Aim: Although in the scientific literature there is a substantial evidence that psychodynamic psychotherapies produce significant symptomatic improvements, the dynamic of change occurring in the process of care remain mainly obscure. Hence, in this work we examine four good-outcome and four poor-outcome psychodynamic psychotherapies pertaining to one of the most renowned research studies: the York Depression Study. The main aim is to investigate the processes of change related to the use of language characterising good and poor outcome psychotherapies.
Method: Every segment of 150 words of brief psychodynamic psychotherapies (i.e. 18-20 sessions) was transcribed and coded. Each segment or statistical unit is clustered according to the K-means Algorithm and/or Minimum Spanning Tree, in order to show the temporal invariants between different segments. Now, each statistical unit will represent a specific configuration of the therapeutic field (i.e. the therapist-patient relationship) characterised by specific patterns of the above mentioned variables. Hence, each cluster will correspond to a specific relational configuration between patient and therapist, a specific state of the complex psychotherapeutic system. The cluster/states time serie generated will be studied by means of sophisticated mathematical analyses in order to identify the invariants within the good and poor outcome therapies processes. These statistical techniques capturing the non-linearity and non-stationarity of time series are, mainly: the Markov Transition Matrix, the Recurrence and the Cross-Recurrence Quantification Analysis, the Network Analyses.
Conclusions: this study for the first time in literature aim to analyse by means of statistical techniques that consider the non-linearity and non-stationarity of data the linguistic longitudinal dynamics of change, trying to understand their evolution characterising four poor and four good outcome brief psychodynamic psychotherapies.
According to the previous literature we can depict four main expected results.
- The network of good outcome cases is ergodic, Halfon, Andreassi, de Felice et al. 2016.
-If the system goes towards phase transitions, these last are preceded by an increase of system¿s variability, Schiepek et al. 2008, 2010, Gumz et al. 2008, and an intra-systemic correlation peak, Gorban et al. 2010.
-The patient-patient and therapist-therapist components are correlated with therapist-patient ones only in good outcome cases, pilot study on George and Lisa.
-The temporal evolution of poor cases will present dysfunctional attractors (at this time there is no such longitudinal study with poor outcome cases, so we deduce it by our previous findings on three good outcome cases aggregated, Halfon, Andreassi, de Felice et al. 2016).
Conclusions
Considering the linguistic aspects of psychotherapy, what are the indexes within the process able to predict good and poor outcome cases?
This study for the first time in literature aim to analyse in detail the linguistic longitudinal variables in order to understand their temporal invariants characterising good and poor outcome cases of brief psychodynamic psychotherapy. Specific attention will be paid to the dynamics of change occurring in their processes, the factors that precede it and those who seem to impede it.