Gorman, I., Belser, A. B., Jerome, L., Hennigan, C., Shechet, B., Hamilton, S., Yazar‐Klosinski, B., Emerson, A., & Feduccia, A. A. (2020). Posttraumatic growth after MDMA‐assisted psychotherapy for posttraumatic stress disorder. J ournal of Trauma tic Stress, 33(2), 161–170.
Description and explanation for this orrection
During the MP‐4 study (CT.gov Identifier NCT01958593), which took place between October 2014 and October 2016, there was an ethical violation resulting in a deviation from the protocol by a pair of participating therapists with one of the participants. The pair treated four of six participants in total in the MP‐4 study out of the 60 participants included in the pooled analysis reported in the original manuscript.
Due to the aforementioned ethical violation, data from all four participants treated by this pair of participating therapists have been removed from analysis in this publication correction. The removal of the data does not compromise the conclusions.
Summary of corrections made
In the abstract:
The overall sample size was revised from “N = 60” to “N = 56” due to removal of four participants from analyses. The sample size of the active MDMA dose group was revised from “n = 45” to “n = 43,” and the active control was revised to reflect the dose range covered from “0–40mg of MDMA” to “30–40mg of MDMA.” The sample size of the active control group was revised from “n = 15” to “n = 13.”
The Hedges’ g at primary endpoint was revised as follows from “At primary endpoint, the MDMA group demonstrated more PTG, Hedges' g = 1.14, 95% CI [0.49, 1.78], p < .001; and a larger reduction in PTSD symptom severity, Hedges' g = 0.88, 95% CI [−0.28, 1.50], p < .001, relative to the control group” to “At primary endpoint, the MDMA group demonstrated more PTG, Hedges’ g = 1.26, 95% CI [0.60, 1.92], p < .001; and a larger reduction in PTSD symptom severity, Hedges’ g = 0.97, 95% CI [−0.32, 1.60], p < .001, relative to the active control group.”
The number of participants who did not meet the PTSD diagnostic criteria was revised from “and two thirds of participants (67.2%) no longer met the criteria for PTSD” to “36 participants (70.6%) no longer met the criteria for PTSD” due to removal of four participants from analyses.
In Paragraph 1 of the introductory section, the estimated mean difference between groups changed from “−22.0” to “−23.4,” and the standard error changed from “5.17” to “5.4,” due to removal of four participants from the analyses.
In Paragraph 1 of the Methods section under “Participants,” the following text was added: “During the MP‐4 study, there was an ethical violation by a pair of participating therapists that resulted in a deviation from the protocol. Data from all four participants treated by this pair of participating therapists were removed, and all analyses were updated in the current study to reflect this change. Additional details are provided in the Supplementary Materials.”
In Paragraph 1 of the Methods section under “Procedure,” sample sizes were revised due to the removal of four participants from the analyses:
The original text was: “The three studies tested different active doses of MDMA (75 mg, n = 7; 100 mg, n = 9; 125 mg, n = 29) compared to either active control doses of MDMA (30 mg, n = 7; 40 mg, n = 6) or placebo (0 mg, n = 2). Procedures are described for pooled data from the active MDMA dose group (MDMA group: 75–125 mg of MDMA, n = 45) or the placebo/active control group (control group: 0–40 mg of MDMA, n = 15)”
This has been changed to: “The three studies tested different active doses of MDMA (75 mg, n = 7; 100 mg, n = 9; 125 mg, n = 27) compared to the active control doses of MDMA (30 mg, n = 7; 40 mg, n = 6). Procedures are described for pooled data from the active MDMA dose group (MDMA group: 75–125 mg of MDMA, n = 43) or the active control group (active control group: 30–40 mg of MDMA, n = 13).”
These changes were due to the removal of four participants, two of whom were randomized to the placebo condition, from the analyses. As such, all placebo patients were removed from the analysis, and it is not appropriate to use “placebo” when describing the control group. This change has been implemented throughout the manuscript.
In Paragraph 3 of the Methods section under “Procedure,” the control group name was revised as follows: “After unblinding, all participants randomized to the active control group (30–40 mg)…”
In Paragraph 4 of the Methods section under “Procedure,” the control group name was revised as follows: “All measures were administered at baseline, prior to the first session, when participants received MDMA or an active control. The primary endpoint was 1 month after the second blinded MDMA or control session.”
In Paragraph 1 of the Methods section under “Data analysis,” the control group name and sample size were revised as follows: “Analyses for this paper aggregated participants into either an active MDMA dose group (“MDMA group;” 75–125 mg of MDMA, n = 43) or an active control group (“active control group;” 30–40 mg of MDMA, n = 13)” due to the removal of four participants from analyses.
The following changes were necessary to clarify the 12‐month follow‐up analysis, as only observed values were included in the MMRM analysis and no missing data were imputed, in both the original publication and the corrected analyses. In Paragraph 2 of the “Data analysis” subsection:
The method of missing data handling was revised from “12‐month follow‐up outcomes were compared to scores at baseline and treatment exit (i.e., 2 months after the last active‐dose MDMA session or the last observation carried forward in cases where a participant had completed at least one active‐dose session).” to “12‐month follow‐up outcomes were compared to scores at baseline and treatment exit (i.e., 2 months after the last active‐dose MDMA session).”
The 12‐month follow‐up data were analyzed using a within‐subject mixed model, which was clarified as being conducted only among participants who had received active doses of MDMA in the blinded or open‐label stages (N = 53). The following text was revised from “Because participants from both groups had received active doses of MDMA in the blinded or open‐label stages, outcome scores from all participants were aggregated at three time points and analyzed with a mixed‐effect repeated measure model (MMRM).” The corrected text is: “Among participants from both groups who had received active doses of MDMA in the blinded or open‐label crossover stages ( N = 53), outcome scores from these participants were aggregated at three time points and analyzed with a separate mixed‐effect repeated measure model (MMRM).”
In the “Sample characteristics” subsection within Results, the demographic and baseline characteristic values were updated to reflect the removal of the four participants:
The original text was: “The sample in the pooled dataset (Table 1) was nearly equally split by gender (29 women and 31 men), with a nearly homogenous racial sample (51 White, three Latino/Hispanic, two Native American, and four “other” participants) and a mean participant age of 40 years (range: 22–66). Qualifying traumatic events included war‐related trauma (n = 21), childhood sexual abuse (n = 7), childhood physical abuse (n = 7), adult sexual abuse (n = 8), adult physical abuse (n = 2), accident (n = 2), natural disaster (n = 2), and “other” (n = 9). At enrollment, the mean participant CAPS‐IV total and PTGI scores were 89.4 (SD = 16.71) and 37.8 (SD = 22.78), respectively… No outliers were identified for baseline PTGI or CAPS‐IV total scores in either group (i.e., treatment or control). At baseline, there were no between‐group differences for CAPS‐IV total score, p = .721, or PTGI total score, p = .565.”
This has been changed to: “The sample in the pooled dataset (Table 1) was nearly equally split by gender (29 women and 28 men), with a nearly homogenous racial sample (48 White, three Latino/Hispanic, two Native American, and three “other” participants) and a mean participant age of 40 years (range: 22–66). Qualifying traumatic events included war‐related trauma (n = 21), childhood sexual abuse (n = 7), childhood physical abuse (n = 6), adult sexual abuse (n = 7), adult physical abuse (n = 2), accident (n = 2), natural disaster (n = 1), and “other” (n = 8). At enrollment, the mean participant CAPS‐IV total and PTGI scores were 89.1 (SD = 17.10) and 38.3 (SD = 22.92), respectively…No outliers were identified for baseline PTGI or CAPS‐IV total scores in either group (i.e., treatment or active control). At baseline, there were no between‐group differences for CAPS‐IV total score, p = .491, or PTGI total score, p = .388.”
TABLE 1.
Demographics and baseline characteristics.
| Control | Active | |||||||
|---|---|---|---|---|---|---|---|---|
| (n = 13) | (n = 43) | |||||||
| Variable | M | SD | n | % | M | SD | n | % |
| Age (years) | 39.56 | 10.22 | 40.49 | 12.65 | ||||
| Sex | ||||||||
| Male | 6 | 46.2 | 22 | 51.2 | ||||
| Female | 7 | 53.8 | 21 | 48.8 | ||||
| Race/ethnicity | ||||||||
| White/Caucasian | 11 | 84.6 | 37 | 86.0 | ||||
| Latino/Hispanic | 1 | 7.7 | 2 | 4.7 | ||||
| ative American | 1 | 7.7 | 1 | 2.3 | ||||
| Other/Biracial | 0 | 0.0 | 3 | 7.0 | ||||
| Prestudy ecstasy use | ||||||||
| Yes | 3 | 23.1 | 16 | 37.2 | ||||
| No | 10 | 76.9 | 27 | 62.8 | ||||
| BMI, kg/m2 | 28.88 | 6.99 | 25.99 | 5.05 | ||||
| PTSD duration (months) | 157.23 | 145.15 | 236.51 | 232.04 | ||||
| Baseline CAPS‐IV score | 86.23 | 11.33 | 90 | 18.52 | ||||
| Baseline PTGI score a | 33.67 | 19.48 | 39.61 | 23.89 | ||||
Note. BMI = body mass index; PTSD = posttraumatic stress disorder; CAPS = Clinician‐Administered PTSD Scale; PTGI = Posttraumatic Growth Inventory.
At baseline, 12 participants in the control group and 41 participants in the active group completed the PTGI.
Table 1 was revised to show the updated demographic and baseline characteristic values (all changed data are highlighted below) due to the removal of four participants from the analyses.
In Paragraph 1 of the “Primary outcome” subsection in Results, data were updated to reflect the removal of the four participants:
The original text was: “Hedges' g between‐group effect sizes were large for both measures: CAPS‐IV, 0.88, 95% CI [−0.28, 1.5]; PTGI: 1.14, 95% CI [0.49, 1.78]. At the primary endpoint, more participants in the active group (52.3%) did not meet the PTSD diagnostic criteria than in the control group (33.3%).”
This has been changed to: “Hedges' g between‐group effect sizes were large for both measures: CAPS‐IV, 0.97, 95% CI [−0.32, 1.60]; PTGI: 1.26, 95% CI [0.60, 1.92]. At the primary endpoint, more participants in the active group (54.8%) did not meet the PTSD diagnostic criteria than in the active control group (30.8%).”
In the “Primary outcome” subsection, Paragraph 2, in Results, data were updated to reflect the removal of the four participants:
The original text was: “At baseline, neither group had significant correlations of PTGI and CAPS‐IV scores: placebo group (n = 14): r = −.09, p = .773; MDMA group (n = 43): r = −.01, p = .956. In the MDMA group (n = 41), the CAPS‐IV score change from baseline to primary endpoint was negatively correlated with the change in PTGI score, r = −.42, p = .006, 95% CI [−0.64, −0.13], showing that PTSD symptoms decreased as PTG increased. There was no significant correlation for the control group (n = 14), r = −.04, p = .895. However, a comparison of correlation coefficients between groups did not reach statistical significance, p = .234.”
This has been changed to: At baseline, neither group had significant correlations of PTGI and CAPS‐IV scores: active control group (n = 12): r = −.18, p = .578; MDMA group (n = 41): r = −.01, p = .964. In the MDMA group (n = 40), the CAPS‐IV score change from baseline to primary endpoint was negatively correlated with the change in PTGI score, r = −.40, p = .011, 95% CI [−0.63, −0.09], showing that PTSD symptoms decreased as PTG increased. There was no significant correlation for the active control group (n = 12), r = .14, p = .662. However, a comparison of correlation coefficients between groups did not reach statistical significance, p = .130.
Table 2 was revised to show the updated CAPS‐IV and PTGI total scores at the primary endpoint after two blinded sessions (all changed data are highlighted below) due to the removal of four participants from the analyses.
TABLE 2.
Outcome measures at primary end point after two blinded sessions.
| Control | Active | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| (n = 13) | (n = 43) | ||||||||||
| Variable | n | % | M | SD | n | % | M | SD | p | Hedges' g | 95% CI |
| CAPS‐IV total | |||||||||||
| Baseline | 13 | 86.23 | 11.33 | 43 | 90 | 18.52 | |||||
| Primary endpoint | 13 | 74.77 | 22.92 | 42 | 52.93 | 33.01 | |||||
| Change a PTGI total b | 13 | −11.46 | 16.36 | 42 | −36.17 | 27.66 | < .001 | 0.97 | [− 0.32, 1.60] | ||
| Baseline | 12 | 33.67 | 19.48 | 41 | 39.61 | 23.89 | |||||
| Primary endpoint | 13 | 26.77 | 25.28 | 42 | 61.17 | 25.16 | |||||
| Change a | 12 | −5.17 | 19.09 | 40 | 23.8 | 24.07 | < .001 | 1.26 | [0.60, 1.92] | ||
| PTSD diagnostic met primary endpoint | |||||||||||
| Yes | 9 | 69.2 | 19 | 45.2 | — | — | — | ||||
| No | 4 | 30.8 | 23 | 54.8 | — | — | — | ||||
Note. CAPS = Clinician‐Administered PTSD Scale; PTGI = Posttraumatic Growth Inventory; PTSD = posttraumatic stress disorder.
Change from baseline to primary endpoint.
Different participants were missing data at baseline and primary endpoint. The change was calculated only for participants with scores from both visits.
Figure 1, Panel A was revised to show updated PTGI scores at baseline and all endpoints due to the removal of four participants from the analyses.
- The original Figure 1, Panel A:
- Figure 1, Panel A changed to:
Figure 1, Panel B was revised to show updated CAPS‐IV total scores at baseline and all endpoints due to the removal of four participants from the analyses.
- The original Figure 1, Panel B:
- Figure 1, Panel B changed to:
In the Figure 1 legend, the second sentence was revised to reflect the update to “active control” group. It was changed from “The primary endpoint occurred 1 month after the second blinded MDMA/placebo session” to “The primary endpoint occurred 1 month after the second blinded MDMA/active control session.”
Figure 2 was revised to show the updated correlation of change from baseline PTGI total score to change from baseline CAPS‐IV total score at the primary endpoint due to the removal of four participants from analyses.
- The original Figure 2:
- Figure 2 changed to:
In the “Treatment exit and 12‐month follow‐up” section within Results, the sample size was updated due to the removal of four participants. The second sentence was changed from: “The results of the MMRM (Table 3) demonstrated that PTSD symptom severity and PTG significantly improved at treatment exit compared to baseline (n = 54) for the CAPS‐IV, p < .001, and PTGI, p < .001, with positive gains sustained at the 12‐month follow‐up.” to “The results of the MMRM (Table 3) demonstrated that PTSD symptom severity and PTG significantly improved at treatment exit compared to baseline (n = 53) for the CAPS‐IV, p < .001, and PTGI, p < .001, with positive gains sustained at the 12‐month follow‐up.”
TABLE 3.
Within‐subject outcome measure scores at baseline, treatment exit, and 12‐month follow‐upa.
| Total (N = 53) | |||
|---|---|---|---|
| Variable | LS | SE | p |
| CAPS‐IV total score | |||
| Baseline | 89.67 | 2.40 | |
| Treatment exit | 39.21 | 3.78 | |
| Change: Baseline to treatment exit | −50.45 | 3.63 | < .001 |
| 12‐month follow‐up | 34.63 | 3.47 | |
| Change: Treatment exit to 12‐month follow‐up | −4.58 | 2.86 | .115 |
| Change: Baseline to 12‐month follow‐up | −55.04 | 3.66 | < .001 |
| PTGI total score | |||
| Baseline | 38.28 | 3.15 | |
| Treatment exit | 70.43 | 3.25 | |
| Change: Baseline to treatment exit | 32.15 | 4.39 | < .001 |
| 12‐month follow‐up | 71.79 | 3.45 | |
| Change: Treatment exit to 12‐month follow‐up | 1.36 | 2.58 | .601 |
| Change: Baseline to 12‐month follow‐up | 33.51 | 4.14 | < .001 |
Note. LS = least squares; CAPS = Clinician‐Administered PTSD Scale; PTGI = Posttraumatic Growth Inventory; PTSD = posttraumatic stress disorder.
At 12‐month follow‐up, 15 participants (29.4%) met diagnostic criteria for PTSD; 36 participants (70.6%) did not meet criteria.
Table 3 was revised to reflect the updated within‐subject CAPS‐IV and PTGI total scores at baseline, treatment exit, and 12‐month follow‐up (all changed data are highlighted below). The significance of these outcomes were unchanged following the removal of the four participants.
In the Discussion section, Paragraph 1, Sentence 5, the description of the control group was changed to “active control group” due to the removal of the four participants as follows: “Second, participants who received the MDMA treatment showed significantly more PTG and improvement in PTSD symptoms than those in the active control group, and this effect was of a large magnitude.”
In the Discussion section, Paragraph 1, Sentence 8, the Hedges’ g values for PTGI and CAPS‐IV scores at the primary endpoint were updated due to the removal of the four participants as follows: “After two sessions, active doses of MDMA combined with psychotherapy resulted in decreased PTSD symptom severity and increased PTG with large between‐group effect sizes: g = 0.97 for the CAPS‐IV and g = 1.26 for the PTGI.”
In the Discussion section, the last paragraph, Sentence 2, the description of the control group was changed to “active control group” due to the removal of the four participants as follows: “At posttreatment assessment, the MDMA group experienced higher levels of PTG and larger reductions in PTSD symptom severity compared to the active control group.”
The following text was added (all new text) to the beginning of the Supplemental Materials section for contextualizing these corrections:
New text added to supplement: “During the MP‐4 study (CT.gov identifier, NCT01958593) there was an ethical violation resulting in a deviation from the protocol by a pair of participating therapists. This pair treated four participants in total in the MP‐4 study out of the 60 participants included in the pooled analysis reported in the original manuscript. Data from all four participants treated by this pair of participating therapists have been removed from the analysis in this correction.”
All study design data reported within Supplemental Table 1 for the MP‐4 study were updated to reflect the new sample size from the reanalysis (all changed data are highlighted below).
SUPPLEMENTARY TABLE 1.
Summary of individual study designs.
| Study | NCT # | Location | Population |
Enrolled (n) |
Therapy Teams (n) | MDMA Doses | CAPS‐IV Inclusion Criteria |
Post Second Session Endpoint |
Post Third Session Endpoint |
|---|---|---|---|---|---|---|---|---|---|
| MP‐4 | NCT01958593 | Vancouver, Canada | Various | 2 | 1 | 125 mg (n = 2) | CAPS ≥ 60 |
1 m post 2nd Exp Session |
2 m post 3rd Exp Session |
| MP‐8 | NCT01211405 | Charleston, South Carolina | Veterans, firefighters, police officers | 26 | 1 | 30 mg (n = 7), 75 mg (n = 7), 125 mg (n = 12) | CAPS ≥ 50 |
1 m post 2nd Exp Session |
2 m post 3rd Exp Session |
| MP‐12 | NCT01793610 | Boulder, Colorado | Various | 28 | 9 | 40 mg (n = 6), 100 mg (n = 9), 125 mg (n = 13) | CAPS ≥ 50 |
1 m post 2nd Exp Session |
2 m post 3rd Exp Session |
Abbreviations: NCT = clinicaltrials.gov identifier; CAPS‐IV = Clinician‐Administered PTSD Scale; n = number of participants; m = months; w = weeks.
Supplemental Table 2 was revised to include updated PTGI subscale scores at baseline and all endpoints due to the removal of the four participants (all changed data are highlighted below).
SUPPLEMENTARY TABLE 2.
Summary of PTGI subscales at baseline and endpoints.
| Control | Active | Total | ||||
|---|---|---|---|---|---|---|
| n = 13 | n = 43 | n = 56 | ||||
| Mean | SD | Mean | SD | Mean | SD | |
| Relating to others | ||||||
| Baseline | 9.2 | 8.47 | 10.6 | 8.65 | 10.3 | 8.55 |
| Primary endpoint | 9.7 | 8.97 | 20.7 | 8.81 | 18.1 | 9.95 |
| Treatment exit | 18.5 | 10.73 | 24.9 | 7.65 | 23.5 | 8.69 |
| 12‐month follow‐up | 16.8 | 11.32 | 24.7 | 7.72 | 23.1 | 9.03 |
| New possibilities | ||||||
| Baseline | 7.0 | 5.64 | 10.3 | 7.06 | 9.6 | 6.86 |
| Primary endpoint | 5.8 | 6.66 | 14.4 | 7.06 | 12.4 | 7.82 |
| Treatment exit | 13.4 | 7.66 | 17.4 | 5.35 | 16.5 | 6.08 |
| 12‐month follow‐up | 14.6 | 9.71 | 18.3 | 5.70 | 17.6 | 6.76 |
| Personal strength | ||||||
| Baseline | 8.0 | 4.69 | 8.6 | 5.92 | 8.5 | 5.63 |
| Primary endpoint | 6.0 | 6.38 | 11.9 | 5.86 | 10.5 | 6.45 |
| Treatment exit | 11.3 | 6.47 | 15.8 | 3.55 | 14.8 | 4.66 |
| 12‐month follow‐up | 10.8 | 7.76 | 14.9 | 4.43 | 14.0 | 5.44 |
| Spiritual change | ||||||
| Baseline | 2.4 | 2.68 | 3.0 | 3.34 | 2.9 | 3.18 |
| Primary endpoint | 1.3 | 1.84 | 4.3 | 3.45 | 3.6 | 3.39 |
| Treatment exit | 3.1 | 3.62 | 5.9 | 3.38 | 5.3 | 3.59 |
| 12‐month follow‐up | 3.5 | 4.30 | 6.2 | 3.31 | 5.6 | 3.65 |
| Appreciation of life | ||||||
| Baseline | 7.1 | 3.23 | 7.0 | 3.89 | 7.0 | 3.72 |
| Primary endpoint | 3.9 | 4.23 | 9.8 | 3.97 | 8.4 | 4.71 |
| Treatment exit | 8.8 | 4.71 | 11.9 | 3.36 | 11.2 | 3.85 |
| 12‐month follow‐up | 8.8 | 5.25 | 12.1 | 2.96 | 11.4 | 3.72 |
Abbreviations: PTGI = Posttraumatic Growth Inventory; SD = standard deviation; n = number of subjects.
Supplemental Table 3 was revised to include updated CAPS‐IV subscale scores at baseline and all endpoints due to the removal of the four participants (all changed data are highlighted below).
SUPPLEMENTARY TABLE 3.
Summary of CAPS‐IV subscales at baseline and endpoints.
| Control | Active | Total | ||||
|---|---|---|---|---|---|---|
| n = 13 | n = 43 | n = 56 | ||||
| Mean | SD | Mean | SD | Mean | SD | |
| CAPS B Score | ||||||
| Baseline | 4.1 | 0.86 | 4.1 | 1.06 | 4.1 | 1.01 |
| Primary endpoint | 3.5 | 1.51 | 2.9 | 1.85 | 3.0 | 1.78 |
| Treatment exit | 1.8 | 1.96 | 1.9 | 1.75 | 1.9 | 1.78 |
| 12‐month follow‐up | 2.1 | 2.07 | 2.1 | 1.76 | 2.1 | 1.81 |
| CAPS C Score | ||||||
| Baseline | 5.2 | 1.21 | 5.7 | 1.21 | 5.6 | 1.22 |
| Primary endpoint | 4.5 | 1.61 | 3.2 | 2.50 | 3.5 | 2.37 |
| Treatment exit | 2.6 | 1.88 | 2.2 | 2.09 | 2.3 | 2.03 |
| 12‐month follow‐up | 2.3 | 2.28 | 1.9 | 1.82 | 2.0 | 1.91 |
| CAPS D Score | ||||||
| Baseline | 4.3 | 0.75 | 4.2 | 1.11 | 4.3 | 1.03 |
| Primary endpoint | 3.9 | 1.19 | 2.8 | 1.64 | 3.1 | 1.61 |
| Treatment exit | 2.2 | 1.70 | 2.3 | 1.67 | 2.3 | 1.66 |
| 12‐month follow‐up | 1.9 | 2.07 | 2.0 | 1.35 | 2.0 | 1.51 |
Abbreviations: CAPS‐IV = Clinician‐Administered PTSD Scale; SD = standard deviation; n = number of subjects.
We apologize for this error.