Chronic pain is a significant public health concern. It is one of the most common reasons for seeking medical care [1], with considerable societal and financial burden [2,3] in addition to human suffering. Chronic pain treatment is challenging due to limitations of existing pharmacological approaches, particularly intolerability of many medicines [4], health risks [5-9], and lack of sustainable availability. In contrast, behavioral medicine approaches are low risk and have been shown to have low-to-moderate efficacy for a range of chronic pain conditions [5,7,10-12], with cognitive behavioral therapy (CBT)–pain as the established evidence-based treatment for chronic pain [8,9]. However, several barriers prevent broad access to CBT-pain, including out-of-pocket costs, burdensome travel, lack of availability in remote or rural areas, and a lack of adequately trained clinicians [6,13]. The length of treatment (ie, eight sessions or more for CBT-pain) is associated with increased dropout [14], which may perpetuate the overuse of medical services and health disparities. The COVID-19 pandemic has further challenged the feasibility of multisession in-person treatment options [15], and this creates the need for accessible solutions that are low-cost, low-burden, and remove the need for face-to-face meetings [16].

Brief web-based and remotely delivered psychological interventions have been demonstrated to accomplish this goal and have been found to be as effective as face-to-face therapy, while decreasing health care use and barriers to treatment [17]. Effective web-based treatments now exist for various psychological disorders, including depression, panic disorder, and social phobia [18]. Web-based CBT has been effective for a number of chronic health conditions, including irritable bowel syndrome, tinnitus, and headache [19,20]. In a systematic review of web-based interventions for chronic back pain, Garg et al [21] reported that web-based interventions, including CBT, effectively reduced pain catastrophizing scale (PCS) scores. Effect sizes for web-based interventions for chronic pain have been estimated to range from small to moderate (Cohen d range 0.04-1.23) [18]. In addition to CBT, existing web-based interventions for patients with pain include compassionate mind training [22], social media–based web-based community intervention [23], pain self-management [24], and hypnosis [25]. Despite the promise of web-based interventions for pain, multisession treatments typically range from 20 days to 12 weeks in duration, which can be costly, burdensome, and limit access to adequate care [26].

Single-session interventions (SSIs), which have been defined as the intentional delivery of a specific, structured program involving a single visit with a clinic, provider, or program [27], have the potential to reduce the burden of traditional multisession treatment approaches and expand access to behavioral medicine. SSIs have been found to be effective for a variety of psychological and health conditions, including severe mental illness [28], anxiety and youth conduct disorder [27], acute insomnia [29], heavy alcohol consumption in college students [30], trauma and adversity [31,32], postsurgical pain [33], and chronic pain [11,12,34-36]. SSIs that have been further optimized via web-based delivery have demonstrated feasibility and efficacy similar to face-to-face interventions [11,12,37] and include interventions for multiple sclerosis pain [38], disordered gambling [39], and adolescent mental health [40-42]. In addition to reducing treatment burdens, web-based SSIs have the vital advantage of ease of scalability, as they can be completed by patients in any location [37]. Effective web and digital-based SSIs would eliminate or reduce many of the existing barriers to treatment, such as cost, lack of trained therapists, and insurance limits on the length of treatment. However, despite the promise of SSIs, rigorously designed and randomized controlled studies are needed to establish their efficacy [43].

Building on the limited literature on web-based SSIs for chronic pain, we aim to evaluate a Zoom-delivered version of a previously developed and efficacious, skills-based, behavioral medicine SSI empowered relief (ER) [11,36]. Previously, our group conducted a three-arm randomized controlled trial conducted in 263 adults with chronic low back pain comparing in-person ER to a health education class or 8-week CBT [12]. However, web-based delivery of ER has not been tested, particularly in mixed-etiology chronic pain, and the COVID-19 context underscores the importance of providing effective home-based chronic pain care. Extending this work to the digital platform, in this study, we conducted a parallel-group, randomized (1:1) comparative efficacy trial to assess the impact of a single-session videoconference-delivered ER group pain relief skills class versus waitlist control (WLC). We hypothesize that at 3 months posttreatment, (1) ER would be superior to WLC for reductions in pain catastrophizing, an index of pain coping; (2) ER would be superior to WLC for reductions in pain intensity, bothersomeness, and sleep disturbance; and (3) ER would be superior to WLC in reducing anxiety, depression, and physical function.

This study was a parallel-group, randomized, clinical trial comparing ER, a 2-hour single-session videoconference-delivered class, to a WLC. Enrolled participants (N=104) aged 18-80 years with mixed chronic pain etiology were randomized (1:1) to one of the two study arms. Participants completed outcome assessments at pretreatment, week 2, and 1, 2, and 3 months posttreatment. The primary outcome was pain catastrophizing levels at the 1-month follow-up [35], with pain intensity, pain, bothersomeness, and sleep disturbance as secondary outcomes. The tertiary outcomes included PROMIS (Patient-Reported Outcomes Measurement Information System) measures of anxiety, depression, and physical function at the 1-month follow-up time-point. We also report a more rigorous test of the durability of treatment effects at 3 months posttreatment as highlighted in the Results section. A crossover intervention class was offered to participants who were initially randomized to the WLC and was conducted in February 2021. Of 48 participants who were offered the treatment, 20 (42%) attended the crossover class at no cost, and no data were collected. This study was conducted between August 2020 and February 2021.

Participants were recruited remotely through targeted emails to lists of patients who agreed to be contacted for research purposes through two databases: (1) the Stanford’s Collaborative Health Outcomes Information Registry comprising patients who have received care at the Stanford Pain Management Center (a tertiary referral pain clinic in Redwood City, California) and (2) the Stanford Systems Neuroscience and Pain Lab database of individuals with chronic pain. Interested individuals were directed to a web-based screening form, and those who met the initial eligibility criteria were contacted by research staff by phone to confirm their eligibility. Approximately 18.3% (19/104) of the study sample were clinic patients, and 81.7% (85/104) were from the larger community. Eligible and interested participants were invited to participate in a pain-skills treatment class compared with a WLC and completed an electronic informed consent form (see Multimedia Appendix 1 for the study consent form). After informed consent was obtained, study participants were randomized one-to-one using a REDCap (Research Electronic Data Capture; Vanderbilt University) [44] cloud random number generator and allocated to the study group. All study procedures were completed remotely, and no in-person visits were required. Figure 1 displays the participants’ study activities.

All the study procedures were performed remotely and electronically. Enrolled participants completed a baseline survey, which was conducted 4 days before the scheduled class, followed by posttreatment surveys at 2 weeks, and 1, 2, and 3 months. The intervention was conducted on Zoom [45], and details of the Zoom session were emailed to participants before the scheduled class time. Overall, in the ER arm, two cohorts were recruited, each consisting of 24-26 participants per class. The class sizes were consistent with studies on web-based interventions [46-51]. Participants received US $10 in the form of an Amazon electronic gift code for each survey they completed and could receive up to US $50 following completion of all study procedures. The research procedures were approved by the Stanford institutional review board.

All study recruitment efforts directed interested individuals with chronic pain to the web-based screening form to complete an automated eligibility form that screened for initial inclusion and exclusion criteria. Participants who met initial eligibility were then contacted by the research staff to confirm study eligibility according to the inclusion and exclusion criteria provided in Textbox 1. Internet and computer literacy were the implicit de facto eligibility criteria. If interested participants indicated that they had never used the Zoom platform, the research coordinator contacted them by phone to schedule an individual meeting; their skill to participate in the web-based class would increase because they were more familiar with the Zoom platform and navigating the tool in a supervised setting. Of the 9 participants who indicated that they had never used Zoom, none requested to schedule an individual meeting.

Participants were randomized 1:1 (no blocking applied) and allocated to one of two study arms: (1) ER or (2) WLC. REDCap was used to apply an automatic and blinded randomization program and ensure equal allocation to both groups. Participants were not blinded to the study arm, given the nature of WLC. Statisticians performed blinded analysis of data sets that were randomly labeled as group 1 and group 2, with statistician unblinding occurring only after posttreatment month 3 data were analyzed. The study coordinator (LGC) was unblinded to the individual study arm assignments and was not involved in any data analyses. The study protocol was reviewed and approved [52].

All surveys were collected through REDCap [44], a web-based electronic data capture platform, which is a secure (password protected), HIPAA (Health Insurance Portability and Accountability Act)–compliant platform, and hosted by the Stanford University School of Medicine.

Data collection consisted of electronically collected participant-reported measures across the four phases of the study: screening, pretreatment, treatment, and posttreatment. For both groups, posttreatment survey collection was conducted at 2 weeks and 1-3 months posttreatment. The ER group completed an additional brief survey 1-week posttreatment to assess participant satisfaction with the intervention and Zoom platform. Table 1 shows the data collection time-points for both the groups. Baseline surveys included demographic information, pain intensity, pain bothersomeness, pain catastrophizing, and PROMIS measures. Posttreatment surveys included the same measures, excluding the demographics.

The planned enrollment accounted for 4% attrition in each study arm. The enrolled sample size ensured significant statistical power (80%) to detect medium-to-large treatment effects between the two treatment groups. Of the 104 enrolled participants (ER: 50/104, 48.1%; WLC: 54/104, 51.9%), intent-to-treat analyses were used to analyze the data of 101 (97.1%) participants (ER: 50/101, 49.5%; WLC: 51/101, 50.5%) who met the study criteria and completed at least one survey.

Baseline scores on continuous variables were summarized using mean (SD), and categorical variable scores were summarized by count and proportions. The differences between conditions at baseline for continuous and categorical variables for the treatment groups were assessed using one-way analysis of variance and Fisher exact test, respectively. For primary analyses, we used a between-within mixed design to assess the effect of treatment, time, and treatment × time, where time represented the within-subjects factor and treatment was specified as a between-subjects factor. Within-subjects’ dependency was modeled by specifying participants as a random effect with a heterogeneous compound symmetry covariance structure. Similar analyses were used to examine the treatment effects for all secondary and tertiary outcomes at each posttreatment month. Intention-to-treat analysis was used to investigate the causal effects of the treatment. All statistical significances were applied at a two-sided level of 0.05, with Benjamini-Hochberg adjustment for multiple comparisons. Cohen d effect sizes for the between-subjects’ factor were calculated from the between-subjects’ F statistic adjusting for sample size. Responder analyses were conducted by calculating the proportion of participants with 15% (minimal), 30% (moderate), and 50% (substantial) improvement from baseline for pain-related outcomes. All data processing and statistical analyses were performed using the SAS version 9.4 (SAS Institute, Inc).

Figure 2 shows the CONSORT (Consolidated Standards of Reporting Trials) diagram for the study. In total, 196 individuals were assessed for initial eligibility, and 92 were excluded because they did not meet the study criteria. In total, 104 participants were enrolled and randomized, and 101 (97.1%) participants completed the baseline surveys (ER: 50/101, 49.5%; WLC: 51/101, 50.5%). A total of 49 participants received the ER intervention, of whom 2 (4%) were lost to follow-up due to feeling unwell or being too busy to complete follow-up surveys. A final sample of 94 participants provided complete follow-up data. Nearly 94% (47/50) of participants in the ER group and 86% (44/51) of participants in the WLC group completed their 3-month follow-up surveys (end of study).

Table 2 shows the baseline demographic characteristics of the study arm. The sample included 101 participants recruited primarily from northern California. The sample was predominantly female (70/101, 69.3%), White (75/101, 74.3%), and with at least some college education (66/101, 65.3%), and a mean age of 49.76 years (SD 13.90; range 24.00-78.00). No significant between-group differences were observed for any demographic variables, demonstrating that randomization was effective.

Table 3 presents the baseline pain and clinical characteristics of the sample by the study arm. The mean pain intensity T score was 62.2 (mean 50, SD 10). No significant differences were observed between the two study groups for any of the variables assessed, except for sleep disturbance (P=.04), which was controlled for in the analysis.

For satisfaction with ER treatment on a scale of 0 (completely dissatisfied) to 10 (completely satisfied), participants (n=47) reported high satisfaction with all items related to overall satisfaction with the class (mean 8.26, SD 1.57), likelihood of recommending the class (mean 8.77, SD 1.43), class relevance (mean 7.94, SD 2.03), the usefulness of the presented information (mean 8.37, SD 1.79), and likelihood of using the skills and information learned (mean 8.66, SD 1.58).

Table 4 reports the between-group comparisons at 2 weeks and 1-3 months posttreatment for all outcomes. Figures 3-12 compare group effects over time for pain-related outcomes. For all figures, the trend of the pain-related variables is displayed over time for participants in both groups. Values in the x-axis refer to the number of weeks in the study, where 0 represents the baseline (roughly 4 days preintervention) time-point. The color band represents the 95% CI for the mean outcome variable. Overlapping bands indicate nonsignificant treatment group differences (P value) of simple main effects within each time-point. The corresponding model effects for each outcome are displayed in Figures 3-12.

We observed a significant treatment effect (P=.005) on pain catastrophizing; on average, the ER group had lower PCS scores than the WLC group (Cohen d=0.50). Separately, we observed a time effect; the average PCS significantly decreased over time for both study groups (time effect; P<.001). Most importantly, the decrease was greater for ER versus WLC (group×time effect; P<.001). At 3-month posttreatment, clinically meaningful reductions in PCS were found for ER but not for WLC (ER: PCS −8.72, 42.25% reduction; WLC: PCS −2.25, 11.13% reduction). Figure 3 displays the average PCS at 3 months by study arm, and Table 4 displays the between-group comparisons for PCS at baseline and posttreatment months. ER was superior to WLC (difference −6.05, 95% CI −9.92 to −2.18; P=.002 at 3-month posttreatment.

The ER group effect size was 0.89, with combined results showing a large effect size and moderate clinical importance. As much as 62% (31/50) of ER and 24% (12/51) of WLC participants achieved a 30% or more reduction in PCS. For ER, 46% (23/50) achieved ≥50% in PCS reduction, whereas for WLC, 12% (6/51) reached that threshold.

For social isolation, we observed a time effect; average social isolation scores significantly decreased over time for both study groups (time effect; P=.02). Most importantly, this was qualified by a significant group×time interaction (P=.02); there was a decrease in social isolation in the ER group but not in the WLC group. No time or interaction effects were observed for pain interference or anger.

Participants in the ER arm provided feedback on their experience using the Zoom platform on a scale from 1 (strongly disagree) to 7 (strongly agree), with higher scores indicating a higher endorsement of each item. Overall, participants reported high satisfaction with the platform, including the ease of operating Zoom (46/50, 92% satisfaction: mean 6.45, SD 1.43), engaging with class material (mean 6.02, SD 1.45), comfort in engaging with Zoom instructor and participants (mean 5.68, SD 1.45), and feeling connected to the instructor (mean 5.43, SD 1.64).

We conducted the first Zoom-delivered randomized controlled trial of the group-based, SSI ER in a sample of individuals with mixed-etiology chronic pain. The primary goal of this study was to evaluate the feasibility, acceptability, and engagement of a single-session Zoom-delivered intervention for chronic pain. We also aimed to determine the preliminary efficacy of Zoom-delivered ER for reducing average pain intensity, pain catastrophizing, pain bothersomeness, and sleep disruption at 1-month posttreatment. Secondary and tertiary pain-related outcomes included physical function, anxiety, depression, and physical function. We also report a more rigorous test of the durability of treatment effects at 3-month posttreatment, as well as report the clinical meaningfulness of effects.

ER demonstrated high feasibility as indexed by high enrollment rate and excellent engagement; 94% (47/50) of participants in the ER group attended the class and completed all surveys through the 3-month duration of the study. The enrollment rate was 53.1% (104/196), which is five times higher than that observed in other studies [12,77] and stands as an index of both the inclusivity of the study protocol (few exclusionary criteria; all types of chronic pain) and the convenience of a single-session pain treatment study. Importantly, participants reported high overall satisfaction with the class (mean 8.26-8.77), the usefulness of presented information, and the likelihood of using the skills and information learned. In addition, participants reported high satisfaction with operating the Zoom platform (46/50, 92% satisfaction), in addition to high satisfaction with engaging with class material, other attendees, and the class instructor. Importantly, of the 9 participants who indicated they had never used the Zoom platform during screening, none requested additional or individual training with the platform. These metrics indicate superior feasibility, accessibility, and engagement with the Zoom-delivered ER intervention.

With respect to preliminary efficacy, time effects showed reductions in pain intensity and most pain-related indices, except for sleep disruption. ER demonstrated superior treatment effects on average pain intensity and pain-related catastrophizing, pain bothersomeness, and sleep disruption over the 3 months following treatment. The between-groups Cohen d effect sizes at 3-month posttreatment ranged from 0.60 to 0.91 for primary and secondary variables. For ER, pre- or posttreatment Cohen d effect sizes ranged from 0.37 to 0.89, demonstrating no clinical importance to moderate clinical importance for reducing pain-related catastrophizing, bothersomeness, and sleep disruption at the end of treatment. The treatment effect sizes across the primary and secondary outcomes suggested that a single-session, digital, skills-based intervention may have clinically meaningful effects on patient-reported outcomes. The durability of the treatment effects was demonstrated at the 3-month follow-up time-point. A greater proportion of participants in the ER group exceeded the thresholds for the clinical importance of effects at 3 months posttreatment.

First, for moderate importance in PCS reduction (≥30% reduction in PCS), 62% (31/50) in ER versus 24% (12/51) in WLC met this threshold. For substantial clinical importance in PCS reduction (≥50% reduction in pain), 46% (23/50) in ER versus 12% (6/51) in WLC met this threshold. PCS reductions for ER (mean −8.72, SD 1.03) exceeded the clinically meaningful threshold of 6.8 reported in the literature [78] and is notably greater than that reported in other studies examining in-person 8-week CBT [7,10,79]. Our findings on PCS are expected and aligned with the literature suggesting that pain catastrophizing, an index of pain coping, is highly responsive to behavioral treatments [7,10,79], including SSI [11,12]. For context, multidisciplinary pain rehabilitation research has shown that PCS reductions of 38% are clinically important and associated with less disability and work status at one-year follow-up [80].

Second, for minimal clinical importance in pain reduction (≥15% reduction in pain), 30% (15/50) in ER versus 6% (3/51) in WLC met this threshold. The moderate clinical importance of pain reduction was low in both groups. Importantly, this study used a PROMIS pain measure with a stricter range (5-point scale) versus a traditional 11-point Numeric Pain Rating Scale. As such, the lack of variability in this measure may have negatively impacted our ability to detect the effects of pain intensity.

The tertiary outcomes yielded mixed findings. Reductions in anxiety and depression over time were found in both groups, and the decline was significantly greater for ER than WLC. However, these reductions did not reach the threshold for clinical significance. Notably, for depression, we observed regression to the mean at 3 months with the treatment effect lost, thus suggesting that more intensive treatment might be needed for durable reductions in depressive symptoms. This may be contextualized within widespread social isolation due to COVID-19, which has been shown to worsen depression symptoms over time [81,82], or other stressors related to COVID-19 or the seasonal context of this study.

Interestingly, no treatment or time effects were observed for physical function, only a trend toward significance (P=.09), which suggests that the study may be underpowered to detect differences or identify subgroups where effect may exist. In addition, due to a coding error, the short form used in this study did not capture upper extremity function, and scale items focused on lower mobility. As this was a mixed-etiology chronic pain study, comprehensive assessment of physical function is warranted to definitively explore its effects. Exploratory analyses revealed interaction effects for reductions in social isolation, which is a notable and important therapeutic target during the COVID-19 pandemic social restrictions. Although the two groups did not differ at 3 months, the interaction effect revealed a decrease in social isolation in the ER group but not in the WLC group. These findings lend support to ER serving as a buffer against the social threat perpetuated by the pandemic [16,83], which can ameliorate its impact on people’s overall health status [65,66].

Although a growing body of research exists on the efficacy of ER for chronic and acute pain management [12,33,36,52], the use of Zoom as a platform to deliver behavioral medicine for chronic pain, particularly single-session treatment, remains novel and understudied. In addition, the extant literature on digital behavioral health research has reported participant treatment engagement rates ranging from 20% to 60% [33,84-86]. Strikingly, the current trial demonstrated a 94% (47/50) engagement rate in the treatment arm. These results also highlight the public interest in the web-based delivery of behavioral medicine interventions as a home-based chronic pain treatment modality. Combined with high participant engagement data and high satisfaction with using the Zoom platform, these data extend prior work [18-25] supporting the utility, user satisfaction, and efficacy of videoconference-delivered interventions for chronic pain.

Several findings should be considered within a number of limitations. First, given that the participant population was on average of higher socioeconomic status, as evidenced by their average education and income levels, the results may not be generalizable to populations of different demographics. No data were available on medical diagnoses, current prescriptions, over-the-counter medication use, or concurrent treatments for either study group. In addition, all data were self-reported, and we did not control for receipt of medical care. In light of COVID-19 restrictions, participants were likely to receive their care via web-based delivery as well. In addition, the pain intensity measure used was a 5-point Numeric Pain Rating Scale, as opposed to the 11-point Numeric Pain Rating Scale [87] typically used in research, limiting variability and may have obscured treatment effects and clinically meaningful findings. Finally, the participants were unblinded to the study group assignment. However, given that ER is low risk and low burden, placebo effects are less concerning. Replication studies were indicated with participants with more demographic variability over a longer follow-up period. Finally, future studies should characterize pain diagnoses and pain types.

Despite these limitations, this web-based study evidenced strong interest and participant engagement with single-session skills-based pain treatment, thus supporting further research and extension of web-based ER into clinical care. Second, the COVID-19 context and web-based delivery modality support the ecological validity of the study findings. Notably, the study was conducted remotely and did not benefit from any in-person contact that occurred when research was conducted in medical treatment settings (ie, halo effects). Additional aspects of methodological rigor included analyst blinding, intention-to-treat analyses, and randomization. Finally, our sample included minimal exclusions, which rendered it highly generalizable to other real-world individuals.

This study examined class cohorts comprising approximately 25 participants, and in-person class sizes ranged between 20 and 85, thus illustrating the promise of scalable pain care. Indeed, ER may help address pain disparities by ensuring rapid and more equitable access to pain treatment. Our findings provide initial evidence that the web-based delivery of ER may efficiently reduce the burden of chronic pain and improve symptom management. Additional research is needed to test the pragmatic comparative effectiveness of web-based ER to other treatments, such as gold-standard 8-week CBT, and to determine the heterogeneity of treatment effects.

In 2019, the US Health and Human Services cited ER as a promising scalable behavioral pain treatment [88]. For the first time, this study determined the preliminary efficacy of the Zoom-delivered class in reducing the burden of chronic pain and improving symptom management. Importantly, the web-based single-session class addresses the rapidly expanding need for alternatives to face-to-face encounters due to the COVID-19 pandemic. Web-based delivery of ER stands to largely improve patient access and engagement because it is adaptable to medical or community settings, is readily extendable to underserved populations (eg, rural locations), and may be offered at a low cost.