correlate strongly with functional impairment (Kroenke et al., 2001; Spitzer et al., 2006).
ISI
The ISI is a 7-item self-report measure used to assess insomnia (Bastien et al., 2001). Respondents answer items related to their insomnia problems, sleep satisfaction, insomnia-related functional impairment, noticeability of sleep problems to others, and worry caused by sleep problems on a scale of 0–4, with higher scores indicating more severe sleep problems. Total scores range from 0–28, with a score of 15 or above indicating clinically significant insomnia (Bastien et al., 2001). The ISI shows good internal consistency (α = .74) and convergent validity with other insomnia measures, such as sleep diaries (Morin et al., 2011).
Statistical Analysis
Since the purpose of this study was to investigate how cannabis use and cannabis use characteristics impact PTSD-related outcomes in Canadian veterans, only participants who reported experiencing a traumatic event via the PC-PTSD-5 screening question were included in the analyses. We did not require participants to meet a certain cutoff score on the PCL-5 for inclusion, as the effects of trauma exist on a spectrum and cannot be cleanly delineated by a single cutoff score.
Basic descriptive analysis was performed on participants’ demographic features, military history, and cannabis use. Due to violations of univariate and multivariate normality, robust non-parametric statistical methods were used for our MANOVAs (Friedrich et al., 2019; Wilcox, 2017). Non-parametric MANOVAs were performed with the R package “MANOVA.RM” (Friedrich et al., 2019). Specifically, a modified MANOVA-type statistic (MATS) was run using a parametric bootstrap procedure to calculate test statistics and p values. If a MANOVA had any significant main effects or interactions (p < .05), additional MANOVAs were performed separately for each outcome variable with the significant factor or interaction included as the only predictor variable. The Benjamini-Hochberg method was used to adjust for the false discovery rate when performing these additional MANOVAs (Benjamini & Hochberg, 1995). If the outcomes of the additional MANOVAs were significant (p < .05) after the Benjamini-Hochberg adjustment, a Tukey post hoc test was performed. MANOVA was chosen over regression as it allows for the testing of multiple dependent variables simultaneously while accounting for interaction effects and correlations among predictor and outcome variables. It also protects against inflated type I error rates due to multiple tests of correlated outcome variables (Tabachnick & Fidell, 2019).
To address the first research question, a 3 × 2 × 2 MANOVA was conducted to infer whether PTSD severity, psychological distress, and insomnia varied as a function of cannabis use status (never, past, or current), talk therapy attendance (yes or no), and sex (male or female). To address the second research question, another 3 × 2 × 2 MANOVA was conducted to infer whether individual PTSD symptom clusters (B, C, D, E) varied as a function of the same predictor variables. Since talk therapy attendance could have a large impact on PTSD-related outcomes, talk therapy attendance (yes or no) was included as a cofactor for these analyses.
For research questions three and four, only the subset of participants who endorsed current cannabis use were included. To address the third research question four separate MANOVAs were conducted to infer whether PTSD severity, psychological distress, and insomnia varied as a function of cannabis prescription (yes or no), route of administration (inhaled, ingested, absorbed, or multiple routes), THC:CBD ratio (low, equal, or high), and grams/day of cannabis consumption (<3 g, 3 g, or >3 g). Sex (male or female) was included as a predictor in each of the four MANOVAs. To address research question four, four additional MANOVAs were conducted to infer whether PTSD symptom clusters (B, C, D, E) varied as a function of the same predictor variables. Talk therapy attendance was not included as a factor for these analyses due to cell restraints.
The md.pattern function in the MICE package was used to assess the pattern of missingness in the outcome variables (e.g., PCL-5, ISI, etc.). There were 531 missing values across all variables in the dataframe. The PCL-C variable had the fewest number of missing values (74), and the ISI variable had the largest number (125). Data were determined to be missing completely at random according to Little’s (1988) MCAR Test, X2 (63, N = 509) = 61.708, p = .522. Using R version 4.3.1 missing values for PHQ-ADS, ISI, and the PCL sub-scores were single imputed via the “imputeEM()” function from the “mvdalab” package (Afanador et al., 2022). The “imputeEM()” function iteratively updates missing values through an expectation-maximization (EM) algorithm. Missing values for categorical variables were single imputed using the “mice()” function from the “Multivariate Imputation by Chained Equations” (“MICE”) package (Van Buuren & Groothuis-Oudshoorn, 2011). Missing values for sex were not imputed as there were no variables that were strong predictors of sex. Thus, participants who did not disclose their sex during the survey (n = 4) were excluded from the MANOVAs due to cell size limitations.
Results
Overall, 600 veterans responded to the original survey. Seven respondents were removed due to low item response (i.e., <25%), clear misrepresentation (e.g., reporting ‘years of service’ greater than age), or failure to report their cannabis use status. Of the 593 participants remaining, 513 (86%) reported experiencing a traumatic event via the PC-PTSD-5 screening question. Only these participants were included in the following secondary analyses.
Descriptive Analysis
Participants’ demographic characteristics and how demographic characteristics differ between cannabis use groups can be found in Tables 1 and 2. Of note, no participants in the survey described their sex as intersex or “other,” their gender as two-spirited, or their race/ethnicity as Arab.
Table 1
Frequency Count and Percentage of Categorical Demographic Variables.
| VARIABLE | NON-CANNABIS USERS | NO CANNABIS USE IN PAST 6 MONTHS | CURRENT CANNABIS USERS | COMBINED | ||||
|---|---|---|---|---|---|---|---|---|
| N | % | N | % | N | % | N | % | |
| Total | 115 | 22.42 | 70 | 13.65 | 328 | 63.94 | 513 | 100 |
| Sex | ||||||||
| Female | 21 | 18.26 | 14 | 20.00 | 45 | 13.72 | 80 | 15.59 |
| Male | 93 | 80.87 | 56 | 80.00 | 280 | 85.37 | 429 | 83.63 |
| Non-disclosed | 1 | 0.87 | 0 | 0.00 | 3 | 0.91 | 4 | 0.77 |
| Race | ||||||||
| White | 104 | 92.86 | 64 | 92.75 | 290 | 89.78 | 458 | 90.87 |
| Black | 0 | 0.00 | 0 | 0.00 | 3 | 0.93 | 3 | 0.60 |
| Indigenous | 5 | 4.46 | 4 | 5.80 | 21 | 6.50 | 30 | 5.95 |
| Asian/Pacific Islander | 0 | 0.00 | 0 | 0.00 | 1 | 0.31 | 1 | 0.20 |
| Hispanic | 0 | 0.00 | 0 | 0.00 | 1 | 0.31 | 1 | 0.20 |
| South Asian | 0 | 0.00 | 1 | 1.45 | 0 | 0.00 | 1 | 0.20 |
| Mixed race | 3 | 2.68 | 0 | 0.00 | 7 | 2.17 | 10 | 1.98 |
| Marital Status | ||||||||
| Single | 9 | 7.96 | 4 | 5.80 | 23 | 7.01 | 36 | 7.06 |
| Common law | 18 | 15.93 | 10 | 14.49 | 33 | 10.06 | 61 | 11.96 |
| Married | 64 | 56.64 | 43 | 62.32 | 203 | 61.89 | 310 | 60.78 |
| Divorced/separated | 17 | 15.04 | 9 | 13.04 | 62 | 18.90 | 88 | 17.25 |
| Widowed | 5 | 4.42 | 3 | 4.35 | 7 | 2.13 | 15 | 2.94 |
| Survey language | ||||||||
| English | 97 | 84.35 | 64 | 91.43 | 313 | 95.43 | 474 | 92.40 |
| French | 18 | 15.65 | 6 | 8.57 | 15 | 4.57 | 39 | 7.60 |
| Highest level of education | ||||||||
| No highschool degree | 15 | 13.39 | 6 | 8.57 | 26 | 7.98 | 47 | 9.25 |
| Highschool | 22 | 19.64 | 22 | 31.43 | 79 | 24.23 | 123 | 24.21 |
| College/trade school | 43 | 38.39 | 30 | 42.86 | 164 | 50.31 | 237 | 46.65 |
| Undergrad degree | 20 | 17.86 | 8 | 11.43 | 42 | 12.88 | 70 | 13.78 |
| Graduate degree | 12 | 10.71 | 4 | 5.71 | 15 | 4.60 | 31 | 6.10 |
| CAF branch | ||||||||
| Army | 71 | 62.83 | 42 | 60.87 | 199 | 62.58 | 312 | 62.40 |
| Navy | 21 | 18.58 | 8 | 11.59 | 57 | 17.92 | 86 | 17.20 |
| Air Force | 21 | 18.58 | 19 | 27.54 | 62 | 19.50 | 102 | 20.40 |
| Rank upon release | ||||||||
| Junior NCM | 43 | 38.39 | 29 | 42.03 | 162 | 52.43 | 234 | 47.76 |
| NCO | 39 | 34.82 | 32 | 46.38 | 117 | 37.86 | 188 | 38.37 |
| Commissioned | 30 | 26.79 | 8 | 11.59 | 30 | 9.71 | 68 | 13.88 |
| Employment | ||||||||
| Unemployed | 18 | 15.65 | 17 | 24.29 | 110 | 33.54 | 145 | 28.72 |
| Student/training | 1 | 0.87 | 1 | 1.43 | 6 | 1.83 | 8 | 1.56 |
| Employed | 27 | 23.48 | 21 | 30.00 | 53 | 16.16 | 101 | 19.69 |
| Retired | 69 | 60.00 | 31 | 44.29 | 159 | 48.48 | 259 | 50.49 |
| Disability | ||||||||
| No | 82 | 71.30 | 44 | 62.86 | 174 | 53.05 | 300 | 58.48 |
| Yes | 33 | 28.70 | 26 | 37.14 | 154 | 46.95 | 213 | 41.52 |
| Mental health condition | ||||||||
| No | 39 | 33.91 | 22 | 31.43 | 72 | 21.95 | 133 | 25.93 |
| Yes | 76 | 66.09 | 48 | 68.57 | 256 | 78.05 | 380 | 74.07 |
[i] Note. NCM = Non-commissioned member, NCO = Non-commissioned officer. All percentages are calculated as valid percentages (i.e., the proportion of non-missing cases that fall into a certain category).
Table 2
Mean and Standard Deviation of Continuous Demographic Variables.
| VARIABLE | NON-CANNABIS USERS | NO CANNABIS USE IN PAST 6 MONTHS | CURRENT CANNABIS USERS | COMBINED |
|---|---|---|---|---|
| Age | 60.07 (8.63) | 58.25 (8.59) | 57.35 (8.35) | 58.08 (8.50) |
| Year joined CAF | 1983 (10.50) | 1984 (10.21) | 1986 (10.36) | 1985 (10.40) |
| Year released from CAF | 2003 (15.49) | 2004 (13.93) | 2005 (12.67) | 2004 (13.50) |
| Total years of service | 20.43 (11.57) | 20.98 (11.12) | 19.72 (10.73) | 20.04 (10.96) |
| Years since release | 20.33 (15.49) | 19.20 (13.93) | 18.07 (12.67) | 18.71 (13.50) |
[i] Notes. Numerical data format: Mean (Standard Deviation). Missing cases were excluded.
Table 3 shows PTSD-relevant conditions that veterans reported using cannabis to treat, and how effective they found cannabis to be for these conditions. Of the 328 participants who used cannabis, 229 (70%) reported using cannabis for at least one mental health condition (i.e., PTSD, anxiety, and/or depression). Of the 328 participants who reported currently using cannabis, 262 (80%) reported using cannabis medicinally, 12 (4%) recreationally, 37 (11%) reported using cannabis both medicinally and recreationally, and 17 (5%) did not answer this question. Similarly, 253 (77%) participants who currently used cannabis reported having a cannabis prescription (pre-imputation). Additionally, the majority of those with a cannabis prescription (n = 246, 97%) reported knowing how much cannabis they were prescribed, with only a few users not knowing their prescribed dose (n = 7, 3%). Similarly, most current users reported knowing how much cannabis they used per day (n = 285, 87%), while some users reported not knowing how much they used (n = 33, 10%), and a few users did not wish to disclose this information (n = 10, 3%). Of those who knew how much cannabis they were prescribed per day (n = 256, 97%) the average amount prescribed per day was 4.55 g (Mode = 3.00g, SD = 2.60 g). Of those who reported how much cannabis they used per day (n = 284, 87%) the average amount of cannabis used per day was 3.52 g (Modes = 2.00 g and 3.00 g, SD = 2.82 g).
Table 3
Frequency and Impact of Cannabis Use on Mental Health Conditions in Canadian Veterans Who Use Cannabis, Have Chronic Pain, and Have a History of Trauma.
| CONDITION | HAS CANNABIS USE HELPED THE FOLLOWING CONDITIONS | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| CANNABIS USE FOR BELOW CONDITIONS | NO CHANGE | HELPFUL | MADE IT WORSE | |||||||
| N | % | N | % | N | % | N | % | |||
| PTSD | 174 | 58.59 | 40 | 23.95 | 127 | 76.05 | 0 | 0.00 | ||
| Depression | 142 | 47.81 | 14 | 10.00 | 125 | 89.29 | 1 | 0.71 | ||
| Anxiety | 185 | 62.29 | 15 | 8.15 | 164 | 89.13 | 5 | 2.72 | ||
| Sleep Problems | 228 | 76.77 | 17 | 7.56 | 207 | 92.00 | 1 | 0.44 | ||
[i] Note. All percentages are calculated as valid percentages (i.e., the proportion of non-missing cases that fall into a certain category).
Of the 328 participants who were current cannabis users, the vast majority reported knowing the THC:CBD ratio of the cannabis they used (n = 311, 95%). Specifically, 118 (36%) reported using cannabis with a high THC:CBD ratio, 139 (42%) reported using cannabis with an equal amount of THC and CBD, and 95 (29%) reported using cannabis with a low THC:CBD ratio.
Multivariate Analysis of Variance
The imputed mean and standard deviation of outcome variables related to research questions one and two can be found in Table 4. The imputed mean and standard deviation of the outcome variables related to research questions three and four can be found in Table 5. See Appendix A for the correlations between the outcome variables. See Appendix B for the individual cell values of each MANOVA.
Table 4
Mean and Standard Deviation of PCL-5, PHQ-ADS, ISI, and PCL-5 Symptom Cluster Scores across Predictor Variables in Canadian Veterans with Chronic Pain and a History of Trauma.
| VARIABLE | n | PCL-5 | PCL-B | PCL-C | PCL-D | PCL-E | PHQ-ADS | ISI |
|---|---|---|---|---|---|---|---|---|
| Cannabis use | ||||||||
| Never | 114 | 32.88 (17.57) | 7.92 (5.03) | 3.67 (2.32) | 11.00 (6.71) | 10.29 (5.32) | 20.55 (10.54) | 14.78 (6.14) |
| Non-current | 70 | 34.54 (17.72) | 7.75 (5.05) | 3.60 (2.59) | 12.11 (6.96) | 11.08 (5.04) | 21.58 (10.65) | 15.24 (4.91) |
| Current | 325 | 37.00 (16.82) | 8.51 (4.73) | 4.01 (2.23) | 12.79 (6.54) | 11.69 (5.20) | 22.47 (10.17) | 15.39 (5.88) |
| Sex | ||||||||
| Female | 80 | 39.09 (15.89) | 8.69 (4.87) | 4.32 (2.29) | 14.12 (6.15) | 11.97 (4.59) | 24.29 (8.99) | 16.77 (4.61) |
| Male | 429 | 35.12 (17.35) | 8.20 (4.84) | 3.80 (2.30) | 11.95 (6.71) | 11.17 (5.34) | 21.48 (10.51) | 14.94 (5.97) |
| Talk therapy attendance | ||||||||
| No | 244 | 28.63 (15.65) | 6.74 (4.55) | 3.12 (2.21) | 9.58 (6.05) | 9.19 (4.74) | 18.22 (9.39) | 14.19 (5.57) |
| Yes | 265 | 42.29 (15.88) | 9.69 (4.68) | 4.58 (2.17) | 14.79 (6.22) | 13.23 (4.91) | 25.32 (9.99) | 16.19 (5.88) |
[i] Notes. Numeical data format: Mean (Standard Deviation). Those who did not disclose their sex (n = 4) were excluded from analysis due to this group’s inclusion violating assumptions of the MATS (i.e., their inclusion created factor level combinations with less than two observations).
Table 5
Mean and Standard Deviation of PCL-5, PHQ-ADS, ISI, and PCL-5 Symptom Cluster Scores across Predictor Variables in Canadian Veterans Who Use Cannabis, Have Chronic Pain, and Have a History of Trauma.
| N | PCL-5 | PCL-B | PCL-C | PCL-D | PCL-E | PHQ-ADS | ISI | |
|---|---|---|---|---|---|---|---|---|
| Cannabis Prescription | ||||||||
| No | 68 | 32.48 (17.01) | 7.38 (4.77) | 3.68 (2.24) | 11.13 (6.37) | 10.29 (4.99) | 21.42 (9.46) | 14.54 (5.79) |
| Yes | 257 | 38.30 (16.60) | 8.82 (4.68) | 4.11 (2.22) | 13.27 (6.52) | 12.10 (5.21) | 22.78 (10.32) | 15.56 (6.04) |
| Preferred route of administration | ||||||||
| Inhaled | 100 | 36.08 (16.34) | 8.38 (4.55) | 3.90 (2.18) | 12.64 (6.30) | 11.16 (5.28) | 22.35 (9.79) | 14.77 (6.27) |
| Ingested | 101 | 36.07 (16.19) | 8.09 (4.49) | 3.71 (2.29) | 12.29 (6.33) | 11.98 (5.16) | 21.90 (9.83) | 15.61 (5.89) |
| Absorbed | 29 | 33.28 (20.62) | 7.06 (5.12) | 4.09 (2.56) | 11.18 (8.35) | 10.95 (5.66) | 20.96 (11.37) | 14.79 (6.02) |
| Multiple | 95 | 40.38 (16.48) | 9.56 (4.92) | 4.46 (2.08) | 14.09 (6.27) | 12.27 (5.03) | 23.75 (10.47) | 15.86 (5.82) |
| THC:CBD ratio | ||||||||
| Low | 82 | 37.55 (16.68) | 8.60 (4.79) | 4.02 (2.23) | 13.11 (6.59) | 11.83 (4.79) | 22.25 (8.66) | 16.08 (5.08) |
| Equal | 123 | 35.66 (17.12) | 8.57 (4.76) | 3.99 (2.27) | 11.91 (6.64) | 11.19 (5.43) | 21.62 (10.81) | 15.08 (6.26) |
| High | 120 | 38.23 (16.66) | 8.41 (4.70) | 4.06 (2.21) | 13.57 (6.33) | 12.18 (5.24) | 23.55 (10.37) | 15.13 (6.29) |
| Grams/day | ||||||||
| <3 grams | 152 | 35.37 (17.35) | 8.08 (5.09) | 3.70 (2.34) | 12.15 (6.62) | 11.44 (5.09) | 22.00 (10.30) | 15.11 (5.95) |
| 3 grams | 64 | 38.70 (17.18) | 8.68 (4.53) | 4.33 (2.16) | 13.33 (6.65) | 12.36 (5.61) | 23.59 (10.02) | 15.81 (6.42) |
| >3 grams | 109 | 38.52 (15.76) | 9.03 (4.30) | 4.29 (2.07) | 13.48 (6.30) | 11.72 (5.14) | 22.55 (10.05) | 15.42 (5.83) |
[i] Notes. Numerical data format: Mean (Standard Deviation). Those who did not disclose their sex (n = 3) were excluded from analysis due to this group’s inclusion violating assumptions of the MATS (i.e., their inclusion created factor level combinations with less than two observations).
Research Questions One and Two
Two MANOVAs were conducted to see if PTSD-related outcomes differed based on cannabis use (see Table 6). Cannabis use and sex were not significantly associated with PCL-5, PHQ-ADS, or ISI scores (p > .05). Conversely, talk therapy attendance was significantly associated with these outcomes (p < .001). Similarly, cannabis use and sex were not significantly associated with PCL-5 symptom cluster scores (p > .05), while talk therapy attendance was (p < .001). When separate MANOVAs were performed with talk therapy attendance as the only factor, it was found that talk therapy attendance significantly correlated with all outcome measures after controlling for family-wise error rate with the Benjamini-Hochberg method (p < .001). Post hoc Tukey tests showed that those who go to talk therapy tended to have higher PCL-5, PHQ-ADS, ISI, and PCL-5 cluster scores than those who do not go to talk therapy. Table 7 shows the mean difference in outcome variable scores between those who attended talk therapy and those who did not.
Table 6
MANOVAs Investigating How Cannabis Use, Talk Therapy, and Sex Impact PCL-5, PHQ-ADS, ISI, and PCL-5 Symptom Cluster Scores in Canadian Veterans Who Have Chronic Pain and a History of Trauma.
| PREDICTOR | MANOVA | |
|---|---|---|
| MATSa | p | |
| Cannabis use MANOVA for PCL-5, PHQ-ADS, and ISI | ||
| Cannabis use | 10.29 | .287 |
| Talk therapy | 88.78 | <.001 |
| Sex | 4.00 | .332 |
| Cannabis use * Talk therapy | 15.00 | .138 |
| Cannabis use * Sex | 9.04 | .340 |
| Talk therapy * Sex | 1.28 | .712 |
| Cannabis use * Talk therapy *Sex | 15.41 | .136 |
| Cannabis use MANOVA for PCL-5 symptom cluster scores | ||
| Cannabis use | 21.95 | .094 |
| Talk therapy | 166.71 | <.001 |
| Sex | 4.15 | .415 |
| Cannabis use * Talk therapy | 29.78 | .043 |
| Cannabis use * Sex | 3.78 | .797 |
| Talk therapy * Sex | 4.69 | .370 |
| Cannabis use * Talk therapy *Sex | 34.75 | .029 |
[i] Note. Those who did not disclose their sex (n = 4) were excluded from analysis due to this group’s inclusion violating assumptions of the MATS (i.e., their inclusion created factor level combinations with less than two observations). Statistically significant p values are bolded (i.e., p < .05).
aA MATS score quantifies the relative difference between two means and performs well with data that has homogeneous variances/covariances and non-normal distributions.
Table 7
Post hoc Analysis on Main Effect of Talk Therapy Attendance on Participants’ PCL-5, PHQ-ADS, ISI, and PCL-5 Symptom ClusterScores in Canadian Veterans Who Have Chronic Pain and a History of Trauma.
| MEASURE | MEAN DIFFERENCE | 95% CI | p |
|---|---|---|---|
| Talk therapy Yes – No | |||
| PCL-5 Total | 13.66 | [10.93, 16.40] | <.001 |
| PCL B | 2.95 | [2.15, 3.75] | <.001 |
| PCL C | 1.47 | [1.09, 1.85] | <.001 |
| PCL D | 5.21 | [4.14, 6.27] | <.001 |
| PCL E | 4.04 | [3.20, 4.88] | <.001 |
| PHQ-ADS | 7.10 | [5.42, 8.79] | <.001 |
| ISI | 2.01 | [1.01, 3.00] | <.001 |
[i] Note. Shown p values were adjusted by the Benjamini-Hochberg adjustment to control for family-wise error rate.
The interaction between cannabis use and talk therapy was significantly associated with PCL-5 symptom cluster scores (p = .043); however, this interaction did not remain significant after performing the Benjamini-Hochberg adjustment and thus no post hoc tests were performed.
Similarly, a significant three-way interaction between cannabis use, talk therapy attendance, and sex was also found to be associated with PCL-5 symptom cluster scores (p = .029). After performing the Benjamini-Hochberg adjustment it was found that this three-way interaction only significantly impacted PCL-D scores (p = .022). However, a Tukey’s post hoc on the PCL-D cluster showed that no cells significantly differed from one another (p > .05).
Research Questions Three and Four
MANOVAs were conducted on the subset of participants who endorsed cannabis use to investigate how factors related to cannabis use such as having a cannabis prescription, preferred route of administration, preferred THC:CBD ratio, and grams/day impact PTSD-related outcomes among participants who are current cannabis users (Table 8).
Table 8
Multiple MANOVAs Investigating the Impact of Cannabis Prescription, Route of Administration, THC:CBD Ratio, and Grams/day on PCL-5, PHQ-ADS, ISI, and PCL-5 Symptom Cluster Scores in Canadian Veterans Who Use Cannabis, Have Chronic Pain, and Have a History of Trauma.
| PREDICTOR | MANOVA | |
|---|---|---|
| MATSa | P | |
| 2 × 2 MANOVAs for cannabis prescription and sex | ||
| Outcomes: PCL-5, PHQ-ADS, and ISI | ||
| Cannabis prescription | 15.66 | .017 |
| Sex | 0.91 | .776 |
| Prescription * Sex | 3.71 | .317 |
| Outcomes: PCL-5 symptom clusters | ||
| Cannabis prescription | 23.29 | .017 |
| Sex | 1.63 | .707 |
| Prescription * Sex | 2.59 | .547 |
| 2 × 4 MANOVAs for route of administration and sex | ||
| Outcomes: PCL-5, PHQ-ADS, and ISI | ||
| Cannabis route | 12.07 | .292 |
| Sex | 1.43 | .639 |
| Route * Sex | 6.02 | .697 |
| Outcomes: PCL-5 symptom clusters | ||
| Cannabis route | 27.51 | .099 |
| Sex | 1.34 | .680 |
| Route * Sex | 9.87 | .571 |
| 2 × 4 MANOVAs for THC:CBD ratio and sex | ||
| Outcomes: PCL-5, PHQ-ADS, and ISI | ||
| THC:CBD | 19.18 | .027 |
| Sex | 5.20 | .182 |
| THC:CBD * Sex | 17.65 | .036 |
| Outcomes: PCL-5 symptom clusters | ||
| THC:CBD | 31.96 | .013 |
| Sex | 3.68 | .377 |
| THC:CBD * Sex | 34.15 | .010 |
| 2 × 4 MANOVAs for grams/day and sex | ||
| Outcomes: PCL-5, PHQ-ADS, and ISI | ||
| Grams/day | 1.08 | .973 |
| Sex | 1.32 | .688 |
| Grams/day * Sex | 3.59 | .723 |
| Outcomes: PCL-5 symptom clusters | ||
| Grams/day | 6.07 | .575 |
| Sex | 2.66 | .511 |
| Grams/day * Sex | 7.50 | .467 |
[i] Note. Those who did not disclose their sex (n = 3) were excluded from analysis due to this group’s inclusion violating assumptions of the MATS (i.e., their inclusion created factor level combinations with less than two observations). Statistically significant p values are bolded (i.e., p < .05).
aA MATS score quantifies the relative difference between two means and performs well with data that has homogeneous variances/covariances and non-normal distributions.
Whether or not active cannabis users had a cannabis prescription was significantly associated with all outcome variables (p = .017). After performing several MANOVAs with cannabis prescription as the sole factor and performing the Benjamini-Hochberg adjustment, it was found that cannabis prescription was significantly associated with PCL-5 total scores, as well as scores on the PCL-B, PCL-D, and PCL-E clusters (p < .05). However, whether participants had a cannabis prescription did not significantly correlate with PHQ-ADS, ISI, or PCL-C subscale scores (p > .05). Post hoc Tukey tests showed that participants with a cannabis prescription tend to have higher PCL-5 total scores, as well as higher PCL-B, PCL-D, and PCL-E cluster scores. Table 9 shows the mean difference in significant outcome variables between participants with, and without, a cannabis prescription.
Table 9
Post hoc Analysis of the Main Effect of Cannabis Prescription on PCL-5 and Relevant PCL-5 Symptom Clustersin Canadian Veterans Who Use Cannabis, Have Chronic Pain, and Have a History of Trauma.
| MEASURE | MEAN DIFFERENCE | 95% CI | p |
|---|---|---|---|
| Cannabis prescription Yes – No | |||
| PCL-5 total | 5.83 | [1.28, 10.38] | .042 |
| PCL B | 1.44 | [0.16, 2.72] | .041 |
| PCL D | 2.14 | [0.42, 3.86] | .034 |
| PCL E | 1.81 | [0.46, 3.16] | .034 |
[i] Note. Shown p values were adjusted by the Benjamini-Hochberg adjustment to control for family-wise error rate.
Participants’ preferred THC:CBD ratio and the interaction between THC:CBD ratio and sex were significantly correlated with PCL-5, PHQ-ADS, ISI, and PCL symptom cluster scores (p < .05). However, neither remained significant after correcting for family-wise error rate. Preferred route of cannabis administration, grams/day, and sex were not significantly correlated with any outcome variables (p > .05).
Discussion
We investigated whether differences in cannabis use and cannabis use characteristics were associated with differences in PTSD-related outcomes in a sample of Canadian veterans with chronic pain. We found that cannabis use was not associated with PTSD severity, psychological distress, or insomnia. This finding is consistent with past research on Canadian veterans that did not find a relationship between cannabis use and PTSD severity (e.g., Sterniczuk & Whelan, 2016). However, this finding is inconsistent with research that has found cannabis use to be associated with both worsened (e.g., Manhapra et al., 2015) and improved PTSD-related outcomes (e.g., Smith et al., 2017). Overall, our results add to the inconclusive literature surrounding the relationship between PTSD and cannabis use.
Cannabis use not being associated with PTSD-related outcomes is inconsistent with the observation that most participants who reported using cannabis for PTSD-related conditions found it to be helpful (Table 3). Past research has also observed that Canadian veterans perceived medicinal cannabis as helpful for mental health conditions such as PTSD (Valikhanova et al., 2023). The inconsistency between veterans’ perceived effect of cannabis on mental health and the measured effect could be interpreted in several ways. First is that cannabis use has no impact on PTSD severity but those who use cannabis perceive that it does. Second, cannabis use may be effective at treating PTSD, but those who use cannabis have more severe PTSD that becomes comparable to nonusers as a result of cannabis use. However, this explanation may be unlikely as cannabis’s efficacy in treating PTSD has not been found to differ from placebo (Bonn-Miller et al., 2021).
Finally, cannabis use may provide temporary symptom relief but not affect long-term PTSD severity. This would explain why those who use cannabis report that it helps treat PTSD symptoms but why validated measures such as the PCL-5 often fail to find a relationship between cannabis use and PTSD. This explanation is consistent with research that has reported that acute cannabis use lowers PTSD-related symptoms in the short-term but has no observable effect on PTSD severity in the long-term (LaFrance et al., 2020). Because of this, future studies may benefit from the inclusion of measures that specifically focus on the short-term effects of cannabis use, or by using techniques such as momentary ecological assessment to investigate the immediate effects of cannabis use in natural environments.
Talk Therapy
Participants who attended talk therapy had more severe symptoms in all outcomes measured. Previous studies have reported that trauma-focused therapy attendance is associated with exacerbation of PTSD symptoms in a minority of individuals (Larsen et al., 2016; Walker et al., 2020). However, this exacerbation is temporary and those who experience it often show a reduction in PTSD severity over the course of treatment. So, although the higher symptoms found in the talk therapy attendance group may be partially explained by trauma-focused therapy attendance, it is likely that this trend is also due to those with higher-severity PTSD symptoms being more likely to seek treatment.
It was also found that the interaction between cannabis use and talk therapy attendance, and the three-way interaction between cannabis use, talk therapy attendance, and sex, were significantly associated with PCL-5 symptom clusters. However, these interactions did not remain significant during follow-up analyses. One possible explanation behind these interactions is that cannabis use may modulate the effect of talk therapy. For example, previous studies have found that acute and chronic cannabis use may facilitate and suppress fear extinction respectively (Papini et al., 2017; Raymundi et al., 2020). However, other studies have found that cannabis use has no impact on the efficacy of trauma-focused psychotherapies (Ruglass et al., 2017). Given veterans’ interest in cannabis as a form of PTSD treatment, more research is needed on how cannabis use impacts different forms of psychotherapy.
Cannabis Use Characteristics
Participants with cannabis prescriptions tended to have more severe PTSD symptoms than those who use cannabis without a prescription. Specifically, those with cannabis prescriptions had more severe intrusive symptoms (PCL-B), negative alterations to cognition and mood (PCL-D), and arousal symptoms (PCL-E). This could suggest that medicinal cannabis use (i.e., prescribed cannabis) worsens PTSD outcomes, however, it seems more likely that those with severe PTSD are more prone to seek out and/or be approved for prescribed cannabis. However, Additional longitudinal research would be needed to confirm this explanation. Interestingly, prescription cannabis use was not associated with the avoidance (PCL-C) symptom cluster. This may indicate that prescription cannabis users are trying to treat PTSD symptoms rather than using cannabis as a means of avoidance. However, the restricted range of the PCL-C cluster may not have allowed enough variability between groups to detect meaningful differences in avoidance symptoms.
Cannabis prescription was used instead of self-reported medicinal cannabis use to avoid response bias due to potential cannabis stigma. Indeed, more respondents claimed to be using cannabis medicinally, or medicinally and recreationally (n = 299) than had a cannabis prescription (n = 257). However, it is also possible that some cannabis users may be trying to self-medicate but are unable to obtain a cannabis prescription. Thus, the higher number of participants claiming medicinal cannabis use compared to those with a cannabis prescription may be a result of both self-medication and stigma avoidance. Conversely, it is also possible that prescribed cannabis could be used recreationally instead of medicinally. When considering all these factors, our ability to accurately distinguish recreational use from medicinal use may be limited. Thus, differences observed between these groups should be interpreted with caution.
Participants’ preferred route of administration and grams/day of cannabis consumption were not significantly associated with any PTSD-related outcomes. However, preferred route of administration did approach significance for the PCL-5 symptom clusters. As per Table 5, a numerical trend can be seen where a preference for absorbed cannabis is associated with a lower level of global psychopathology. This is likely because those who are using cannabis via absorption are doing so for physical rather than psychological effects. Although veterans’ average cannabis consumption was not associated with PTSD severity, dosage may still influence the short-term effects of cannabis on PTSD. For example, higher doses of cannabis have been found to produce larger short-term reductions in anxiety and intrusive symptoms (LaFrance et al., 2020).
Demographics
There were no significant sex differences or interactions observed. However, genuine differences between male and female participants may not have been detectable due to the low number of women in our sample. Additionally, men and women may differ in their response to cannabis in other ways. For example, the acute effects of cannabis on PTSD symptoms differ by gender (LaFrance et al., 2020). Furthermore, although not investigated here due to cell size restraints, the relationship between cannabis use and PTSD may differ in gender minorities. Because of these reasons, future studies should continue to investigate the role sex and/or gender may play in the relationship between cannabis use and PTSD.
Interestingly, current cannabis users showed a higher rate of unemployment (33.54%) than past (24.29%) and never users (15.65%; Table 1). However, the association between these two variables may be explained by the relationship between cannabis use and disability (Table 1), as current cannabis users were more likely to have a disability that precluded their ability to work (46.95%) than past (37.14%) and never users (28.70%).
Limitations
This study had several limitations. The largest limitation was our cross-sectional design, which precludes our ability to determine the presence or direction of causality between cannabis use and PTSD. Future studies may benefit from longitudinal designs which would provide stronger evidence of causality and show how PTSD and cannabis use can impact each other’s development over time. The generalizability of the current study’s findings may be limited to Canadian veterans with chronic pain and PTSD and may not be applicable to Canadian veterans in general or other populations suffering from PTSD. Generalizability could also be limited due to participants self-selecting, and thus our results may not be representative of Canadian veterans with chronic pain at large. Our use of self-report measures may have obfuscated differences between PTSD symptom clusters as self-report measures of PTSD are not as effective at differentiating between similar symptoms as diagnostic interviews are (Rosencrans et al., 2024). Our current results are also limited by our binary operationalization of talk therapy attendance (i.e., yes or no). Thus, we were not able to investigate how different forms of talk therapy (e.g., group vs. individual, manualized vs. non-manualized, etc.) and one’s adherence to talk therapy impacts the relationship between talk therapy, PTSD, and cannabis use. Finally, we did not test to see if participants’ responses were impacted by socially desirable responding, which may have affected responses to questions concerned with stigmatized topics such as mental health and substance use.
Conclusion
Several promising avenues for future research were identified in the current study. For example, both talk therapy attendance and prescription cannabis use were associated with more severe PTSD-related outcomes. Although it is likely that veterans with more severe symptoms are more prone to seek out mental health support through talk therapy and/or medicinal cannabis, more longitudinal research is needed to better understand these associations. Additionally, longitudinal designs may help clarify the efficacy and suitability of cannabis in treating PTSD, as well as clarify which cannabis use characteristics are associated with better or worse outcomes. Investigating how cannabis use mediates the effectiveness of talk therapy may also be promising. Finally, future research may benefit from the inclusion of measures that specifically assess the short-term effects of cannabis use, that measures such as the PCL-5 are not designed to detect.
Additional Files
The additional files for this article can be found as follows:
Ethics and Consent
The current study, and the study on which this secondary analysis was based, were both approved by the Hamilton Integrated Research Ethics Board (HiREB; project # 15279). All participant data was stored on an encrypted Qualtrics file format on a secure server at Hamilton Health Sciences (HHS) before being anonymized after data collection had been completed.
Acknowledgements
David Storey is supported in part by funding from the Social Sciences and Humanities Research Council.
Funding Information
The present study was conducted as part of the research project: “Veterans Cannabis Therapy Outcome Registry” (VECTOR), a multi-phase project designed to study the efficacy of cannabis as a treatment for chronic pain in Canadian veterans and was conducted in conjunction with the Hamilton Health Sciences Pain Clinic and funded by the Chronic Pain Center of Excellence for Canadian Veterans.
Competing Interests
Mitchell Sheehy and David Storey, as veterans of the CAF, both receive financial benefits through VAC. However, neither have previously nor are currently receiving reimbursement for medicinal cannabis through VAC. The authors have no other competing interests to declare.