Dialectical Behaviour Therapy to Treat Emotion Dysregulation in Autistic Adults without Intellectual Disability: A Randomised Controlled Trial

Abstract

Introduction

Emotion dysregulation is prevalent in autistic adults without intellectual disability, whereby it has been associated with heightened non-suicidal self-injury and suicidal behaviours. Dialectical behaviour therapy (DBT) has shown to be feasible and preliminary findings suggest that it might reduce emotion dysregulation in this population. Yet studies evaluating the efficacy of DBT in this context are lacking.

Methods

Sixty-three autistic adults presenting with emotional dysregulation as well as self-harm and/or suicidal behaviours were randomised either to the DBT condition (18-week treatment) or to the waiting list condition. Participants completed self-report scales, including emotion dysregulation, alexithymia, depression and quality of life, at 4 time points (pre-, mid-, post-therapy, 6-month follow-up).

Results

Emotion dysregulation improved in the DBT condition relative to the waiting list condition mid-therapy (β₀₁ = −18.59 [−27.67 to −9.44], Pr [β₀₁ < 0] = 1.000), post-therapy (β₀₂ = −31.91 [−41.67 to −22.30], Pr [β₀₂ < 0] = 1.000), with lasting improvements at follow-up. Alexithymia improvement mediated the therapy’s effects on emotion dysregulation. Moreover, depressive symptoms and quality of life improved in the DBT condition relative to the waiting list condition post-therapy, with improvements lasting at follow-up.

Conclusion

DBT was found to be effective in reducing emotional dysregulation in autistic adults presenting with self-harm and/or suicidal behaviour. Additionally, improvements in depression and quality of life were observed post-therapy. Interestingly, the improvements in emotion dysregulation were mediated by a decrease in alexithymia, consistent with research showing that alexithymia is a central mechanism of emotion dysregulation in autistic adults.

Introduction

A growing body of research indicates high rates of non-suicidal self-injury (NSSI) [1, 2] and suicidality (i.e., 3.7 to nine-fold increase in death by suicide relative to the general population) [3, 4] in autistic adults without intellectual disability (henceforth referred to as “autistic adults”) [5]. Recent studies show that emotion dysregulation (ED) and heightened levels of camouflaging – i.e., efforts to mask and/or compensate for autistic traits – might contribute to the high rates of co-occurring psychopathology, NSSI and suicidal behaviours in autistic adults, especially women [1, 6, 7]. For instance, Moseley et al. [1] reported that NSSI is used by autistic adults to regulate painful emotions which are strongly linked to difficulties labelling one’s emotions (i.e., alexithymia), a prerequisite to effective emotion regulation [8]. Importantly, beyond alexithymia, ED in autism spectrum condition (ASC) has been linked to a variety of factors (e.g., temperamental emotional vulnerability, ASC-related intrinsic features, and adverse life experiences) [7], leading to several consequences on adaptive functioning and quality of life [6].

Few studies have focused on treatments targeting ED in autistic people. Specifically, psychotropic medication has only shown short-term efficacy in treating ED [9] and evidence-based psychotherapies are limited, especially for autistic adults [10, 11]. Dialectical behaviour therapy (DBT) [12] has assembled a large body of evidence to treat ED and life-threatening behaviours in borderline personality disorder (BPD) [13]. More recently, DBT has shown to efficiently treat ED in several psychiatric disorders and neurodevelopmental conditions other than autism (e.g., ADHD) [14]. In autistic adults, preliminary findings by Bemmouna et al. [15] support the feasibility and high acceptability of comprehensive DBT, which includes four modes: i.e., individual sessions, skills training groups, consultation team and phone coaching [12]. Importantly, the latter feasibility study found a significant decrease in self-reported ED and life-threatening behaviours in autistic adults following an 18-week comprehensive DBT program. Similarly, in a recent randomised controlled trial (RCT), Huntjens et al. [16, 17] showed that a 26-week comprehensive DBT program led to a reduction in suicide ideation and suicide attempts in a sample of autistic adults. However, their RCT did not explore the effect of DBT on ED, considered to be the underlying cause of suicidal behaviour, nor did it investigate the mechanisms through which DBT improves ED in this population.

The aim of our RCT was to evaluate the efficacy of an 18-week comprehensive DBT program [15] in reducing ED and associated psychopathology compared to a waiting list (WL) control condition for autistic adults presenting with ED and life-threatening behaviours. We hypothesise that self-reported ED – our main outcome – will significantly decrease in the DBT condition compared to the WL condition. Additionally, clinical dimensions associated with ED (i.e., depression, anxiety, impulsivity, alexithymia, suicide ideation, and quality of life) are expected to improve in the DBT condition relative to the WL condition and these results are expected to remain significant at a 6-month follow-up. We also expect increased use of emotion regulation skills and decreased alexithymia and impulsivity levels to mediate the effects of DBT on ED, akin to results in DBT for BPD [18]. Furthermore, since previous findings have outlined the impact of gender [19], autistic traits [19], camouflaging [20], and BPD traits [7] on ED, we hypothesise that these variables will moderate the effect of DBT on ED. This also applies to treatment credibility as this variable has been shown to possibly influence psychological treatment outcomes [21].

Methods

Study Design

The study was a single-centre two-arm parallel RCT conducted from October 2020 to September 2023 at the University Hospital of Strasbourg (CONSORT 2010 checklist is available in online suppl. material S1; for all online suppl. material, see https://doi.org/10.1159/000544717). Recruitment of participants started on April 8, 2021, and ended on November 17, 2022. The study was advertised among mental health professionals, associations for autistic adults, and the special needs unit of the local university hospital. The study was approved by the regional Ethics Committee of the East of France (SI 21.01.21.41923). Written informed consent was obtained from each participant. The study has been preregistered on clinicaltrials.gov (NCT04737707).

Participants

The inclusion criteria were (a) being ≥18 years old; (b) having a formal diagnosis of ASC without intellectual disability supported by the Autism Diagnostic Interview-Revised (ADI-R) [22] and the Autism Diagnostic Observation Schedule, Second Edition revised module 4 (ADOS-2) [23], as well as an IQ assessment based on the Wechsler Adult Intelligence Scale, Fourth Edition (WAIS-IV) indicating an IQ above 80; (c) having a DERS total score above 96 [24, 25] at baseline; (d) presenting with NSSI and/or suicidal behaviours (i.e., suicide attempts) and/or suicide ideation in the 6 months prior to inclusion, whose occurrence was verified during the inclusion interview using a semi-structured interview, and (e) being able to understand and consent to the research aims. Exclusion criteria included (a) having already received DBT, and (b) having a lifetime diagnosis of schizophrenia, schizoaffective disorder, or any unspecified psychotic disorder. Co-occurrent disorders were assessed using the Mini-International Neuropsychiatric Interview (M.I.N.I.) [26]. Recruitment was carried out as the study progressed. Participants received no financial incentives.

Procedure

Individuals eligible for inclusion were randomised with a 1:1 allocation ratio into one of the two conditions: DBT (experimental condition), whereby participants received DBT directly after randomisation and the WL (control condition), whereby participants were on a 5-month WL after randomisation before receiving DBT. The participants in the WL continued their usual psychological and/or pharmacological treatments during the waiting period. Those who had no psychiatric care were referred to a psychiatrist to receive standard suicide prevention care while in the WL.

Randomisation was carried in four blocks of sixteen participants, resulting in eight groups. D.B. and L.W. enrolled eligible participants who have consented to participate. They performed randomisation using the online central randomisation service provided by the eCRF CLEANWEB^® (2019 Telemedicine Technologies™). Randomisation occurred as participants’ recruitment progressed, therefore preventing stratification using baseline variables.

DBT Treatment

The intervention consisted of an 18-week comprehensive DBT [12, 15, 24]. The treatment encompassed the four modes of DBT including a weekly 2 h 15 skills training group session, a weekly 1-h individual session, access to telephone coaching, and a weekly 2-h therapist consultation. The skills training covered the four DBT modules, i.e., mindfulness, emotion regulation, distress tolerance, and interpersonal effectiveness. Key adaptations of DBT to autistic adults are like those implemented in the pilot study by Bemmouna et al. [15]. For further details on the therapy content, see online supplementary material S2.

The therapy was provided by four clinical psychologists including a senior psychologist. All of them were extensively trained in DBT. The senior psychologist provided weekly supervision to the team. The first and/or the last author were the leader facilitators of the group sessions. Both are extensively trained in DBT and relied on the DBT Skills Training Manual [27] to prepare the sessions and maintain optimal adherence. The therapists also participated in weekly consultation team to discuss complex cases and increase adherence to the DBT model.

Outcomes

To evaluate DBT’s efficacy, we administered self-report questionnaires pre-treatment (T0), mid-treatment (T1), post-treatment (T2), and at 6-month follow-up (T3). The WL participants completed the T1 and T2 questionnaires at the same time points as the DBT group.

The DERS was our primary outcome measure, assessing self-reported ED [28]. The DERS has been validated within autistic samples [29] and there is a validated French version [30]. The scale is a self-report measure consisting of 36 items that is easy to apply in clinical trials. It includes 6 dimensions of ED: non-acceptance of emotional responses (Non-acceptance), difficulty engaging in goal-directed behaviours when distressed (Goals), impulse control difficulties (Impulse), lack of emotional awareness (Awareness), limited access to regulation strategies (Strategies), and lack of emotional clarity (Clarity). Higher scores indicate greater difficulties. The DERS has been found to have adequate internal consistency (Cronbach’s α = 0.93), test–retest reliability (r = 0.88), and construct and predictive validity among a college sample [28]. The DERS scores were calculated based on McVey et al.’s [29] adaptation for autistic adults. In the absence of an official cut-off, we used the one by Neacsiu et al. [24] to set the score of 97 as the threshold for high ED.

Secondary outcome measures included the Barratt Impulsiveness Scale-short form (BIS-15) [31], Eight-item General Alexithymia Factor Score (GAFS-8) [32], the Beck Anxiety Inventory (BAI) [33], the Beck Depression Inventory – Second Edition (BDI-II) [34], the Beck Scale for Suicide Ideation (BSS) [35], the DBT Ways of Coping Checklist (DBT-WCCL) [36] to assess DBT strategies and dysfunctional coping strategies, the Five Facet Mindfulness Questionnaire (FFMQ) [37], and the Abbreviated World Health Organization Quality of Life Questionnaire (WHOQoL-BREF) [38]. The BIS-15 (impulsivity), GAFS-8 (alexithymia), DBT-WCCL (DBT skills use), and FFMQ (mindfulness) were also used to explore potential mediation effects on the DERS outcomes. In addition to self-report scales, suicidal behaviours and hospitalisations were tracked throughout the intervention. Participants in the WL condition were asked to report any suicidal behaviours’ occurrence during the WL period. At the 6-month follow-up, participants reported whether suicide attempts had occurred and their subjective evaluation of the potential impacts of DBT on NSSI and suicide ideation. We collected these clinical outcomes only at follow-up to tackle sustainable changes post-DBT.

Moreover, the following dimensions were assessed pre-DBT to explore potential moderation effects: gender, autistic traits measured with the Autism Spectrum Quotient (AQ) [39], BPD symptoms measured with the short form of the Borderline Symptom List (BSL-23) [40], camouflaging behaviour assessed with the Camouflaging Autistic Traits Questionnaire (CAT-Q) [20], and the treatment credibility and client expectancy for improvement evaluated via the Credibility/Expectancy Questionnaire (CEQ) [41].

Statistical Analyses

The minimal sample size was estimated based on assumptions on the Difficulties in Emotion Regulation Scale (DERS) [28] scores distributions based on Bemmouna et al.’s [15] pilot study. The margin of error was set at 0.05 and the minimal statistical power at 90%. A sample of at least 48 participants was required to demonstrate a reduction of at least ten points in the DERS mean score mi-therapy (T1) compared to baseline (T0), with a power of 95%, and a reduction of at least 15 points post-therapy (T2), and at 6-month follow-up (T3) compared to T0, with a power of 91%. Sixteen additional participants were included (i.e., a third of the sample) to account for potential dropouts.

The data were analysed according to a Bayesian paradigm using R software, Version 4.3.1 (The R Foundation for Statistical Computing; https://www.r-project.org) and JAGS software, Version 4.3 [42]. Between-condition comparisons were made at T0, T1, and T2, but not at T3. Indeed, when collecting T3 data for the DBT groups, the WL participants had switched to the DBT condition.

To meet the primary objective (i.e., to evaluate the changes in ED), inferential analysis was conducted to study the evolution of the DERS mean score over time, using Bayesian mixed linear regression including time (effect of treatment) as fixed effect and participants as random effect. To meet the secondary objectives (i.e., to evaluate the changes in depression, anxiety, impulsivity, alexithymia, suicide ideation, and quality of life), Bayesian mixed linear regressions were used to study the evolution of the different variables. Sensitivity analyses were carried out by varying the parameters of the prior to obtain an optimistic prior (in favour of an effect of DBT) and a pessimistic a prior (against an effect of DBT).

For each analysis, the posteriors of the parameters of interest (e.g., proportion, mean, regression coefficient) were estimated using the Markov Chain Monte Carlo (MCMC) method. The default number of iterations was 100,000, after deleting the first 10,000 and a thinning of two (210,000 iterations were therefore performed). Convergence was estimated graphically. Autocorrelation was estimated graphically and, if necessary, the number of iterations was increased to raise the step size of the values retained, with the aim of reducing autocorrelation as much as possible.

Results are presented as coefficients, together with 95% credibility intervals (CrI), and the probability that the coefficient would be higher than 0 was calculated based on the posterior distribution (probability coefficient >0, hereinafter abbreviated “Pr > 0”). It should be noted that these probabilities must not be confused with the p value of classical (frequentist) statistical analyses. The reported Pr > 0 value indicates an effect for values close to 1 (with a high probability if greater than 97.5%) [43]. The Bayesian framework does not imply a strict cut-off for this probability but rather a contextualised probability interpretation.

Feasibility and acceptability were assessed via the attrition rate (the percentage of dropouts), the attendance rate at group sessions, and satisfaction measured post-DBT using the Client Satisfaction Questionnaire (CSQ-8) [44]. For further details on the assessment tools, see online supplementary material S2.

Results

Sample Description

One hundred and nineteen individuals were assessed for eligibility for the study. Fifty-five (46%) were excluded (Fig. 1). Sixty-four (54%) were randomised. Autistic traits, measured via the AQ, were high within our sample (M = 33.4, SD = 12.4) [39]. Participants’ demographic and clinical variables are presented in Table 1.

Fig. 1.

Study participants’ inclusion flow chart.

Table 1.

Sample description

	Total sample	DBTa	WL
Demographics
n (%)	63 (100)	58 (92)	31 (49)
Mean age (SD), years	29 (9.19)	29 (9.49)	31 (10.10)
Age range (min-max)	18–67	18–67	18–67
Gender, n (%)
Women	30 (48)	27 (47)	11 (35)
Men	28 (44)	26 (45)	19 (61)
Non-binary	5 (8)	5 (9)	1 (3)
Marital status, n (%)
Single	33 (52)	32 (55)	16 (52)
Married/in relationship	27 (43)	23 (40)	8 (26)
Divorced	3 (5)	3 (5)	2 (6)
Having children, n (%)	8 (13)	6 (10)	6 (19)
Professional status, n (%)
Professionally active	21 (33)	17 (29)	11 (35)
Student	21 (33)	21 (36)	11 (35)
Unemployed	20 (32)	19 (33)	8 (26)
Retired	1 (2)	1 (2)	1 (3)
Educational status, n (%)
College graduate	44 (70)	41 (71)	23 (74)
High school degree or less	19 (30)	17 (29)	8 (26)
Living situation, n (%)
Alone	28 (44)	26 (45)	13 (42)
With parents	17 (27)	16 (28)	8 (26)
With partner with or without children	13 (21)	12 (21)	6 (19)
Flatsharing	3 (5)	3 (5)	3 (10)
Alone with children	2 (3)	1 (2)	1 (3)
Clinical variables
Recent ASC diagnosis (<1 year), n (%)	38 (60)	37 (64)	15 (48)
Co-occurring BPD, n (%)	6 (10)	5 (9)	2 (6)
Co-occurring ADHD, n (%)	27 (43)	25 (43)	17 (55)
Current psychotropic medication, n (%)	51 (81)	47 (81)	26 (84)
Antidepressant	31 (49)	30 (52)	16 (52)
Neuroleptics	17 (27)	17 (29)	7 (23)
Benzodiazepines	14 (22)	12 (21)	6 (19)
Psychostimulants	14 (22)	13 (22)	9 (29)
Mood stabilisers	4 (6)	4 (7)	2 (6)
Other psychotropic medication	8 (13)	6 (10)	4 (13)
Current psychological and psychiatric care, n (%)
Psychological + psychiatric follow-up	25 (40)	22 (38)	12 (39)
Psychiatric follow-up only	26 (41)	25 (43)	15 (48)
Psychological follow-up only	2 (3)	1 (2)	2 (6)
No follow-up	10 (16)	10 (17)	2 (6)
Current NSSI, n (%)	37 (59)	34 (63)	17 (55)
Current skin cutting	18 (29)	16 (29)	6 (19)
Frequent skin cuttingb	7 (11)	6 (10)	3 (10)
Current suicide ideation (SI), n (%)	59 (94)	54 (93)	30 (97)
History of suicidal behaviours (SB), n (%)	32 (51)	28 (48)	14 (45)
SB in the year prior to inclusion	12 (19)	11 (19)	5 (16)
History of hospitalisation in psychiatry, n (%)	28 (44)	25 (43)	13 (42)
Last hospitalisation for NSSI and/or SI and/or SB	24 (38)	21 (36)	11 (35)
Hospitalisation in the last year prior to inclusion	14 (22)	12 (21)	5 (16)
Participants with NSSI only, n (%)	4 (6)	4 (7)	1 (3)
Participants with SI only, n (%)	15 (24)	14 (24)	9 (29)
Participants with SI and SB, n (%)	11 (17)	10 (16)	5 (16)
Participants with NSSI and SI, n (%)	12 (19)	12 (21)	7 (23)
Participants with NSSI, SI and SB, n (%)	21 (33)	18 (31)	9 (29)

^aBoth participants who received DBT directly and those who received it after the WL period.

^bDaily/many times a week.

Efficacy Outcomes

Condition × Time Interaction Effects on the Efficacy Measures

Results between the two conditions in T1 and T2 are presented in Table 2. The evolution of the questionnaire scores for the DBT condition is presented in Table 3. Sensitivity analyses conducted (with optimistic and pessimistic priors) revealed similar effects of DBT on the DERS.

Table 2.

Comparison of the efficacy measures’ evolution between conditions at T1 and T2

	DBT						WL						DBT vs. WL
	T0		T1		T2		T0		T1		T2		difference at T1				difference at T2
	M	SD	M	SD	M	SD	M	SD	M	SD	M	SD	β01b	95% CrI	Pr (β01 < 0)c	Cohen’s d	β02b	95% CrI	Pr (β02 < 0)c	Cohen’s d
DERS	125.84	16.08	100.63	22.99	84.06	17.98	129.45	14.85	125.18	20.19	122.09	15.71	−18.59	[−27.67 to −9.44]	1.000*	−1.113 a	−31.91	[−41.67 to −22.30]	1.000*	−2.198 a
Clarity	9.53	2.34	7.84	3.02	6.49	2.54	9.68	2.60	9.25	3.19	8.44	2.00	−1.18	[−2.48 to 0.10]	0.965	−0.457	−1.44	[−2.80 to −0.07]	0.980*	0.816 a
Awareness	31.60	6.18	25.80	7.73	22.14	7.04	30.19	7.09	30.29	6.78	30.13	6.78	−5.64	[−8.56 to −2.72]	1.000*	−0.605	−9.15	[−12.25 to −6.05]	1.000*	−1.147 a
Impulse	20.15	5.05	14.84	4.91	12.08	4.42	21.74	4.32	20.61	4.61	19.22	4.45	−4.13	[−6.54 to −1.51]	0.999*	−1.199 a	−5.78	[−8.45 to −3.16]	1.000*	−1.611 a
Non-acceptance	20.00	6.67	16.51	6.56	13.39	5.41	22.26	5.98	21.18	6.72	20.52	6.22	−2.06	[−4.99 to 0.83]	0.920	−0.706	−4.45	[−7.55 to −1.34]	0.998*	−1.257 a
Goals	22.06	2.88	18.98	4.70	16.16	4.39	22.13	3.90	21.39	4.35	21.78	3.42	−2.43	[−4.31 to −0.52]	0.994*	−0.527	−5.39	[−7.41 to −3.39]	1.000*	−1.367 a
Strategies	22.51	4.30	16.56	5.60	13.80	4.64	23.45	2.82	22.46	4.66	22.00	4.22	−4.53	[−6.92 to −2.19]	1.000*	−1.110 a	−7.28	[−9.82 to −4.73]	1.000*	−1.818 a
BIS-15	35.31	7.29	34.84	7.33	33.65	7.33	37.26	8.40	38.18	8.31	36.35	7.81	−0.31	[−2.61 to 1.99]	0.605	−0.434	−0.28	[−2.74 to 2.23]	0.582	−0.350
Non-planning	11.74	3.05	11.60	3.71	11.10	4.17	12.52	3.81	13.29	3.43	12.57	3.68	−0.62	[−2.02 to 0.78]	0.809	−0.466	−0.41	[−1.90 to 1.08]	0.709	−0.363
Motor impulsivity	10.33	3.22	10.30	3.72	9.90	3.20	11.45	3.58	11.50	3.50	11.00	2.78	0.41	[−0.82 to 1.62]	0.256	−0.329	0.01	[−1.29 to 1.30]	0.493	−0.358
Attentional impulsivity	13.24	3.00	12.94	3.05	12.65	3.04	13.29	3.18	13.39	3.13	12.78	3.29	−0.16	[−1.22 to 0.90]	0.615	−0.147	0.09	[−1.02 to 1.21]	0.432	−0.041
GAFS-8	30.82	6.47	25.14	9.17	21.89	7.51	30.27	5.66	28.82	7.87	28.87	6.88	−3.86	[−6.90 to −0.79]	0.994*	−0.422	−5.91	[−9.06 to −2.76]	1.000*	−0.953 a
BAI	28.59	12.12	25.71	12.96	20.37	10.88	27.17	12.47	27.82	14.64	23.50	12.17	−5.01	[−10.4 3 to 0.37]	0.966	−0.156	−4.58	[−10.39 to 1.13]	0.943	−0.277
BDI-II	25.36	11.65	20.24	12.45	13.90	9.44	26.97	7.67	23.14	13.92	22.24	10.29	−1.20	[−5.89 to 3.45]	0.697	−0.223	−5.89	[−10.88 to −0.84]	0.989*	−0.860 a
BSS	9.88	8.75	7.35	9.04	4.16	6.25	11.55	8.03	8.33	8.92	6.95	8.11	1.06	[−2.65 to 4.76]	0.716	−0.109	0.25	[−3.53 to 3.94]	0.554	−0.407
DBT-WCCL
General dysfunctional coping	27.64	8.18	25.42	9.03	21.72	8.91	29.13	6.98	27.85	6.46	24.62	7.69	−1.06	[−5.27 to 3.18]	0.691	−0.296	−0.24	[−4.76 to 4.23]	0.534	−0.337
Blaming others	5.68	4.20	5.88	4.29	5.25	4.50	6.97	4.34	5.70	4.84	5.19	4.56	1.47	[−0.42 to 3.38]	0.065	−0.039	1.50	[−0.54 to 3.53]	0.073	−0.013

													β01b	95% CrI	Pr (β01 > 0)c	Cohen’s d	β02b	95% CrI	Pr (β02 > 0)c	Cohen’s d
DBT-WCCL
Skills use	49.75	18.35	66.82	21.12	71.46	19.43	50.47	16.98	52.07	19.29	55.71	20.64	14.98	[6.03 to 23.96]	0.999*	0.719	16.56	[6.87 to 26.22]	1.000*	0.795
FFMQ	101.45	17.07	112.69	23.27	123.56	19.94	102.39	15.74	104.22	17.78	107.17	15.96	7.70	[−0.48 to 15.78]	0.968	0.393	17.33	[8.73 to 25.92]	1.000*	0.873 a
Observation	27.60	6.41	27.23	8.20	28.27	6.96	28.39	7.57	28.52	7.78	28.09	7.10	−0.01	[−3.00 to 3.08]	0.494	−0.159	1.47	[−1.77 to 4.60]	0.820	0.026
Description	18.20	6.69	21.00	6.64	24.06	7.06	19.13	6.27	20.33	7.49	21.09	6.47	1.53	[−0.84 to 3.90]	0.897	0.096	3.49	[0.99 to 6.02]	0.997*	0.433
Aware actions	20.36	5.32	22.71	7.87	23.71	8.08	21.48	6.58	21.59	7.01	22.04	5.32	1.55	[−1.04 to 4.12]	0.882	0.147	1.72	[−1.06 to 4.45]	0.89	0.227
Non-judgemental inner critic	20.73	7.73	23.59	8.89	27.73	8.01	19.10	6.58	19.19	6.96	20.35	7.19	2.29	[−1.32 to 5.81]	0.892	0.533	5.97	[2.11 to 9.78]	0.999*	0.951 a
Non-reactivity	14.56	4.60	18.16	5.87	19.79	5.08	14.29	3.82	14.59	5.10	15.61	4.04	2.70	[0.11 to 5.27]	0.979*	0.634	4.95	[2.31 to 7.61]	1.000*	0.876 a
WHOQoL-BREF
Physical health	17.62	4.12	19.36	4.69	21.12	4.64	18.29	3.88	18.33	4.34	18.30	3.61	1.39	[−0.58 to 3.60]	0.920	0.224	2.87	[0.83 to 4.94]	0.997*	0.648
Psychological health	13.13	3.99	15.16	4.65	17.06	4.86	13.16	3.51	13.67	4.27	13.70	4.38	1.31	[−0.75 to 3.41]	0.892	0.330	3.13	[0.94 to 5.32]	0.997*	0.714
Social relationships	7.62	2.63	8.58	2.78	9.20	3.13	7.61	2.43	7.52	2.23	8.00	2.37	1.04	[−0.30 to 2.38]	0.937	0.408	1.09	[−0.32 to 2.51]	0.938	0.413
Environment	24.40	6.74	25.32	6.80	26.45	7.83	25.65	5.45	26.74	5.86	26.17	5.88	0.13	[−2.53 to 2.77]	0.460	−0.219	1.02	[−1.81 to 3.82]	0.761	0.038

*Pr > 0.975.

^aLarge effect size.

^bNumber of points of difference between DBT and WL (for instance, at T1 the DERS mean score decreased by 18.52 points more at the DBT condition than WL).

^cPr (β < 0) fo scales expected to decrease and Pr (β > 0) expected to increase following treatment.

Table 3.

Efficacy measures’ evolution over time compared to T0 in the DBT condition

	T0		T1		T2		T3		Difference at T1			Difference at T2			Difference at T3
	M	SD	M	SD	M	SD	M	SD	β01a	95% CrI	Pr (β01 < 0)b	β02a	95% CrI	Pr (β02 < 0)b	β03a	95% CrI	Pr (β03 < 0)b
DERS	125.84	16.08	100.63	22.99	84.06	17.98	89.47	22.58	−24.11	[−29.95 to −18.23]	1.000*	−41.02	[−46.85 to −35.19]	1.000*	−36.71	[−42.80 to −30.53]	1.000*
Clarity	9.53	2.34	7.84	3.02	6.49	2.54	6.47	2.33	−1.64	[−2.44 to −0.84]	1.000*	−2.92	[−3.75 to −2.09]	1.000*	−3.03	[−3.89 to −2.17]	1.000*
Awareness	31.60	6.18	25.80	7.73	22.14	7.04	22.03	6.14	−5.82	[−7.65 to −3.99]	1.000*	−9.46	[−11.37 to −7.57]	1.000*	−10.10	[−12.03 to −8.14]	1.000*
Impulse	20.15	5.05	14.84	4.91	12.08	4.42	13.92	6.01	−5.20	[−6.78 to −3.64]	1.000*	−8.13	[−9.75 to −6.52]	1.000*	−6.34	[−8.02 to −4.68]	1.000*
Non-acceptance	20.00	6.67	16.51	6.56	13.39	5.41	13.55	5.62	−3.46	[−5.29 to −1.62]	1.000*	−6.71	[−8.57 to −4.82]	1.000*	−6.50	[−8.44 to −4.57]	1.000*
Goals	22.06	2.88	18.98	4.70	16.16	4.39	17.82	4.91	−3.27	[−4.47 to −2.06]	1.000*	−5.93	[−7.13 to −4.72]	1.000*	−4.51	[−5.78 to −3.23]	1.000*
Strategies	22.51	4.30	16.56	5.60	13.80	4.64	15.68	4.77	−5.62	[−7.11 to −4.11]	1.000*	−8.87	[−10.41 to −7.35]	1.000*	−7.16	[−8.74 to −5.58]	1.000*
BIS-15	35.31	7.29	34.84	7.33	33.65	7.33	33.74	7.06	0.27	[−1.22 to 1.74]	0.643	−1.38	[−2.88 to 0.09]	0.033	−1.30	[−2.84 to 0.25]	0.048
Non-planning	11.74	3.05	11.60	3.71	11.10	4.17	11.24	3.40	0.02	[−0.88 to 0.90]	0.521	−0.44	[−1.33 to 0.45]	0.165	−0.36	[−1.29 to 0.57]	0.224
Motor impulsivity	10.33	3.22	10.30	3.72	9.90	3.20	9.61	3.20	0.31	[−0.48 to 1.07]	0.784	−0.43	[−1.22 to 0.34]	0.138	−0.65	[−1.46 to 0.17]	0.059
Attentional impulsivity	13.24	3.00	12.94	3.05	12.65	3.04	12.89	2.58	−0.11	[−0.79 to 0.57]	0.372	−0.53	[−1.22 to 0.15]	0.062	−0.32	[−1.03 to 0.41]	0.189
GAFS-8	30.82	6.47	25.14	9.17	21.89	7.51	21.82	7.46	−5.35	[−7.24 to −3.46]	1.000*	−8.34	[−10.27 to −6.44]	1.000*	8.45	[10.46 to −6.49]	1.000*
BAI	28.59	12.12	25.71	12.96	20.37	10.88	24.47	12.24	−4.03	[−7.39 to −0.71]	0.991*	−7.86	[−11.30 to −4.42]	1.000*	−4.86	[−8.47 to −1.20]	0.996*
BDI-II	25.36	11.65	20.24	12.45	13.90	9.44	16.61	10.34	−4.34	[−7.24 to −1.43]	0.998*	−10.31	[−13.35 to −7.35]	1.000*	−8.55	[−11.61 to −5.48]	1.000*
BSS	9.88	8.75	7.35	9.04	4.16	6.25	4.00	7.10	−2.13	[−4.63 to 0.16]	0.966	−4.63	[−6.88 to −2.41]	1.000*	−4.97	[−7.28 to −2.66]	1.000*
DBT-WCCL
General dysfunctional coping	27.64	8.18	25.42	9.03	21.72	8.91	23.00	8.71	−2.20	[−4.83 to 0.44]	0.949	−5.27	[−7.93 to −2.56]	1.000*	−4.36	[−7.11 to −1.53]	0.999*
Blaming others	5.68	4.20	5.88	4.29	5.25	4.50	5.55	4.00	0.34	[−0.83 to 1.51]	0.289	−0.10	[−1.29 to 1.10]	0.562	0.24	[−1.02 to 1.48]	0.35

									β01a	95% CrI	Pr (β01 > 0)b	β02a	95% CrI	Pr (β02 > 0)b	β03a	95% CrI	Pr (β03 > 0)b
DBT-WCCL
Skills use	49.75	18.35	66.82	21.12	71.46	19.43	66.00	16.77	16.72	[11.07 to 22.31]	1.000*	22.31	[16.64 to 28.13]	1.000*	18.09	[12.21 to 24.14]	1.000*
FFMQ	101.45	17.07	112.69	23.27	123.56	19.94	120.55	22.14	10.51	[5.30 to 15.60]	1.000*	23.20	[18.00 to 28.39]	1.000*	20.64	[15.27 to 26.07]	1.000*
Observation	27.60	6.41	27.23	8.20	28.27	6.96	28.50	6.32	−0.34	[−2.27 to 1.56]	0.362	0.92	[−1.01 to 2.83]	0.828	1.46	[−0.52 to 3.45]	0.926
Description	18.20	6.69	21.00	6.64	24.06	7.06	23.47	7.14	2.85	[1.36 to 4.33]	1.000*	5.64	[4.14 to 7.16]	1.000*	5.24	[3.67 to 6.80]	1.000*
Aware actions	20.36	5.32	22.71	7.87	23.71	8.08	23.63	7.27	1.91	[0.31 to 3.51]	0.991*	3.19	[1.54 to 4.84]	1.000*	3.16	[1.46 to 4.86]	1.000*
Non-judgemental inner critic	20.73	7.73	23.59	8.89	27.73	8.01	26.58	8.63	2.75	[0.52 to 4.96]	0.992*	7.72	[5.44 to 10.03]	1.000*	6.28	[3.92 to 8.62]	1.000*
Non-reactivity	14.56	4.60	18.16	5.87	19.79	5.08	18.37	6.18	3.52	[1.94 to 5.09]	1.000*	5.91	[4.32 to 7.50]	1.000*	4.68	[2.96 to 6.34]	1.000*
WHOQoL-BREF
Physical health	17.62	4.12	19.36	4.69	21.12	4.64	20.11	5.72	1.30	[0.09 to 2.52]	0.982*	2.88	[1.66 to 4.09]	1.000*	2.01	[0.74 to 3.26]	0.999*
Psychological health	13.13	3.99	15.16	4.65	17.06	4.86	16.76	4.20	1.73	[0.42 to 3.03]	0.995*	3.84	[2.53 to 5.14]	1.000*	3.77	[2.40 to 5.13]	1.000*
Social relationships	7.62	2.63	8.58	2.78	9.20	3.13	9.00	3.03	0.94	[0.11 to 1.77]	0.987*	1.53	[0.69 to 2.37]	1.000*	1.45	[0.57 to 2.32]	0.999*
Environment	24.40	6.74	25.32	6.80	26.45	7.83	26.92	6.59	0.81	[−0.86 to 2.45]	0.831	1.76	[0.07 to 3.44]	0.980*	2.32	[0.54 to 4.07]	0.995*

*Pr > 0.975.

^aNumber of points of difference between time points (for instance, at T1 the DERS mean score decreased by 23.96 points compared to T0).

^bPr (β < 0) fo scales expected to decrease and Pr (β > 0) expected to increase following treatment.

Regarding our primary outcomes, the DERS total mean score decreased significantly more (with large effect sizes) in the DBT condition than the WL condition mid-therapy (β₀₁ = −18.59 [−27.67 to −9.44], Pr [β₀₁ < 0] = 1.000) and post-therapy (β₀₂ = −31.91 [−41.67 to −22.30], Pr [β₀₂ < 0] = 1.000; Fig. 2). The DERS scores were no longer above the severity cut-off of 96 [24, 25] at post-therapy, and they remained stable at 6-month follow-up (Table 3). Moreover, the “Awareness,” “Impulse,” “Goals” and “Strategies” mean subscale scores decreased significantly more in the DBT condition compared to the WL condition at mid-therapy and post-therapy.

Fig. 2.

Condition × time interaction effects on the DERS total mean over time.

Regarding the secondary outcomes, the BIS-15 total mean score, measuring impulsivity, did not differ significantly between the DBT and the WL conditions mid-therapy (β₀₁ = −0.31 [−2.61 to 1.99], Pr [β₀₁ < 0] = 0.605) and post-therapy (β₀₂ = −0.28 [−2.74 to 2.23], Pr [β₀₂ < 0] = 0.582). The GAFS-8 mean score, measuring alexithymia, decreased significantly (with a large effect size) more in the DBT condition compared to the WL condition mid-therapy (β₀₁ = −3.86 [−6.90 to −0.79], Pr [β₀₁ < 0] = 0.994) and post-therapy (β₀₂ = −5.91 [−9.06 to −2.76], Pr [β₀₂ < 0] = 1.000), with lasting results at follow-up (Table 3). The BAI mean score, measuring anxiety, did not differ between the DBT and the WL conditions mid-therapy (β₀₁ = −5.01 [−10.4 3 to 0.37], Pr [β₀₁ < 0] = 0.966) and post-therapy (β₀₂ = −4.58 [−10.39 to 1.13], Pr [β₀₂ < 0] = 0.943). This was also the case for the BSS mean score, measuring suicide ideation (β₀₁ = 1.06 [−2.65 to 4.76], Pr [β₀₁ < 0] = 0.716; β₀₂ = 0.25 [−3.53 to 3.94], Pr [β₀₂ < 0] = 0.554). However, the BDI-II mean score, measuring depression, decreased significantly more in the DBT condition compared to the WL condition post-therapy (β₀₂ = −5.89 [−10.88 to −0.84], Pr [β_{0 2} < 0] = 0.989), and the scores remained stable at the 6-month follow-up assessment (Table 3). Regarding the DBT-WCCL, measuring the use of DBT skills and maladaptive coping strategies, only the “Skills use” subscale score increased significantly (with a large effect size) more in the DBT condition compared to the WL condition mid-therapy (β₀₁ = 14.98 [6.03–23.96], Pr [β₀₁ > 0] = 0.999) and post-therapy (β₀₂ = 16.56 [6.87–26.22], Pr [β₀₂ > 0] = 1.000), with lasting scores at the follow-up (Table 3). The “General dysfunctional coping” and “Blaming others” subscales did not differ between the two conditions. The FFMQ total mean score, measuring mindfulness skills, increased significantly more in the DBT condition compared to the WL condition only post-therapy (β₀₂ = 17.33 [8.73–25.92], Pr [β₀₂ > 0] = 1.000), and the scores remained stable at the 6-month follow-up assessment (Table 3). Regarding the WHOQoL-BREF, the “Physical health” and “Psychological health” subscales increased more in the DBT condition compared to the WL condition post-therapy (for “Physical health”: β₀₂ = 2.87 [0.83–4.94], Pr [β₀₂ > 0] = 0.997; for “Psychological health”: β₀₂ = 3.13 [0.94–5.32], Pr [β_{0 2} > 0] = 0.997), and remained stable at the 6-month follow-up assessment. No significant differences were found in the “Social relationships” and “Environment” subscales mid- and post-therapy.

Suicidal Behaviours and Hospitalisations during the Therapy

One suicide attempt occurred in each condition (DBT and WL). Both did not require medical intervention. There was no suicide during the study.

Evolution of ED Behavioural Correlates at Follow-Up

Among the 34 therapy completers presenting with NSSI at baseline, 13 (38%) reported no NSSI during the 6-month follow-up period and 13 (38%) reported notable improvements in the frequency of NSSI. Eight (24%) participants did not provide this information.

Among the 54 therapy completers who presented with suicide ideation, 13 (24%) reported no suicide ideation during the follow-up period, 19 (35%) reported notable reduction in the frequency of suicide ideation, and 11 (20%) reported no subjective improvement. Eleven (20%) participants did not provide this information (online suppl. material S3, Additional Table 1). Two participants reported one suicide attempt that occurred during the follow-up period that did not require psychiatric and medical care.

Feasibility and Acceptability Outcomes

The CSQ-8 mean score, measuring satisfaction, was very high, i.e., 3.58 (±0.37, ranging from 2.63 to 4) (Fig. 3). Six (10.34%) participants out of the 58 who started DBT dropped out of the therapy. Only one dropout was related to acceptability challenges, specifically due to anxiety in the group. Another dropout was related to cumulated absences. Three dropouts were related to reasons unrelated to the therapy (i.e., one moved to another town, one due to a physical issue, one due to family problems) and one with no reason given.

Fig. 3.

Mean scores of the CSQ-8 items (out of 4).

The mean attendance rate of the 58 participants who started DBT was 87.82% (±13%, ranging from 29 to 100%). For the 52 participants who completed DBT, the mean attendance rate was 89.38% (±9%, ranging from 67 to 100%).

Mediation and Moderation Effects

The GAFS-8 scores significantly mediated the effect of DBT on the DERS score mid-therapy (ACME = −5.36 [−9.49 to 0.31], p = 0.04) and post-therapy (ACME = −10.02 [−14.93 to −4.52], p < 0.0001). The proportion of the total effect of the DBT on the DERS score mediated by alexithymia (GAFS-8) was, respectively, 22% and 26% mid- and post-therapy. The FFMQ scores significantly mediated the effect of DBT on the DERS score mid-therapy (ACME = −6.23 [−13.68 to −1.53], p < 0.0001) and post-therapy (ACME = −10.49 [−16.16 to −3.67], p < 0.0001). The proportion of the total effect of the DBT on the DERS score mediated by mindfulness (FFMQ) was, respectively, 24% and 27% mid- and post-therapy (online suppl. material S3, Additional Table 2).

Moderation analyses showed that only autistic traits assessed with the AQ moderated the effect of the DBT intervention. Indeed, a higher AQ score was significantly associated with better ED outcomes at follow-up. Specifically, a one-unit increase in the AQ score at T0 was associated with a decrease in the DERS score by 0.786 [−1.310 to −0.257] points at T3 (Pr [β₀₃ > 0] = 0.096).

Discussion

To the best of our knowledge, our study is the first to evaluate the efficacy of DBT in reducing ED in autistic adults presenting with NSSI and/or suicidal behaviours and to investigate the potential mechanisms of change in this context. Our results support previous findings showing the feasibility, acceptability, and efficacy of DBT for autistic adults [15, 16, 45]. Furthermore, they suggest that DBT might be effective in reducing ED involving NSSI and/or suicidal behaviours/ideation in this population.

Interestingly, our findings show that improvement in mindfulness skills and alexithymia might be key mechanisms of change of DBT for ED in autistic adults. This is in line with previous findings pointing to the reduction of alexithymia as a key mechanism of change in DBT in people with BPD [18]. In autistic adults, mindfulness-based treatments other than DBT have been found to be useful for ED [46], which may be explained by the fact that mindfulness skills may enhance the awareness of one’s emotional experience and access to effortful emotion regulation [8, 47]. In other words, mindfulness skills are likely to play a major role in the reduction of alexithymia which is suggested to be key to ED in autistic adults [7, 43, 48], especially in case of co-occurring NSSI and suicidal behaviours [1].

This finding is crucial as several studies support the complex link between alexithymia and autistic traits, with alexithymia being both a difficulty that contributes to and emanates from autistic traits [48, 49]. Notably, even though alexithymia is prevalent in autistic people, concerning up to 60% of them [49], it does not entirely overlap with autistic traits as some autistic people do not present with alexithymia (around 40% of them). Hence, it is plausible to consider that alexithymia is a frequent co-occurrent difficulty that significantly contributes to ED and the diminished well-being of autistic people [7]. Importantly, alexithymia, not autistic traits, has been found to predict social interaction difficulties in autistic people, highlighting its importance as a therapeutic target to mitigate some of the challenges involved in the diminished quality of life of autistic adults [50].

Notwithstanding these improvements in ED and alexithymia, DBT was not effective in improving anxiety and suicide ideation compared to the WL. Similar improvements in the WL condition and the DBT condition were observed, which might be due to the positive impact of anticipating the start of DBT at the end of the waiting period. Indeed, similar findings have been reported in WL control conditions in research with the same design as ours [51], suggesting that WL controls may underestimate the effects of psychotherapy. Therefore, it is possible that differences would have emerged in anxiety and suicide ideation if no therapy had been planned for the WL condition [51]. Additionally, regarding anxious symptoms, our results are in line with the study by Huntjens et al. [16] which found no effects of DBT on social anxiety in autistic adults. Hence, another explanation is that anxiety might be inherently associated with autistic functioning and may be less amenable to change following DBT. At follow-up, however, our participants reported a subjective reduction in suicide ideation and NSSI, which is consistent with the findings by Huntjens et al. [16] showing a decrease in suicide ideation and suicidal behaviours following DBT.

By contrast, our findings suggest that DBT significantly alleviates depressive symptoms in autistic adults. This is in line with several findings in BPD reporting decreased depressive symptoms following DBT, with this improvement being sustained over time [52, 53]. These results are extremely important given the high rates of depression in autistic adults (i.e., 37%) [54] compared to the general population (i.e., 15–21%) [55]. Although depressive symptoms are not the primary target of DBT, decreased depressive symptoms might reflect an enhanced orientation towards goals relative to participants’ “life worth living,” which is the ultimate aim of DBT [12]. Interestingly, this is consistent with the significant improvement in the “psychological health” dimension of quality of life observed post-DBT, which encompasses items on satisfaction with one’s life and finding a meaning to one’s life. Thus, these results indicate that ED might be a relevant target when treating autistic adults with depression similar to other clinical conditions, whereby DBT has been found to be effective in treating depression [16, 56].

Finally, our moderation analyses suggest that autistic traits moderated the ED improvement at follow-up. Specifically, a higher AQ score at baseline was associated with a higher likelihood of improvement at follow-up. This might seem surprising as previous findings have supported a positive correlation between autistic traits and ED [19], suggesting that higher autistic traits might be associated with increased ED severity. Yet, similar results have been found in BPD, whereby higher levels of BPD symptoms were associated with higher improvements following DBT [57], likely due to a regression toward the mean. Interestingly, other factors (i.e., camouflaging, gender, age, borderline traits) likely to be involved in the heterogeneity of ED in ASC [58] did not moderate the efficacy of DBT on ED. Nevertheless, future studies should consider assessing specific variables more thoroughly (e.g., IQ, social functioning, executive functioning) to determine which subgroups of autistic people might benefit most from DBT.

Limitations

Our study has several limitations. First, our control condition is a WL condition, which precludes any conclusion regarding the specific effects of DBT relative to other interventions. Indeed, in future studies, we recommend comparing DBT to an active control condition like a discussion group [14] as well as to other interventions targeting ED in autistic adults, such as the mindfulness program by Conner and White [46]. Second, the self-reported measures used include a risk of bias due to difficulties related to the self-assessment of one’s functioning, especially when alexithymia is present [49]. Nevertheless, numerous studies have found that autistic adults are able to accurately assess their difficulties, including ED and alexithymia [29, 59]. Third, to circumvent the problems related to the classification of NSSI, NSSI was considered an indicator of high ED in our study, with ED being the primary target of DBT [7, 24]. Indeed, given that NSSI included different types of behaviours with different frequencies and severity, we collected data on the perceived effects of DBT on NSSI at follow-up (and not post-DBT) and the frequency of NSSI was not considered a therapy outcome. While NSSI and suicide ideation were assessed daily using diary cards, participants scored these behaviours idiosyncratically, limiting the ability to consider them as group measures (unlike suicide attempts). Moreover, existing standardised assessment tools (e.g., the Non-Suicidal Self-Injury Assessment Tool [60]) are based on retrospective assessments, which are subject to several biases, particularly recall biases [61]. To address these problems, future studies should consider investigating changes in daily measures of NSSI following DBT. Fourth, six (10%) of our autistic participants had a diagnosis of BPD. Hence, ED in these participants could partially stem from co-occurring BPD [7]. However, as these participants were also autistic, DBT seems particularly recommended [62]. It should be noted that differential diagnosis between BPD and ASC is complex, sometimes requiring long-term assessments [62]. Therefore, excluding a BPD co-occurring diagnosis could hinder the feasibility of future trials. Fifth, participants were encouraged to keep their psychotropic medication stable throughout their participation. However, due to the length of participation, some participants in both conditions made minor modifications to their psychotropic medication, and these changes were not tracked to be used as a covariable in the analyses. Nevertheless, previous RCT findings have shown that pre-to-post changes in psychotropic medication were uncorrelated with pre-to-post-DBT changes in the psychotherapy outcomes [63] and medications have shown limited long-term efficacy in reducing ED in several disorders, including ASC [9]. Sixth, the DSM-5 severity level ratings were not considered within our sample of participants, and it is likely that ASC-severity is heterogeneous. Yet, recent studies have found that the DSM-5 specifier is established based on factors other than ASC-severity (e.g., age, IQ), limiting its clinical utility [64]. Hence, the use of other ASC-severity clinical measures, such as the AQ, which was used here as a moderator, seems more pertinent.

The results of this RCT suggest that DBT might be effective in reducing ED in autistic adults presenting with suicidal behaviour and/or self-injury, with the reduction of alexithymia being a potential key mechanism of change. Given the high rates of suicidal behaviours, NSSI and ED in this population, DBT should be further studied in this context and increasingly considered a first-line psychological treatment for autistic adults presenting with ED.

Acknowledgments

The authors thank the Neuroglia Endowment fund, who fully funded the study, and Strasbourg University Hospitals for sponsoring it. The authors thank all participants who took part in the trial. They also thank the general practitioners, psychiatrists, psychotherapists, clinics, associations, and hospitals that supported the recruitment process.

Statement of Ethics

All procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2013. Our study was approved by the Regional Ethics Committee of the East of France (SI 21.01.21.41923). Written informed consent was obtained from each participant.

Conflict of Interest Statement

All authors declare no competing interests.

Funding Sources

Neuroglia endowment fund fully funded the study from October 2021 to September 2024. Strasbourg University Hospital sponsored the study by supplying insurance, legal and statistical support, therapy rooms, and handbook printing. The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The Strasbourg University Hospital provided statistical support to analyse and interpret data.

Author Contributions

L.W., D.B., and S.W. contributed to the conception and the design of the study and acquisition of funding and had final responsibility for the decision to submit for publication. L.W. was responsible for the administration of the project and was responsible for supervising the therapists. D.B. and L.W. were responsible for the study organisation and prepared the original draft. D.B., L.W., S.W., E.R., and R.C. were responsible for the recruitment of participants. D.B. was responsible for data collection and management. E.R. and R.C. were responsible for carrying out ASC diagnostic assessments for participants referred without a formal ASC diagnosis. F.L. was the trial statistician responsible for the formal statistical analysis and was responsible for the randomisation procedure and statistical analysis. D.B., L.W., S.W., and F.L. verified the data, had access to the raw study data, and approved the statistical analysis plan. All authors contributed to the drafting and revision of the final study protocol. All authors confirm that they had full access to all data in the study and confirm responsibility for the decision to submit for publication.

Funding Statement

Neuroglia endowment fund fully funded the study from October 2021 to September 2024. Strasbourg University Hospital sponsored the study by supplying insurance, legal and statistical support, therapy rooms, and handbook printing. The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The Strasbourg University Hospital provided statistical support to analyse and interpret data.

Data Availability Statement

Anonymised datasets used and/or analyses may be available from the corresponding author upon reasonable request. The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants but are available from L.W. upon reasonable request.