Trial design and procedures
This was a pilot randomized, two-arm (Mindfulness and IE) study. The data presented here was obtained in a pilot randomized clinical trial exploring the effects of MT versus IE on core borderline symptoms and mindfulness-related capacities (Elices, Pascual, Portella, Feliu-Soler, Martin-Blanco, Carmona, Soler: Impact of mindfulness training on borderline personality disorder: A pilot randomized trial, submitted). (Elices et al.Impact of mindfulness training on borderline personality disorder: A pilot randomized trial, Submitted). (. Here, we report secondary data from this trial. The study was approved by the ethics committee of the Hospital de la Santa Creu i Sant Pau and carried out in accordance with the Declaration of Helsinki. Participants were informed of the study procedures and signed informed consent prior to randomization.
Patients were allocated to either mindfulness or IE training. Research Randomizer software was used to obtain randomization sets (www.randomizer.org). Sixteen sets of 4 numbers each were generated. To ensure the same number of subjects in each group, random allocation was forced and thus, each group comprised 8 individuals. The team responsible for enrolment was blind to randomization. To guarantee that participants met inclusion criteria, trained psychiatrists and psychologists blinded to the treatment assignment conducted the clinical interviews. All assessments were collected at the research centre in the presence of a supervising psychologist from our unit.
Participants
Participants were recruited from the outpatient BPD Unit at the Department of Psychiatry from the Hospital de la Santa Creu i Sant Pau from December 2011 to May 2014. A total of 92 participants were screened for eligibility and 64 were randomized, 32 to each treatment group (Fig. 1). Inclusion criteria included fulfillment of BPD diagnostic criteria (two diagnostic interviews: the Structured Clinical Interview for DSM-IV Axis II Disorders [SCID-II] and the Diagnostic Interview for Borderlines Revised [DIB-R]) and aged from 18–45 years (inclusive). Participants were excluded if they: a) had a diagnosis of drug-induced psychosis, organic brain syndrome, bipolar or psychotic disorder or mental retardation; b) were participating in any sort of psychotherapy during the study or had participated in DBT skills groups in the past. Patients who were under pharmacological treatment were included into the study, although changes in type or dose were not allowed.
Treatment conditions
Interventions consisted of group therapy sessions of 120 min each. Participants in both conditions attended group therapy on a weekly-basis for 10 weeks. Both interventions were conducted in agreement with the DBT skills-training manual [28], with the exception that the mindfulness intervention included formal mindfulness practices. Each skill was trained separately during the clinical trial to ensure no overlap between the various DBT skills modules. Each session followed the same structure: 1) review of homework, 2) presentation and practice of a new skill and 3) new homework assignment. In accordance with the DBT framework, skills training in both groups included step-by-step instructions for each skill, rehearsal exercises, and role-playing. To provide feedback and supervision, other team members followed group sessions via a close-circuit television. Video cameras transmitted a signal but did not record.
Mindfulness skills
Two different sets of mindfulness skills (denominated “what” and “how” skills) were taught [28]. “What” skills refer to what is being trained during mindfulness (observe, describe and participate) and “how” skills refer to the attitudinal component of the practice (non-judgmental attitude, focusing on one thing at a time, and being effective). Other mindfulness skills oriented to increasing acceptance of painful life events and emotions were also taught. Formal mindfulness exercises were taught, including: observing the breath, observing sounds, describing thoughts, and physical sensations and walking meditation. Exercises were practiced first during the session, and then at home. All participants received a CD with all formal meditations for home practice. Participants were instructed to decide wisely (i.e., using wise mind) the length of home-practice that was most appropriate for them. The therapist encouraged participants to practice for as long as they could and formal practice was monitored and reinforced in each group session.
Interpersonal effectiveness skills
The aim of these skills is to teach patients how to act more effectively in interpersonal interactions, to achieve their own goals without damaging relationships without losing self-respect. For that purpose, core skills of this module include: objective effectiveness, relationship effectiveness, and self-respect effectiveness [28]. In addition, skills aimed at improving the patient’s ability to ask others to do things or to say no to unwanted requests were also taught.
Instruments
Diagnostic instruments
The SCID-II [29, 30] and DIB-R [2, 31] were used to ensure an accurate BPD diagnosis, as well as to screen for other personality disorders. The Spanish SCID-II is a good instrument to discriminate between Axis II personality disorders with good inter-rater reliability (Kappa of .85). The DIB-R provides BPD diagnosis over the last two years, with scores ranging from 0 to 10 (the cut-off point for BPD diagnosis in the Spanish version is 6). The DIB-R has shown good psychometric properties for internal consistency (Cronbach’s alpha: .89), sensitivity (.81) and specificity (.94). To screen for current Axis I disorders, patients completed the Psychiatric Diagnostic Screening Questionnaire (PDSQ) [32]. The PDSQ is answered in a yes/no format and contains 13 sub-scales screening for several Axis I diagnosis.
Self-reported impulsivity
The BIS-11 was used to assess trait impulsivity [4]. The BIS is a 30-item scale presented on a four-point Likert scale (1 = rarely/never to 4 = always) that measures three aspects of impulsivity: (1) motor impulsiveness (acting without forethought); (2) attentional impulsiveness (the tendency to make quick, non-reflexive decisions); and (3) non-planning impulsiveness (failure to prepare for future events).
Laboratory tasks
Response inhibition
To evaluate response inhibition, subjects underwent the Continuous Performance Test-II (CPT-II) [33] and the GoStop Impulsivity Paradigm [13]. In the CPT-II, participants are instructed to press the “space” bar whenever a letter appears on the computer screen, except when the letter X is display. The entire task includes six blocks, each of which includes 20-trial sub-blocks. The following CPT-II parameters were selected to represent impulsivity indexes [35]: (1) response style (β), in which higher β values indicate a more cautious response style; (2) commissions, which are responses given to non-targets and (3) hit reaction time, which is the average speed of correct responses. A composite impulsivity index was also calculated to provide a general measure of impulsivity [34]. This impulsivity index was calculated as: (1 ⁄ hit reaction time) × (commissions ⁄ omissions). Additionally, CPT-II profiles (i.e., the attention deficit/hyperactivity disorder [ADHD] clinical profile index) were used to obtain a measure of ADHD symptomatology.
The GoStop Impulsivity Paradigm is designed to assess the capacity to inhibit an already initiated response [13]. The task presents three trial types: 1) non-stop (go); 2) stop; and 3) novel trials. In a non-stop trial, a number identical to the previous number is presented in black. A stop trial presents a stimuli that matches the one before it, but it changes from black to red, and a novel trial consists of a different, random number: 5-digit numbers appear on the screen and subjects have to either respond (by pressing a key) if a “go” signal is presented, or withhold the response if a “stop trial” or a “novel trial” is being presented. Responses have to be made while the stimulus is still on the screen. Two primary dependent measures can be derived from this task: 1) failure to inhibit responses: number of responses made on strop trials divided by the total number of stop trials and 2) latency to respond: length of time between the onset of the go stimulus and a response.
Tolerance for delayed reward
Two laboratory tasks assessed tolerance for delayed rewards: the two choice impulsivity paradigm (TCIP) and the single key impulsivity paradigm (SKIP) [13]. In the TCIP, two different shapes appear on the computer monitor, and participants must choose one of them (by clicking with the mouse). Each shape is associated with a different delay-reward contingency; one shape is associated with a shorter delay and a smaller reward (5 points and 5 sec), and the other with a longer delay and a bigger reward (15 points, 15 s). Once the participant has chosen one of each figures, the reward (points) earned appear on the screen. Impulsive participants are characterized by a preference for smaller-sooner rewards instead of larger-later rewards. The number of immediate choices was used as the dependent variable, the higher the number of immediate choices, the higher the impulsivity. Like the TCIP, the SKIP also evaluates tolerance for delayed rewards. The main difference between these two tasks is that the SKIP is a free-operant procedure, meaning that participants are free to make as many responses as they want, taking into consideration that the longer the time between consecutive responses, the bigger the reward. A counter at the bottom of the screen shows the participant the number of points received with each response, thus allowing the subject to infer that faster responses receive fewer points. Another counter (at the top of the screen) shows the total points earned. The total number of responses made during the session are analyzed, with more responses associated with greater impulsivity.
Time perception
The Time Paradigm test [35] was used to assess time estimation. In this task, participants press a key to start a timer and are told to press it again when they think that 1 min has passed. The task includes 5 trials; after each trial, the real time elapsed is shown on the screen, providing subjects with feedback on performance. The average of time estimated was used as the dependent variable.
Statistical analyses
Patient demographic and baseline characteristics were compared using the chi-square test (or Fisher’s exact test if frequencies were < five) for categorical variables and t-test for continuous variables. Normal distributions were tested by visual inspection and a normality test (Kolmogorov-Smirnov, p < .05); for non-normal distributions, the variables were log-transformed. All analyses were performed in the completer population, defined as patients who completed the treatment according to the study protocol. Participants who missed more than 4 consecutive sessions or who abandoned after the first session were considered non-completers.
To explore associations between self-reported and neuropsychological impulsivity measures, Pearson’s correlations were calculated. For this correlation analysis, Bonferroni’s correction was used to control for multiple comparisons.
To evaluate treatment impact on self-report impulsivity, a repeated-measures multivariate analysis of variance (MANOVA) was performed, entering each of the BIS-11 subscale scores as dependent variables and the treatment arm (MT and IE) as the between-subject factor, and time (pre-and post-intervention) as the within-subject factor. To assess the treatment effects on neuropsychological impulsivity another repeated-measures MANOVA was performed as follows: CPT-II (commissions, HitRT, response style and impulsivity index), SKIP (number of immediate choices), TCIP (total number of responses) and TIME (time estimation mean) variables were entered as dependent variables, treatment arm was the between-subject factor, and time (pre-post intervention) was the within-subject factor. If significant main effects were observed, post-hoc group comparisons were computed using the t-test. Secondarily, to explore whether ADHD symptoms might have influenced the results, the CPT-II clinical confidence index was included as a covariate for both repeated-measure MANOVAs. This index provides an estimate of the probability that any given CPT II result resembles that of a clinical profile. All analyses were conducted with SPSS for Windows, Version 19.