Does Cannabinoid Administration Affect High Levels of COVID-19–Related Daily Perceived Stress? A Multiple Baseline Study

Introduction

The global COVID-19 pandemic has negatively impacted the psychological well-being of people worldwide; empirical research documents increased stress, anxiety, depression, and distress.^1-3 Development of interventions that can alleviate these problems is a global public health priority. The current study evaluates perceived stress, anxiety, and distress among people experiencing high levels of COVID-19–related distress.

Although these constructs can be difficult to differentiate due to their conceptual overlap, theoretical evidence suggests that perceived stress refers specifically to an individual’s perception that environmental challenges may exceed one’s ability to cope with those challenges successfully,⁴ anxiety is a state of future-oriented worry,^5,6 and distress reflects a general, negatively valenced subjective experience.

Cannabidiol (CBD) is a nonintoxicating, phytochemical cannabinoid that can potentially reduce anxious arousal and related states.^7,8 Although studies have documented that CBD reduces cardiovascular reactions associated with physical stress (eg, blood pressure in response to vigorous exercise),^9,10 we are not aware of studies designed to test the effects of CBD on perceived stress in humans. Evidence suggests that CBD can reduce stress-relevant psychological arousal.⁹ In preclinical studies, CBD has been shown to reduce anxiety and fear.^8,11-14

Similarly, research suggests that CBD reduces acute anxious reactivity among healthy and anxious participants and chronic anxiety symptoms in adolescents with social anxiety disorder.^15-19 Evidence also suggests that CBD reduces activity in arousal-generating structures, including but not limited to the anterior cingulate cortex and amygdala.^20,21 Reducing activity in these structures (eg, perhaps with CBD administration) is likely to effect downstream systems, including the hypothalamic–pituitary–adrenocortical axis, thought to partially mediate stress responses.^22,23

The current study aimed to test the hypothesis that CBD could reduce perceived stress, anxiety, and distress, which we collectively refer to as psychological stress, among participants experiencing elevated COVID-related distress.

Methods

Study Design

Although randomized controlled trials are the gold standard in measuring treatment efficacy, a preliminary small-N design is an optimal approach for assessing an intervention on the individual level.²⁴ Thus, a small-N, multiple baseline design is particularly advantageous in this context as it provides preliminary data for informing large-scale trials. More specifically, a nonconcurrent multiple baseline design allows for continuous enrollment and random assignment to variable baseline links before beginning CBD administration.

This approach provides a methodological control for potential history effects. Moreover, establishing a stable baseline measure of psychological stress for randomly determined periods, followed by CBD administration, and continued assessment of psychological stress, offers insight into the effects of CBD on psychological stress. Implementing a stable baseline prior to intervention increases the likelihood that intervention effects (if any) are related to the intervention, rather than symptom improvement that may naturally follow the baseline.²⁵

To our knowledge, this was the first human study in this area to employ a nonconcurrent, multiple baseline design study. The study’s overarching goal was to gauge the effects of total daily administration of 320 mg of CBD using a small-N, multiple-baseline design (N= 6; Table 1). We hypothesized that perceived stress would decrease upon initiation of CBD use. Similarly, we tested the hypothesis that CBD administration would also reduce participants’ daily distress and anxiety ratings. The effects of CBD administration on sedation (eg, drowsiness) and cognitive impairment were considered safety-related end points.

Finally, we measured pre- and post-intervention levels of anxiety, depression, and insomnia symptoms using a clinically relevant index. These questions were evaluated among a small sample of people experiencing particularly elevated levels of COVID-19–related distress.²⁶ As such, the current study provides preliminary experimental evidence of the potential of CBD for reducing elevated psychological stress related to the COVID pandemic.²⁶

Study Sample

Table 2 lists the eligibility criteria designed for participant safety (eg, absence of relevant allergic conditions), to address potential confounds (eg, current use of anxiety medication, delta-9-tetrahydrocannabinol, CBD, and markedly high or low body mass, which can affect CBD absorption), and pragmatic concerns (eg, ability to pick up the product on campus, access to stable internet connection).

These eligibility criteria are standard in CBD administration studies to protect human subjects and increase study rigor. Table 2 includes detailed demographic information of participants who completed the intervention phase. The sample was comprised primarily of single, heterosexual women in their mid-20s who earned less than $60,000 USD annually. All participants had completed at least some college. Three participants identified as White/Caucasian, 1 as Black/African American, and 2 as Hispanic/Latino.

Procedures

Figure 1 presents information regarding participant flow from screening through the study. Potential participants were recruited from a mid-sized college town in the Southeastern United States via online and paper advertisements during early fall 2020. Study enrollment occurred during fall 2020, when COVID-19 cases in the region were relatively high and stable for the initial wave of the pandemic.²⁷

Participants were directed to a Qualtrics website to learn more about the study and to complete the initial screening form, which included the COVID-19 Stress Scale (CSS), Perceived Stress Scale 4 (PSS-4), and the eligibility criteria listed in Table 1. Of the 115 participants who completed the screener, 97 were ineligible based on the 20 eligibility criteria. An additional 9 participants screened in, but the study was complete (ie, the target of 6 participants was reached), so they were omitted.

As has been done in prior work,²⁸ an N of 6 was selected to ensure that 2 participants were randomized to each of the 3 baseline periods. This process took place over 3 months. Potentially eligible participants scheduled an online baseline interview, which was the preferred format rather than an in-person interview due to COVID-related health and safety concerns.

First, we obtained written informed consent using an electronic consent form. Next, eligibility criteria were reassessed via an interview that covered all 20 criteria and was administered to confirm screener accuracy. Eligible participants were administered a baseline survey including the Depression Anxiety Stress Scale-21 (DASS-21) and the Insomnia Severity Index (ISI) randomly assigned to a 3-, 5-, or 7-day baseline period.²⁹ The rationale for the multiple baseline design (eg, addressing key threats to internal validity) is detailed in the Methods section under Study Design.

Between 5 and 7 pm each evening during the baseline phase, participants completed an electronic survey comprising the PSS-4 and our 4-factor Visual Analog Mood Scale (VAMS). Before beginning the study, we informed participants they might have to restart this baseline period. To ensure participants provided accurate responses, they were not told of the criteria necessary to move from baseline to the intervention phase.

Once participants’ baseline PSS-4 scores demonstrated stability (defined as PSS-4 scores remaining 1 SD above the normed mean throughout the randomly assigned baseline length), research staff explained, via virtual platform, procedures for the intervention phase. If the participants’ PSS-4 scores demonstrated instability, the data were discarded, and the baseline period was restarted (Figure 1).

Participants began the intervention phase the day after respective baseline phases ended. They were instructed to take 160 mg of CBD twice daily with food, between 8 and 10 am and 12 and 2 pm for 7 days. This administration period was selected to extend the findings of prior work indicating an acute (single) dose of CBD is therapeutic.^16,17,19 The product was a softgel capsule containing 20 mg of hemp-derived CBD isolate in medium-chain triglyceride oil, manufactured according to 21CFR Part 111 current Good Manufacturing Practices.

Participants were prompted via email to complete compliance assessment surveys to check whether they ingested the softgels at 8 am and 12 pm daily during the entire intervention phase. Between 5 and 7 pm, participants were prompted via email to complete the PSS-4 and multipart VAMS measures. Upon completing the 7-day CBD administration period, participants completed a post-intervention survey, including the DASS-21 and ISI, and were fully debriefed.

Specifically, participants were apprised of the study’s purpose and hypotheses, informed about how study findings would contribute to understanding of CBD’s effects, and thanked for their participation. Participants were compensated with a bottle of commercially available CBD softgels for completing the baseline phase and a second bottle of commercially available CBD softgels for completing the study’s intervention phase. The softgels given for participant remuneration were identical to those administered in the study.

Measures

Participant Characteristics

Demographics. Participants were administered a brief demographic questionnaire during screening. Once matriculated, participants reported a more thorough set of demographics to fully characterize the sample.

COVID-19–Related Distress. The CSS was administered during the eligibility screening to identify individuals experiencing high levels of COVID-19–related distress.³⁰ The CSS is a 36-item, multifactorial measure with sound psychometric properties^26,30,31 that assesses the degree of distress related to the COVID-19 pandemic. Items on this measure include, “I am worried about catching the virus,” “I am worried about keeping my family safe from the virus” (0 = not at all to 4 = extremely), and “I had bad dreams about the virus” (0 = never to 4 = almost always). The source article provides a complete list of the measure items.³⁰ A different measure, the PSS (described below), was used as the primary index of perceived stress.

Protocol Compliance

Compliance Assessment. We designed a 5-item compliance assessment that was administered during the intervention stage to confirm participants complied with the requirements for CBD self-administration. One yes/no item captured whether participants ingested the correct amount of CBD (160 mg) twice daily. The remaining items were not of relevance to the current study.

Outcome Measures

Perceived Stress. The 4-item version of the PSS-4 was used to measure participants’ perceived stress in a 24-hour period during the screening, baseline, and intervention phases.⁴ Items were rated on a scale from 0 (never) to 4 (very often). Items from this measure included the question, “In the past 24 hours, how often have you felt that you were unable to control the important things in your life?”

Mood States. The VAMSis a psychometrically reliable scale used to measure participants’ current experience of 16 mood states.^32,33 An additional VAMS item was also included to measure state distress (ie, at ease/distressed; VAMS-D). Participants rated their mood states during the baseline and intervention phases by dragging a slider along a line anchored by 2 contrasting moods; 0 refers to the positively valenced mood, and 100 refers to the negatively valenced mood (eg, 0 = relaxed, 100 = tense). The anxiety factor (eg, relaxed/tense; VAMS-A) was used to assess daily anxiety. The sedation factor (eg, quick-witted/mentally slow;VAMS-S) and the cognitive impairment factor (eg, alert/drowsy; VAMS-C) were employed as safety-related end points.

Anxiety and Depression. The DASS-21 was administered pre- and post-intervention to measure changes in anxiety and depression levels.³⁴ The DASS-21 is a 21-item self-report measure with 3 factors that assess past-week symptoms of stress (DASS-S), anxiety (DASS-A), and depression (DASS-D).

Sleep Problems. The ISI, a psychometrically strong 7-item questionnaire, was administered pre- and post-intervention to measure sleep problems.^35-37 An example question included, “How satisfied are you with your current sleep pattern? 0 = very satisfied to 4 = very dissatisfied.” Participants were instructed to rate each item based on the past week.

Statistical Analysis

After descriptive analyses and graphical depiction of study variables across the course of the study (Figures 2-7), Tau-U analyses were conducted to examine change from baseline to the intervention phase.³⁸ We chose the Tau-U approach because of its sensitivity to baseline length and stability, ability to index degree of overlap across phases, and allowance for examining the impact of trends across phases. For each outcome variable, we evaluated the following starting with the least conservative analysis and ending with the most conservative analysis:

• A test of nonoverlap across phases, reflecting overall change by comparing scores in baseline (BL) with those in the intervention (INT; [BL vs INT]);
• Nonoverlap across phases while accounting for variability within the intervention phase (BL vs INT + trend I); and
• Nonoverlap across phases while accounting for variability within both the intervention and baseline phases (ie, BL vs INT + trend I – trend B).

Results

The remaining 6 participants reported taking 8 softgels (ie, 160 mg) twice daily, suggesting compliance with the intervention. As detailed in Table 2, CSS scores ranged from 76 to 117 (M = 89, SD = 17.2).

Using the most conservative analysis, PSS-4 score trends within the intervention phase were significantly lower than score trends at baseline when controlling for trends in both phases for participants C (S = 34, Tau = 0.76, SD = 10.92, VAR = 119.33, Z = 3.11, P < 0.01) and E (S = 41, Tau = 0.45, SD = 17.73, VAR = 314.33, Z = 3.33, P < 0.01). A more liberal set of analyses revealed participants B, D, and F evidenced significantly lower PSS-4 score trends within the intervention phase than at baseline.

Figure 2 shows the PSS-4 scores for all assessment points for each participant. Mean PSS-4 scores during the baseline and intervention phases were 10.43 (range: 8-13) and 7.38 (range: 4-11), respectively. Table 3 includes the results of Tau-U analyses of each participant’s PSS-4 scores. These data suggest that when using the most conservative statistical analysis, 2 participants showed improvement in PSS-4 scores, even after accounting for instability in scores during the 2 phases.

Using the most conservative analysis, VAMS-A score trends within the intervention phase were significantly lower than in the baseline phase when controlling for trends in both phases for participant B (S = 22, Tau = 0.33, SD = 14.58, VAR = 212.67, Z = 1.51, P < 0.05). A more liberal set of analyses revealed that participants D and F showed significantly lower VAMS-A score trends within the intervention phase than in the baseline phase.

Figure 3 displays each participant’s VAMS-A scores at all assessment points. Mean scores for the VAMS-A during baseline and intervention phases were 67.07 (range: 17.67-93.67) and 50.56 (range: 19.67-82), respectively. Table 4 presents the results of the Tau-U analyses for VAMS-A scores. These data suggest that when using the most conservative statistical analysis, 1 participant showed improvement in VAMS-A scores, even after accounting for instability in scores during the baseline and intervention phases.

Using the most conservative analysis, VAMS-D score trends within the intervention phase were significantly lower than at baseline when controlling for trends in both phases for participants B (S = 30, Tau = 0.45, SD = 14.51, VAR = 210.67, Z = 2.07, P < 0.01) and F (S = 16, Tau = 0.21 SD = 16.39, VAR = 268.67, Z = 0.98, P < 0.05). A more liberal set of analyses revealed that participant E showed significantly lower VAMS-D score trends within the intervention phase than in the baseline phase.

Figure 4 shows the distress ratings. Mean scores for the VAMS-D during the baseline and intervention phases were 77.53 (range: 45-100) and 52.05 (range: 4-91), respectively. Table 5 presents the results of the Tau-U analyses for distress ratings. These data suggest that when using the most conservative statistical analysis, 2 participants showed improvement in VAMS-D scores, even after accounting for instability in scores during the 2 phases.

VAMS-S score trends within the intervention phase were not significantly higher than score trends in the baseline phase when controlling for trends for any of the 6 participants, indicating that CBD did not increase participants’ sedation. Using the most conservative analysis, participant A showed improvements in VAMS-S score trends during the intervention phase but not at baseline when controlling for trends in both phases (S = 19, Tau = 0.42, SD = 11.18, VAR = 124.99, Z = 1.70, P < 0.05).

No other participants showed changes in VAMS-S. Figure 5 displays the VAMS-S ratings for all assessment points. Mean scores for the VAMS-S during the baseline and intervention phases were 59.83 (range: 32.14-96.57) and 49.89 (range: 21-82.86), respectively. Table 6 presents the results of Tau-U analyses of VAMS-S ratings.

VAMS-C score trends within the intervention phase were not significantly higher than those in the baseline phase when controlling for trends in both phases for any of the 6 participants, indicating that CBD did not increase participants’ cognitive impairment. Figure 6 shows VAMS-C scores for all assessment points. Mean scores for the VAMS-C during the baseline and intervention phases were 55.92 (range: 29.5-100) and 51.2 (range: 0-84), respectively. Table 7 presents results from the Tau-U analyses for VAMS-C scores.

Finally, regarding pre- to post-intervention changes in insomnia and symptoms of anxiety, stress, and depression, Figure 7 displays pre- and post-intervention scores for DASS subscales and the ISI. Descriptive statistics revealed that most participants demonstrated reduced anxiety, stress, depression, and insomnia severity.

Discussion

To our knowledge, this was the first multiple baseline design study designed to test CBD for reducing perceived stress in humans. Our results suggest that 160 mg of CBD taken twice daily with food significantly reduced perceived stress within individuals experiencing elevated stress related to the COVID-19 pandemic, with no negative impact on sedation or cognitive impairment.

Trend comparisons from inferential tests revealed that significantly lower levels of perceived stress were reported in the intervention phase compared with baseline in 5 of 6 participants. Most notably, 2 participants experienced perceived stress reduction, even after statistically accounting for variability in the intervention phase and trends observed during the baseline phase (ie, BL vs INT + trend I – trend B).

Moreover, participants who completed baseline phases that varied in length (ie, 3, 5, or 7 days) reported reduced perceived stress during the intervention phase. As such, it is not likely that reduced stress resulted from repeated assessment or time. Despite methodological controls (ie, having participants restart their baseline if PSS-4 scores were not >1 SD than normative means) to establish a stable baseline in terms of perceived stress, there remained variability in PSS-4 scores during the baseline phase for participants D and F.

That is, although PSS-4 scores were above normative means during the baseline phase, Figure 2 illustrates what appears to be a modest global decline in scores for these participants at baseline. A more stable baseline might have allowed for the detection of reduced stress after accounting for baseline trends. These results suggest daily use of 320 mg of CBD can reduce perceived stress due to the COVID-19 pandemic; however, larger-scale clinical studies will ultimately be needed to suggest CBD does reduce perceived stress in a broader population.³⁹

The present results also suggest that CBD may reduce anxiety and distress during a substantial, prolonged environmental stressor such as the COVID-19 pandemic. Participants B, D, and F reported lower anxiety during the intervention phase. This effect was observed after accounting for variability in the intervention phase for all 3 individuals and after accounting for a trend in the baseline for participant B (see “Statistical Analysis” section for additional details).

Three participants reported reduced distress during the intervention phase, 2 of whom reported reductions after accounting for variability in the intervention phase and trend during the baseline. Notably, only stability in perceived stress was targeted during the baseline phase. As a result, there was more significant variability in baseline levels of anxiety and distress (Figures 3 and 4, respectively), which makes it more challenging to identify an effect of an intervention in a multiple baseline design.²⁹

Figure 7 also displays pre- to post-intervention reductions in anxiety, depression, stress, and insomnia symptoms. The past-week assessment timeframe of the DASS-21 and ISI did not allow enough repeated measurements to apply the Tau-U analyses. Still, the data pattern of Tau-U results suggest reductions in anxiety and stress, and depression and insomnia symptoms. Studies that prioritize anxiety and distress outcomes (eg, establish stable baselines, more frequent assessment) are needed. Indeed, this study sets the stage for addressing several important empirical questions, such as varying CBD doses and intervention timelines.

The only significant change in sedation and cognitive impairment was an improvement in sedation for participant A. Discussions during the debriefing phase of the study indicated that 4 participants identified no drawbacks to using CBD, whereas participants C and E reported drowsiness. These data are broadly consistent with research on the adverse effects of CBD, which suggest somnolence is one of the more common side effects.^40-42

Future studies should incorporate neurobiological work with self-report indicators, as existing neurobiological work points toward the effects of CBD on amygdala activity and downstream activity of the hypothalamic–pituitary–adrenocortical axis.^7,20,21,23 It is possible that CBD directly affects mechanisms underlying perceived stress, anxiety, and/or distress (eg, amygdala activity). It is also possible that reductions in perceived stress led to correlated decreases in anxiety and distress or vice versa. Research that measures the activity of the biological systems underlying each of these subjective experiences to understand how CBD affects these distinct, yet related outcomes is needed.

The present study had additional limitations that warrant consideration. First, limitations inherent to the small-N methodology constrain external validity. The current sample was comprised of a very small population of 6 relatively highly educated, single, young adults (£40 years of age) from South-Central United States. Furthermore, these participants were required to meet a rigorous list of eligibility criteria.

Studies are needed to test the generalizability of these results to other demographics. The present study also only allowed for conclusions to be drawn regarding 320 mg of CBD administered daily for 7 days. The dosage choice was informed by evidence of the anxiolytic effects of daily use of 300 mg of CBD and practical constraints of using 20-mg softgels, which prevented 2 daily doses of 150 mg each.¹⁸

The present study suggests that daily use of CBD for 7 days may reduce perceived stress experienced during a substantial, prolonged environmental stressor such as the COVID-19 pandemic. Randomized placebo-controlled clinical trials are needed to provide additional support for the stress management effects of CBD. The safety profile of CBD⁴¹ and its potential to help individuals manage the psychological effects of environmental stressors could assure its place in the repertoire of interventions for mitigating the psychological and biological impact of chronic stress.