A Pilot Investigation of the Effectiveness of the Sana Device in Management of PTSD: A Blinded Randomized Controlled Trial

Study Procedures

Individuals were recruited from a Southeastern VA Health Care System catchment area. Recruited participants completed informed consent and a baseline interview to determine inclusion/exclusion criteria, including the CAPS-5 PTSD assessment. Participants meeting inclusion/exclusion criteria were randomized to either the Sana plus best practices Treatment as Usual or best practices Treatment as Usual (PE, CPT, or WET without wait times were offered to all participants in both conditions). A decentralized trial vendor, Curavit, was used to administer remote assessments and track visits.

The treatment phase lasted 28 days. During the treatment phase, participants in the Sana arm underwent at least 2 daily sessions of device use, including a usage session immediately prior to bedtime. Additional sessions of device treatment were allowed at the participant’s discretion. During the treatment phase, participants completed telehealth visits on days 1, 3, 7, 14 and 21 by a coordinator to check on progress and adherence to expected device use.

All post-baseline assessments were administered remotely. On Day 14 and Day 28, participants completed self-report assessments. Additionally, on Day 28, participants completed a CAPS-5 assessment via a telehealth video or phone call with a VA Health Care System staff member. Assessments of adverse events (AEs), treatment acceptability, and adherence were also collected.

Each participant participated in the study for up to 49 days, including Screening and Baseline (up to 14 days), Treatment (Sana+TAU or TAU alone; 28 days), and Off-boarding (up to 7 days). Participants received assessments at Day -14 to 0 (Screening and Baseline), Day 14 (Mid-Treatment), and Day 28 (Post Treatment).

Measures

PTSD Checklist 5 (PCL-5)

The PCL-5 is a 20-item self-reported measure that assesses the 20 DSM-5 symptoms of PTSD. The PCL-5 can be used to screen individuals for PTSD as well as monitor symptom changes during and after treatment. The PCL-5 was administered at all assessment time points (Baseline, Day 14, and Day 28).

Clinician-Administered PTSD Scale for DSM-5 (CAPS-5)

The CAPS-5 is a 30-item structured interview that can be used to diagnosis and assess PTSD symptoms and severity. The CAPS-5 is administered by clinicians or trained professionals, the full interview takes 45-60 minutes to administer. The CAPS-5 was administered at Baseline and Day 28. If a CAPS-5 was on file that has been completed within 2 weeks prior to consent, that information was used to determine eligibility

General Anxiety Disorder (GAD-7)

The GAD-7 is a self-reported questionnaire for screening and measuring severity of generalized anxiety disorder. The GAD-7 has 7 questions which assess severity of GAD over the last 2 weeks. The GAD-7 was administered at all assessment time points (Baseline, Day 14, and Day 28).

Patient Health Questionnaire-9 (PHQ-9)

The Patient Health Questionnaire 9-item is a validated tool for screening, diagnosing, monitoring and measuring depression severity and scores each of the 9 Diagnostic and Statistical Manual of Mental Disorder, Fourth Edition (DSM-IV) related criteria. The PHQ-9 was administered at all assessment time points (Baseline, Day 14, and Day 28).

Patient Global Impression of Change (PGIC - QOL)

The Patient Global Impression of Change (PGIC - QOL) is a commonly validated measure that is recommended to assess a patient’s overall change, with regards to their person and in satisfaction with their treatment, over the duration of the study. Using this scale, the participant can indicate any changes in their activity limitations, symptoms, emotions and overall quality of life, as related to their pain condition, since the start of the study (1= No change to 7=A great deal better). The PGIC - QOL was administered at Day 28.Treatments.

Treatments

Sana Device

The Sana device is an externally worn mask that physically contacts the skin of the face. The Sana device delivers Audio Visual Stimulation (AVS) in the form of coordinated pulses of light (through closed eyelids) and sound (through earphones) at targeted frequencies. The Sana device is externally communicating only and meets the definition of a “Non-Significant Risk” device.

Best Practices Treatment as Usual (TAU)

Patients assigned to either arm also received best practices TAU and were asked to maintain the use and dosage of any prescribed medications. As mentioned, best practices TAU comprised a referral to immediately begin PE, CPT, or WET. For Sana+TAU, 2 veterans started PE and 13 started WET; for TAU, the numbers were identical (i.e., 2 started PE and 13 started WET).

Data Analysis

For the CAPS-5, PCL-5, GAD-7, and PHQ-9 mean difference between groups from Baseline to Day 28 was tested for statistical significance using a Linear Mixed model with an alpha criterion of 0.05. The model had a categorical factors for group (Sana+TAU | TAU) and Visit (# of visits in the model depend on the specific outcome) as well as the interaction between Group and Visit. The model used a by-subject intercept to account for within subject correlations across visits. Linear mixed models, with a by-subject intercept, are commonly used in repeated measures trials and comparable results to Repeated Measures ANOVA. The PGIC-QOL scale was only collected at the Day-28 visit, and thus, was analyzed using a simple liner model with a factor for group. Pre-planned contrasts from within the 2-way interaction of group (Sana+TAU | TAU) and Visit (Baseline | Day 28) were examined for all outcomes, beside the PGIC-QOL, which used a pre-planned contrast examining group (Sana+TAU | TAU) at Day-28. All pre-planned contrasts used 2-tailed tests of superiority. In addition, post-hoc analyses of within group change for each outcome were conducted to provide additional context.

Final analyses were conducted in 2 populations. The Intent to Treat (ITT) population included all participants who were randomized and completed the baseline assessment. Due to the properties of the linear mixed model used in the analysis, means were automatically estimated even in the presence of missing data, thus for outcomes using the mixed model, no imputation of scores is required. For the PGIC-QOL within the ITT population, multiple imputations were used to estimate missing scores. The results for 20 imputed data sets were pooled using Rubin’s rules [29]. The Per-Protocol (PP) population included participants who were randomized, completed the Baseline and Week 4 assessments, began treatment, and did not have major any protocol deviations. For the PGIC- QOL scale, participants will have to have completed the Day 28 assessment to be included.

All analyses were conducted in the statistical programming language R [30], within the R-Studio IDE [31]. In addition, the following libraries were used; Diplyr [32], Tidyr [33], Lme4 [34], LmerTest [35], Emmeans [36], and Mice [37].

Results

Participant Flow / Dropouts

A total of 48 subjects were consented into the study. A total of 46 (23 Sana+TAU, 23 TAU) were included in the ITT analysis and a 37 (17 Sana+TAU, 20 TAU) were included in the PP analysis. Please see the Consort Chart (Figure 1) for additional details.

Figure 1: Consort Chart of Participant Flow.

Demographics

Demographics for participants included in the ITT analysis are presented in Table 2.

Table 2: Participant Demographics

Endpoint Results

For the PCL-5, there was a statistically significant improvement the Sana+TAU arm over the TAU arm in the ITT population (Mean difference =11.1, p < 0.001) and the PP population (Mean difference = 11.7, p < 0.001). The mean change from Baseline to Day 28 for the PCL-5 for both analysis populations are shown in Figures 2 and 3. Within and between group summary statistics are presented in Table 3.

Figure 2: Plot of Mean Change (bars) and SEM (error bars from Baseline to Day28 for the PCL-5 within both the Sana + TAU and the TAU arms within the ITT Population.

Figure 3: Plot of Mean Change (bars) and SEM (error bars) from Baseline to Day28 for the PCL-5 within both the Sana + TAU and the TAU arms within the PP Population.

Table 3: Between and within group statistics for all outcomes, populations, planned between group contrasts, and post-hoc within group contrasts

CAPS-5 scores were not statistically significantly different between the Sana+TAU arm and the TAU arm in the ITT population (Mean difference =4.7, p = 0.16) and the PP population (Mean difference = 4.2, p < 0.24). Note: low statistical power insofar as at the calculated effect size of 0.41 it would take 95 participants per group, at 80% power with a 2-tailed T-Test with an alpha criterion of 0.05. The mean change from Baseline to Day 28 for the CAPS-5 for both analysis populations are shown in Figures 4 and 5. Within and between group summary statistics are presented in Table 3.

Figure 4: Plot of Mean Change (bars) and SEM (error bars) from Baseline to Day28 for the CAPS-5 within both the Sana + TAU and the TAU arms within the ITT Population.

Figure 5: Plot of Mean Change (bars) and SEM (error bars) from Baseline to Day28 for the CAPS-5 within both the Sana + TAU and the TAU arms within the PP Population.

For the GAD-7, there was a statistically significant improvement the Sana+TAU arm over the TAU arm in the ITT population (Mean difference =4.8, p < 0.001) and the PP population (Mean difference = 4.8, p < 0.001). The mean change from Baseline to Day 28 for the GAD-7 for both analysis populations are shown in Figures 6 and 7. Within and between groups summary statistics are presented in Table 3.

Figure 6: Plot of Mean Change (bars) and SEM (error bars) from Baseline to Day28 for the GAD-7 within both the Sana + TAU and the TAU arms within the ITT Population.

Figure 7: Plot of Mean Change (bars) and SEM (error bars) from Baseline to Day28 for the GAD-7 within both the Sana + TAU and the TAU arms within the PP Population.

For the PHQ-9 assessment, there was a statistically significant improvement the Sana+TAU arm over the TAU arm in the ITT population (Mean difference =4.1, p = 0.003) and the PP population (Mean difference = 4.4, p = 0.001). The mean change from Baseline to Day 28 for the PHQ-9 for both analysis populations are shown in Figures 8 and 9. Within and between groups summary statistics are presented in Table 3.

Figure 8: Plot of Mean Change (bars) and SEM (error bars) from Baseline to Day 28 for the PHQ-9 within both the Sana + TAU and the TAU arms within the ITT Population.

Figure 9: Plot of Mean Change (bars) and SEM (error bars) from Baseline to Day28 for the PHQ-9 within both the Sana + TAU and the TAU arms within the PP Population.

For the PGIC-QOL assessment, there was a statistically significant improvement the Sana+TAU arm over the TAU arm in the ITT population (Mean difference =2.7, p < 0.001) and the PP population (Mean difference = 2.7, p < 0.001). The mean change from Baseline to Day 28 for the PGIC-QOL for both analysis populations are shown in Figures 10 and 11. Within and between group summary statistics are presented in Table 3.

Figure 10: Plot of Means (bars) and SEM (error bars) at Day28 for the PGIC-QOL within both the Sana + TAU and the TAU arms within the ITT Population.

Figure 11: Plot of Means (bars) and SEM (error bars) at Day28 for the PGIC-QOL within both the Sana + TAU and the TAU arms within the PP Population.

Between and within group statistical results are summarized in Table 2 below. The Estimate is the mean difference between groups for between groups contrasts and is the mean change for within group contrasts.

Safety Results

For this study there were a total of 28 AEs. Only 9 of 28 adverse events were reported to be “possibly” or “probably” related to the study device. The remaining 19 were “unrelated” or “unlikely” to be related to the study device. Of the AEs deemed to be “probably” or “possibly” related to the device 7 were considered “mild”, 2 were considered “moderate”, and none were considered “severe”. All treatment related AEs resolved with only one resulting is discontinuation of device use. From this data, it can be concluded that the Sana Device is safe and does not pose any unknown risks to end-users.

Discussion

This randomized, parallel-arm, and controlled study evaluated the effectiveness of a novel AVS device (Sana) for improving PTSD symptoms, anxiety, depression, and quality of life in 46 patients with PTSD. The results of this trial showed significant improvements in the Sana+TAU arm over TAU for PTSD symptoms (PCL-5), anxiety (GAD-7), depression (PHQ-9), and quality of life (PGIC-QOL). Indeed, considering the ITT sample, PTSD scores as measured by the PCL-5 for Sana+TAU showed a 19 point decrease during the 28 day period, compared to a 7 point decrease in the TAU condition. Similarly, anxiety scores on the GAD-7 showed a 5 point drop for Sana+TAU, relative to a less than 1 point drop for TAU alone. Furthermore, depression scores on the PHQ-9 were reduced 8 points in the Sana+TAU group versus 2 points in response to TAU alone. Finally, the 5 point improvement on the general quality of life measure (PGIC-QOL) was also significantly better for Sana+TAU relative to TAU (3 points). Only the CAPS-5 failed to reach statistical significance, despite the 12 point drop for Sana+TAU relative to the 7 point drop for TAU.

Considering that this small sample study was not powered for efficacy findings and considering that the comparator best practices TAU treatments (i.e., PE, CPT, WET) themselves have extremely powerful effects, the fact that all self-report measures across symptom and functioning evinced significant improvement, over and above referral to best practices treatment, is extremely encouraging. We designated the CAPS-5 as the primary outcome measure a priori, and this measure did not show a significant improvement over TAU. However, this preliminary study was knowingly underpowered at about 25 participants per group. Moreover, we compared Sana’s treatment-augmenting effects to referral to the best available PTSD interventions available, rather than to a ‘straw man’ no treatment controls condition to increase the real-world relevance of our findings. As such, statistical power limitations should not be taken lightly. Indeed, at the calculated effect size of 0.41, it would take 95 participants per group to achieve at 80% power (2-tailed t-test, alpha of 0.05).

As mentioned, both conditions included immediate referral to PE, CPT or WET, as well as continued use of any psychotropic medications already on board. As such, this test of the Sana device was in its capacity as an additive intervention. Other attempts at supplementing or altering evidence-based PTSD interventions exist, and either target treatment retention or treatment efficacy. Considering the former, studies of massed PE and CPT show promise for reducing treatment dropout [38,39] , as does including peers directly during in vivo exposure trials [40].However, utility of supplemental intervention components in increasing treatment effects, measured either in terms of PTSD symptoms or overall functioning, have been less consistent. For example, He reviewed the literature and identified 28 augmentation interventions classified according to their mechanism of action (e.g., cognitive enhancers such as rTMS or attention training; emotional distress reducers such as biofeedback training, fear extinction enhancement such as yohimbine or D-cycloserine; integrated interventions addressing sleep or social rehabilitation, etc.) [41]. These authors found that most augmentation efforts were not fruitful, and that the best augmentation interventions were those that targeted cognitive enhancers. Interestingly, Sana falls into that category. Perhaps future research based on contemporary calls for RCTs using psychedelic augmented PTSD treatments will produce fruitful results, but the legal and side effect considerations of those treatments present far more obstacles to use than an intervention such as the Sana device.

Limitations

Limitations of this study and design were evident and should be noted. To approximate real-world conditions as well as best practices standards, participants in both conditions received referral to best practices PTSD treatments in the form of PE, CPT, or WET (i.e., per their choice). However, participants were not required to engage in these treatments. While increasing generalizability, this introduces an obvious experimental confound with respect to both the type of evidence-based treatment obtained, and differential rates of evidence-based treatment actually engaged in across groups. This concern is somewhat mitigated by the fact that Sana achieved its effects in only 28 days, a period during which standard evidence-based treatments are only 25% completed, and only in the initial phases of engaging in ‘active components’ of treatment. In other words, although it would be methodologically preferable to require all participants in all groups to engage in the same evidence-based treatment either with or without Sana, observed results likely emerged prior to any impact such homogeneity of intervention would have had. Other limitations of our study are similar to those characterizing other PTSD treatment outcome research, including heterogeneity of trauma type, frequency, and time since trauma event; and lack of control over variables such as age and TBI status that may impact outcome insofar as brain changes may play a role in treatment mechanism outcomes.

Conclusion

This signal finding trial showed that the Sana Device + best practices Treatment as Usual was more effective than best practices Treatment as Usual alone for improving PTSD symptoms, anxiety, depression, and quality of life for patients with PTSD. The Sana device shows promise as an effective novel treatment for PTSD that is easy to use, largely free of side effects, and works in alongside existing treatments. Future research involving fully powered designs is now warranted.

References

1. Wisco BE, Nomamiukor FO, Marx BP, Krystal JH, Southwick SM, et al (2022) Posttraumatic stress disorder in US military veterans: Results from the 2019–2020 National Health and Resilience in Veterans Study. J Clin Psychiatry, 83(2):39779.

Indexed at, Google Scholar, Cross Ref

2. Gradus JL (2020) PTSD: National Center for PTSD. US Department of Veterans Affairs.

Google Scholar

3. Inoue C, Shawler E, Jordan CH, Moore MJ, Jackson CA. Veteran and military mental health issues.

Google Scholar

4. Gates MA, Holowka DW, Vasterling JJ, Keane TM, Marx BP, et al (2012) Posttraumatic stress disorder in veterans and military personnel: epidemiology, screening, and case recognition. Psychol Serv. 9(4):361.

Indexed at, Google Scholar, Cross Ref

5. Kessler RC, Sonnega A, Bromet E, Hughes M, Nelson CB (2013) Posttraumatic stress disorder in the National Comorbidity Survey. 22-34.

Indexed at, Google Scholar, Cross Ref

6. Kessler RC, Chiu WT, Demler O, Walters EE (2005) Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry. 62(6):617-27.

Indexed at, Google Scholar, Cross Ref

7. Greenberg G, Hoff R (2016) Veterans with PTSD data sheet: National, VISN, and VAMC tables.

Google Scholar

8. Krysinska K, Lester D (2010) Post-traumatic stress disorder and suicide risk: a systematic review. Arch Suicide Res. 14(1):1-23.

Indexed at, Google Scholar, Cross Ref

9. Cougle JR, Keough ME, Riccardi CJ, Sachs-Ericsson N (2009) Anxiety disorders and suicidality in the National Comorbidity Survey-Replication. J Psychiatr Res. 43(9):825-9.

Indexed at, Google Scholar, Cross Ref

10. McCarthy JF, Bossarte RM, Katz IR, Thompson C, Kemp J, et al (2015) Predictive modeling and concentration of the risk of suicide: implications for preventive interventions in the US Department of Veterans Affairs. Am J Public Health. 105(9):1935-42.

Indexed at, Google Scholar, Cross Ref

11. National Center for Health Statistics. Health, United States, 2016, with chartbook on long-term trends in health.

Google Scholar

12. Forehand JA, Dufort V, Gradus JL, Maguen S, Watts BV, et al (2022) Association between post-traumatic stress disorder severity and death by suicide in US military veterans: retrospective cohort study. Br J Psychiatry. 221(5):676-82.

Indexed at, Google Scholar, Cross Ref

13. Krysinska K, Lester D (2010) Post-traumatic stress disorder and suicide risk: a systematic review. Arch Suicide Res. 14(1):1-23.

Indexed at, Google Scholar, Cross Ref

14. Department of Veterans Affairs. Department of Defense (2017) VA/DoD clinical practice guideline for the management of posttraumatic stress disorder and acute stress disorder.

Google Scholar

15. Seligman ME (1995) The effectiveness of psychotherapy: The Consumer Reports study. Am Psychol. 50(12):965.

Indexed at, Google Scholar, Cross Ref

16. Imel ZE, Laska K, Jakupcak M, Simpson TL (2013) Meta-analysis of dropout in treatments for posttraumatic stress disorder. J Consult Clin Psychol. 81(3):394.

Indexed at, Google Scholar, Cross Ref

17. Van Etten ML, Taylor S (1998) Comparative efficacy of treatments for post‐traumatic stress disorder: A meta‐analysis. Clin Psychol Psychother. 5(3):126-44.

Indexed at, Google Scholar, Cross Ref

18. Bradley R, Greene J, Russ E, Dutra L, Westen D (2005) A multidimensional meta-analysis of psychotherapy for PTSD. Am J Psychiatry. 162(2):214-27.

Indexed at, Google Scholar, Cross Ref

19. Hembree EA, Foa EB, Dorfan NM, Street GP, Kowalski J, et al (2003) Do patients drop out prematurely from exposure therapy for PTSD? J Trauma. Stress. 16:555-62.

Indexed at, Google Scholar, Cross Ref

20. Hoge CW, Grossman SH, Auchterlonie JL, Riviere LA, Milliken CS, et al (2014) PTSD treatment for soldiers after combat deployment: Low utilization of mental health care and reasons for dropout. Psychiatr Serv. 65(8):997-1004.

Indexed at, Google Scholar, Cross Ref

21. Covington, C (2019) Not just for soldiers: Civilians with PTSD struggle to find effective therapy.

Google Scholar

22. Becker CB, Zayfert C, Anderson E (2004) A survey of psychologists’ attitudes towards and utilization of exposure therapy for PTSD. Behav Ther. 42(3):277-92.

Indexed at, Google Scholar, Cross Ref

23. Tanielian T, Farris C, Epley C, Farmer CM, Robinson E, et al (2014) Ready to serve. Community-based provider capacity to deliver culturally competent, quality mental health care to veterans and their families.

Google Scholar

24. Davis LL, Schein J, Cloutier M, Gagnon-Sanschagrin P, Maitland J, et al (2022) The economic burden of posttraumatic stress disorder in the United States from a societal perspective. J Clin Psychiatry. 83(3):40672.

Indexed at, Google Scholar, Cross Ref

25. Greenberg PE, Fournier AA, Sisitsky T, Pike CT, Kessler RC (2015) The economic burden of adults with major depressive disorder in the United States (2005 and 2010). J Clin Psychiatry. 76(2):5356.

Indexed at, Google Scholar, Cross Ref

26. Vigo D, Thornicroft G, Atun R (2016) Estimating the true global burden of mental illness. Lancet Psychiatry. 3(2):171-8.

Indexed at, Google Scholar, Cross Ref

27. Tang HY, Riegel B, McCurry SM, Vitiello MV (2016) Open-loop audio-visual stimulation (AVS): a useful tool for management of insomnia? Appl Psychophysiol Biofeedback. 41:39-46.

Indexed at, Google Scholar, Cross Ref

28. Cahn BR, Polich J (2006) Meditation states and traits: EEG, ERP, and neuroimaging studies. Psychol Bull. 132(2):180.

Google Scholar

29. Rubin DB (2004) Multiple imputation for nonresponse in surveys.

Indexed at, Google Scholar, Cross Ref

30. Team RC (2020) RA language and environment for statistical computing, R Foundation for Statistical. Computing.

Google Scholar

31. Team RS (2021) R Studio: integrated development environment for R.

Google Scholar

32. Wickham H, François R, Henry L, Müller K (2021) dplyr: A grammar of data manipulation (Version 1.0. 8).

Indexed at, Google Scholar

33. Wickham H, Vaughan D, Girlich M (2023) tidyr: Tidy messy data (Version 1.3. 0).

Google Scholar

34. Bates D (2014) Fitting linear mixed-effects models using lme4. arXiv preprint.

Indexed at, Google Scholar, Cross Ref

35. Kuznetsova A, Brockhoff PB, Christensen RH (2017) lmerTest package: tests in linear mixed effects models. J Stat Softw. 82(13).

Indexed at, Google Scholar, Cross Ref

36. Lenth R, Singmann H, Love J, Buerkner P, Herve M (2019) Emmeans: estimated marginal means, aka least-squares means (Version 1.3. 4).
37. Van Buuren S, Groothuis-Oudshoorn K (2011) mice: Multivariate imputation by chained equations in R. J Stat Softw. 45:1-67.

Indexed at, Google Scholar, Cross Ref

38. Galovski TE, Werner KB, Weaver TL, Morris KL, Dondanville KA, et al (2022) Massed cognitive processing therapy for posttraumatic stress disorder in women survivors of intimate partner violence. Psychol Trauma. 14(5):769.

Indexed at, Google Scholar, Cross Ref

39. Peterson AL, Blount TH, Foa EB, Brown LA, McLean CP, et al (2023) Massed vs intensive outpatient prolonged exposure for combat-related posttraumatic stress disorder: a randomized clinical trial. JAMA Netw Open. 6(1):e2249422-.

Indexed at, Google Scholar, Cross Ref

40. Hernandez-Tejada MA, Bruce MJ, Muzzy W, Birks A, Macedo e Cordeiro G, et al (2024) Peer support during in vivo exposure homework increases likelihood of prolonged exposure therapy completion. Mil Psychol. 1-0.

Indexed at, Google Scholar, Cross Ref

41. Metcalf O, Stone C, Hinton M, O’Donnell M, Hopwood M, et al (2020) Treatment augmentation for posttraumatic stress disorder: A systematic review. Clin Psychol. 27(1):e12310.

Indexed at, Google Scholar, Cross Ref