| Study |
Design |
Participant n, gender,
mean age, diagnosis |
Intervention and Follow-Up |
Relevant Measures of Mechanism and Outcome |
Hypothesized Mechanism(s) of Change and Key Findings |
Main Conclusions |
| Hoglend [53] |
Adjunct to clinical trial |
n=15
12 female
3 male
20-39 years
DPD or APD (n=8)
BPD or HPD or NPD (n=7) |
Dynamic Psychotherapy (up to 53 sessions) with or without transference interpretations |
Dynamic Scales
|
Interpretations
Significant negative main effect of high frequency of interpretation on dynamic change, not symptom change. Interpretation group not rated as improved on any outcome measure. |
Transference interpretations not supported as a mechanism of change. |
| McCullough et al. [55] |
Adjunct to clinical trial, exploratory study |
n=16
13 female
3 male
38 years
OCPD
APD
DPD
PAPD
HPD |
STDP
BAP
(up to 53 sessions) |
Process coding system [100] |
Interpretations
Interpretation of therapist-patient relationship followed by patient affect correlated positively with symptom change. All therapist interventions followed by defensiveness significantly negatively associated with outcome. |
Interpretations may be important in the change process for certain patients who respond to them with affect rather than defensiveness. |
| Schut et al. [52] |
Open trial Exploratory study |
n=14
9 female
5 male
35.9 years
APD
Additional personality disorder (64%) |
Psychodynamic Psychotherapy (52 sessions) |
BAI
BDI
GAF
IIP
SASB
WISPI |
Interpretations
Frequency of interpretation was not associated with outcome. Higher concentration of interpretation significantly associated with higher therapist hostility, lower therapist warmth, and poorer GAF and WISPI scores. Unexpected inverse relationship between affiliativeness and symptom outcome. |
Interpretations can foster therapeutic change where made in a positive relational context and when not too frequent.
Further research required to explore link between interpretation, interpersonal context, and symptom change. |
| Clarkinand Levy [1] |
RCT |
n=90
84 female
6 male
18-50 years
BPD |
TFP
DBT
SPT
(12 months) |
AAI
RFS |
Reflective Functioning
Significant increases in RF, attachment security, and attachment coherence in the TFP group compared to the SPT group. Outcomes were reported in Clarkin, Levy, Lenzenweger (2007), which showed all three interventions were associated with positive outcomes. Of 12 outcome variables, TFP predicted improvement in 10 of them, DBT in 5 of them, and SPT in 6 of them. |
RF and attachment organization are possible mechanisms of change, although did not report relationship between change in RF and change in symptoms. |
| Vermote et al. [59] |
Clinical trial |
n=44
31 female
13 male
28 years
Cluster B (n=32)
Cluster C (n=7)
Cluster B and C (n=5) |
Psychodynamic hospitalization-based treatment
(up to 12 months) |
BDI
IPO
RFS
SCID-II
SCL-90
SHI
STAXI |
Reflective Functioning
Significant improvements in outcomes during treatment and follow-up. No linear increases in RF during treatment or follow up, and RF was not associated with symptom change. |
RF not supported as a mechanism of change, but authors acknowledged that the relationship between RF and change may be more complex than a linear relationship. |
| Neacsiu et al. [61] |
Part of a RCT |
n=108
All female
31.44 years
BPD
Additional drug dependence (n=45) |
DBT (12 months)
4 month follow up
Control conditions – CTBE, Treatment as usual, or Validation therapy |
DBT-WCCL
HRSD
SASII
STAXI |
Skills Use
Significantly higher skills use over time in DBT group compared to control conditions.
Skills use fully mediated decrease in depression and suicide attempts, and increased control over anger. Skills use partially mediated decrease in non-suicidal self-injury. |
Skills use is a mechanism of change for suicidal behaviour (attempts and self-injury), anger control, and depression. |
| Perroud et al. [63] |
Exploratory study |
n=53
47 female
5 male
30.5 years
BPD |
4 week Intensive DBT followed by one year DBT |
BDI-II
BHS
KIMS
SCID-II |
Skills Use
Non-significant decrease in BPD symptoms over time.
Significant increase in ‘accepting without judgment’ mindfulness skill was associated with greater decline in BPD symptoms but this was no longer significant after correction for multiple tests.
Other mindfulness skills not associated with symptom change. |
The discrete skill of ‘accepting without judgment’ within Mindfulness skill-set was the only one to increase significantly over treatment and was positively correlated with symptom change, yet not significantly. Use of different skills as mechanisms of change need further study. |
| Axelrod et al.[64] |
Exploratory study |
n=27
All female
38 years
BPD and Substance Dependence |
DBT
(20 weeks) |
BDI
DERS
Substance use frequency |
Emotion Regulation
Significant reduction in substance use by end of treatment. Significant interaction between reduced substance use and improved ER, but not between substance use and BDI. |
Improvements in ER associated with reduced substance use. Improvements in mood did not account for reduced substance use. |
| Berking et al. [67] |
RCT |
n=81
All females
28.9 years |
DBT
CTBE
(one year) |
AAQ
HRSD
BDI |
Experiential Avoidance
Scores on AAQ, HRSD, and BDI improved by end of treatment. AAQ significantly associated with BDI and HRSD: reduced avoidance associated with improved depression scores. |
Reductions in EA were associated with decreased depression scores. Importantly however, depression scores were not associated with subsequent changes on AAQ. |
| Arntzand Bernstein [71] |
Randomized trial |
n=86
80 female
6 male
30 years
BPD |
SFT
TFP
(three years) |
IPO
BDPSI |
Personality Organisation
Significant symptom improvement in SFT and TFP; superior outcomes in SFT. Changes in IPO throughout treatment correlated strongly with changes in BPDSI scores. |
Changes in IPO were associated with improvements in BPD symptomatology. |
| Brown et al. [76] |
Clinical trial |
n=32
28 female
4 male
29 years
BPD
Additional personality disorder (72%) |
One year of weekly CT
6 month follow up |
BDI-II
BHS
PBQ
SSI |
Cognitive Change
Significant reductions over time on all outcomes.
PBQ scores were significantly reduced by the end of treatment, and between end of treatment and follow-up.
Small associations between changes on PBQ and symptom outcome measures. |
Beliefs changed significantly during CT, yet this was not strongly associated with change in symptoms. |
Note: N: number; RCT: Randomised Controlled Trial