Drawing the borderline: Predicting treatment outcomes in patients with borderline personality disorder.
Borderline personality disorder
Dropout
Effectiveness
Predictors
Premature treatment discontinuation
Treatment outcome
Journal
Behaviour research and therapy
ISSN: 1873-622X
Titre abrégé: Behav Res Ther
Pays: England
ID NLM: 0372477
Informations de publication
Date de publication:
17 Jul 2020
17 Jul 2020
Historique:
received:
29
05
2019
revised:
22
04
2020
accepted:
26
06
2020
pubmed:
18
8
2020
medline:
18
8
2020
entrez:
18
8
2020
Statut:
aheadofprint
Résumé
A routinely collected big data set was analyzed to determine the effectiveness of naturalistic inpatient treatment and to identify predictors of treatment outcome and discontinuation. The sample included 878 patients with borderline personality disorder who received non-manualized dialectic behavioral therapy in a psychosomatic clinic. Effect sizes (Hedge's g) were calculated to determine effectiveness. A bootstrap-enhanced regularized regression with 91 potential predictors was used to identify stable predictors of residualized symptom- and functional change and treatment discontinuation. Results were validated in a holdout sample and repeated cross validation. Effect sizes were small to medium (g = 0.28-0.51). Positive symptom-related outcome was predicted by low affect regulation skills and no previous outpatient psychotherapy. Lower age, absence of work disability, high emotional and physical role limitations and low bodily pain were associated with greater improvement in functional outcome. Higher education and comorbid recurrent depressive disorder were the main predictors of treatment completion. The predictive quality of the models varied, with the best being found for symptom-related outcome (R While the exploratory process of variable selection replicates previous findings, the validation results suggest that tailoring treatment to the individual patient might not be based solely on sociodemographic, clinical and psychological baseline data.
Sections du résumé
BACKGROUND
BACKGROUND
A routinely collected big data set was analyzed to determine the effectiveness of naturalistic inpatient treatment and to identify predictors of treatment outcome and discontinuation.
METHODS
METHODS
The sample included 878 patients with borderline personality disorder who received non-manualized dialectic behavioral therapy in a psychosomatic clinic. Effect sizes (Hedge's g) were calculated to determine effectiveness. A bootstrap-enhanced regularized regression with 91 potential predictors was used to identify stable predictors of residualized symptom- and functional change and treatment discontinuation. Results were validated in a holdout sample and repeated cross validation.
RESULTS
RESULTS
Effect sizes were small to medium (g = 0.28-0.51). Positive symptom-related outcome was predicted by low affect regulation skills and no previous outpatient psychotherapy. Lower age, absence of work disability, high emotional and physical role limitations and low bodily pain were associated with greater improvement in functional outcome. Higher education and comorbid recurrent depressive disorder were the main predictors of treatment completion. The predictive quality of the models varied, with the best being found for symptom-related outcome (R
CONCLUSION
CONCLUSIONS
While the exploratory process of variable selection replicates previous findings, the validation results suggest that tailoring treatment to the individual patient might not be based solely on sociodemographic, clinical and psychological baseline data.
Identifiants
pubmed: 32801095
pii: S0005-7967(20)30146-7
doi: 10.1016/j.brat.2020.103692
pii:
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
103692Informations de copyright
Copyright © 2020 Elsevier Ltd. All rights reserved.
Déclaration de conflit d'intérêts
Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.