Randomized controlled trial on the effectiveness of psychometric feedback on treatment success (DFG-funded)
A large number of studies have been able to show the positive effect of continuous psychometric surveys and feedback of the processed information on the progress of therapy. This is especially true for high-risk patients who are at risk of experiencing no change or a negative change. However, in a recent meta-analysis, the findings are mixed, so further research is needed (Kendrick et al., 2016). Since early 2017, a DFG-funded (LU 660/9-1) randomized controlled trial evaluating the effects of personalized treatment prediction and personalized treatment adjustment on treatment outcomes has been ongoing. In this project, therapies are compared in which the therapist was provided with (1) no feedback, (2) feedback regarding patient self-assessment, and (3) feedback regarding patient self-assessment and additional clinical support tools in the computer-based feedback portal. The aim is to examine to what extent psychometric feedback that therapists receive about their patients' therapy progress has a positive impact on therapy outcome and leads to personalization (Lutz, Zimmermann, Müller, Deisenhofer, & Rubel, 2017). In practice, therapists personalize their treatment offering by adapting therapies to the patient's needs intuitively and based on their clinical experience. However, this personalization is unsystematic and raises the question of whether this personalization could be supported by an evidence-based approach and result in improved treatment, especially for high-risk patients. Accordingly, the results of the study should provide new insights in the field of patient-centered care research and promote further development of the profession as well as the discipline.
A freely available version of the Trier Therapy Navigator can be found here: https://github.com/Psykli/Trier_Treatment_Navigator
The research group (Lutz) participated in the SMART Tournament (Stratified Medicine Approaches foR Treatment Selection (SMART) - Mental Health Prediction Tournament), a scientific competition to determine the best method for predicting treatment outcomes and the optimal allocation of patients to different treatment conditions (funded by UK Wellcome Trust). In addition, the project aimed to evaluate different methodological approaches for advantages and disadvantages and thus contribute to fundamental knowledge enhancement in the field of psychotherapy research.
A total of 13 different research groups from different institutions (e.g. University of Pennsylvania, Queens College, Yale University, University of California, Leiden University, University of Trier, etc.) competed. Each research group received a data set consisting of 3840 cases routinely collected by the Improving Access to Psychological Therapies (IAPT) program. Using these data, appropriate predictive models were to be developed. In particular, machine learning algorithms were used to create a predictive model based on certain mathematical statistical procedures. These were tested by independent scientists by applying the models to a so-called "hold-out" data set. The results showed that different models lead to similarly good results, with the models of the Trier research group being among the best across different comparisons (rank 2). The SMART Tournament results were presented at the Treatment Selection Idea Lab conference in London in June 2018. The research group participated with its own contributions and also won a poster prize (Dr. Deisenhofer). The next meeting of the "Treatment Selection Idea Lab" will take place at the University of Trier in 2021.
Many studies have shown that online therapy can be effective as a treatment option for various disorders. So far, there are only a few studies in Germany that examine the extent to which online therapies can be effectively implemented in routine care. In this project, patients waiting for a therapy place at the psychotherapy outpatient clinic PALF are given access to an online portal. This includes useful exercises (e.g., on resource and activity building) from behavioral therapy manuals. So far, 60 patients have already used the exercises to bridge the waiting time. Evaluation of usage data, satisfaction, and symptom changes over time of use is expected in January 2019. The goal is to gain initial insights into whether, and if so in what form, implementing online therapy as a bridging service for waiting time in routine care is useful.
Ecological Momentary Assessment (EMA) is a form of outpatient data collection that takes place in real time in the individuals' natural environment, usually using electronic data collection devices. This new method of measurement allows novel questions to be investigated in the context of psychotherapy research. In the EMA study of the psychotherapy outpatient clinic PALF, diagnostic information regarding the everyday stresses and resources of patients is already obtained while they are waiting for an outpatient therapy place. For this purpose, patients are presented with questions on their own Android smartphones several times a day regarding their current state of health and their experiences during the last 3-4 hours. Well-being, positive and negative affect, experience and behavior as well as emotion regulation skills are asked. The aim of the study is to use the data from the waiting period to develop predictive models that describe the development in later psychotherapy and from which interventions can be derived and an individualized treatment concept can be created.
For the evaluation and quality assurance of the routine care of outpatient psychotherapy, the system in the Trier Research Outpatient Clinic is continuously developed further, financed by various health insurance companies (including Techniker Krankenkasse).
Similar to navigation systems in everyday life, the "Trier Treatment Navigator" enables computer-assisted help with concrete therapeutic decisions (at the beginning and in the course of therapy) based on the findings of patient-oriented psychotherapy research on differential and adaptive indication (cf. Lutz, Clausen & Deisenhofer, 2019).
It thus represents a further development of the PALF|Feedback system and is an internet-based system for quality assurance of psychotherapeutic treatments. It enables continuous monitoring of the psychotherapeutic process. The data input takes place directly in the therapy room at available touch screen PCs. The evaluation is automated and can be viewed by the therapists in an online portal. A detailed description of the components of the system as well as concrete case examples from patient treatments can be found in Lutz, Neu & Rubel (2019).
The system is routinely used in the Polyclinic Psychotherapy Outpatient Clinic at the University of Trier and is being investigated in a recent controlled clinical trial (Lutz, Zimmermann, Müller, Deisenhofer & Rubel, 2017).
In the context of the use of this system in the psychotherapy outpatient clinic, various questions regarding routine care are being investigated in the context of dissertations and master's theses (including therapist effects, services for persons on the waiting list, training research).
In a pilot project of the Techniker Krankenkasse (TK), in which the working group was part of the evaluation team, the influence of routine and continuous progress feedback on the therapeutic progress (feedback) in an outpatient psychotherapy sample was investigated for the first time in Germany in the sense of patient-oriented health care research. Between May 2005 and June 2010, data were collected from a total of 1708 patients who were treated by one of 245 psychotherapists in private practice with different therapeutic orientations (behavioral therapy, depth psychology-based psychotherapy, analytic psychotherapy) participating in the study. In the intervention group (IG), compared to a control group (KG), the therapists received continuous feedback on the progress of therapy in the form of change graphs and characteristic values during the course of therapy. Central research questions of the model project included testing the applicability of quality assurance and feedback systems in outpatient psychotherapies under routine conditions as well as their influence on outcome quality (e.g., Wittmann, Lutz et al., 2011; Lutz et al., 2011; 2012; 2013).
The inpatient feedback project takes place in cooperation with the psychosomatic specialist clinic of the MediClin Bliestal Kliniken in Blieskastel. Within the framework of the project, the feedback effect described above is to be tested for the first time for inpatient routine care. The question is operationalized by means of the survey of four different groups, three control groups and one intervention group each, whose therapy success is to be recorded by means of different questionnaires.
Patients of the psychotherapy outpatient clinic PALF are interviewed again about their psychological condition at regular intervals after the end of therapy, up to three times within 3 years. This enables the working group to investigate to what extent the short-term therapy effects found remain stable and sustainable in the long term. The results of studies in the literature so far are promising. Especially in comparison to drug treatments, the effects of psychotherapy have been shown to be more stable and sustainable (Wucherpfennig, Boyle, Rubel, Weinmann-Lutz, & Lutz, in preparation).
Continuous monitoring of changes in psychotherapeutic interventions is of particular importance in the sense of a personalized treatment approach and scientifically supported psychotherapy. For the regular measurement of therapeutic progress, measuring instruments are needed that fulfill various preconditions. For example, they should be short and inexpensive due to the frequency of data collection. At the same time, they should cover a broad spectrum of relevant aspects and be as sensitive to change as possible.
The working group is continuously engaged in the development of change-sensitive curative forms of common psychometric instruments (e.g. Lutz et al., 2006; Böhnke & Lutz, 2010) as well as new instruments for the assessment of therapy progress (e.g. Lutz & Böhnke, 2008; Lutz et al., 2009, Schürch, Lutz, & Böhnke, 2009).
Within patient-focused psychotherapy research, questionnaires are widely used to predict therapy success and feedback on process variables (e.g., on the quality of the therapeutic relationship). Research shows that nonverbal indicators can provide valuable additional information about dyadic features of the therapeutic relationship. They offer the advantage that they are subject to objective criteria through automated video analyses and are thus not influenced by possible subjective biases. In a DFG-funded project of the research group (DFG, grant nos. STR 306/28-1 to Bernhard Strauß, LU660/8-1 and LU660/10-1 to Wolfgang Lutz) the predictive value of nonverbal synchrony of movements for therapy success was investigated. Results show that nonverbal synchrony of movements is related to therapy success and drop-out (Paulick et al., 2017). Furthermore, in depressed patients, for example, an increase in dyadic synchrony can be detected during the course of therapy, providing information beyond the quantity of movements (e.g., general slowing down) (Paulick et al., 2018). The current research project focuses on synchrony or co-regulation of emotions. Using "Social Signal Interpretation" software, information from facial expressions and prosody will be integrated to enable automated recognition of emotion expression. This methodology is to be tested in the therapeutic setting and examined to determine whether it can contribute to the prediction of therapeutic success.
In psychotherapy research, treatment integrity refers to the extent to which a therapy is implemented as intended and includes, on the one hand, the selection of interventions based on a treatment protocol (adherence) and, on the other hand, the professional application of techniques in the interpersonal context of therapy (competence). A particular challenge arises in recording treatment integrity under routine conditions compared with randomized-controlled trials, since here the treatment is not kept as uniform as possible but is sensibly adapted to the individual patient and course. Nevertheless, the question of treatment integrity is of high relevance in this context in order to be able to investigate therapy processes as well as to assure quality. The aim of the current project is therefore to develop and test an inventory that can be used to evaluate therapy videos with regard to treatment integrity. After the reliability of the inventory has been sufficiently proven, the results are to be used to investigate therapy processes and treatment outcomes as well as to optimize supervision in the long term.
This research group is concerned with the investigation of individual discontinuous courses of therapy and their relationship to therapy outcome. It has been shown that a portion of individual patient trajectories do not improve in a linear and continuous fashion, but rather changes occur discontinuously. These discontinuous changes, i.e. large symptom improvements or worsening between two consecutive sessions, are referred to in the literature as Sudden Gains or Sudden Losses. One of the areas of interest of this research group is the possible causes of positive and negative jumps in change during the therapy process and their relationship to therapy success (Lutz, et al., 2013; 2007; Lutz & Tschitsaz, 2007; Tschitsaz-Stucki & Lutz, 2009). For example, therapy sessions prior to a Sudden Gain or a Sudden Loss have been shown to differ qualitatively from one another and to be influenced by cognitive changes, interpersonal aspects of the therapists, the therapeutic relationship, the emotional status of the patients, and extra-therapeutic factors.
Language as a means of social communication and interaction serves to express personal experience and thus reflects, for example, feelings, thoughts and psychological state. At the same time, our linguistic behavior is influenced by who we are, what makes us tick, and how we feel. Thus, it can be assumed that language both intentionally, through what we say, and non-intentionally, through how we say something, reflects how we are and what is on our mind.
Psychotherapy is largely based on language. Thus, language is a source of data that is present in every psychotherapy session and, because it can be evaluated retrospectively, may be less susceptible to response bias than, for example, questionnaires.
The psychotherapy outpatient clinic PALF is characterized by an area-wide video recording of psychotherapy sessions and a wide range of diagnostics from diagnostic interviews to questionnaires and checklists, so that there are best possibilities for the analysis of language data.
In the project, anonymized transcripts of therapy sessions were created and evaluated using software (LInguistic Inquiry and Word Count - LIWC), which evaluates the frequency of use of word categories such as pronouns, positive emotion words, negative emotion words, cognitive words, social words, and the like. Then, it was compared whether patients with depression, anxiety disorders, or depression and anxiety disorders differed, and it was found that the emotional content differed. For example, depressed patients used more words related to sadness, while anxiety patients used more words related to fear. The group of patients with depression and anxiety disorder can be understood from the language use as a combination of the two groups.
A substantial part of human communication occurs via nonverbal signals. This results in a potentially large relevance for therapeutic interactions and processes. In particular, emotions and their expression are central to most mental disorders and their treatment. New methods of digital video processing for automated recognition of emotion expression will be applied in outpatient psychotherapy in this project. To this end, in cooperation with Prof. André and colleagues at the University of Augsburg, the software NOVA (Nonverbal Behavior Analyzer) is currently being applied to therapy videos and evaluated as part of a pilot study (Clausen, Baur, Lingenfelser, André & Lutz, 2019).
Through this methodology, basic insights into therapy processes are to be expanded and correlations with therapy courses and success are to be investigated. In addition to aspects of the respective emotion expression (valence and arousal), the emotional co-regulation between therapist and patient is also of interest. Based on theory and previous empirical studies, it is expected that more intense emotional activation and moderate to high co-regulation are related to greater therapeutic success.
Clausen, S. A., Baur, T., Lingenfelser, F., André, E. & Lutz, W. (2019, May). Automated emotion recognition in psychotherapy - Pilot study on the applicability of the Nonverbal Behavior Analyzer (NOVA). 11th Workshop Congress on Clinical Psychology and Psychotherapy, Erlangen, Germany.
In cooperation with Bar Ilan University in Tel Aviv, a specially developed ImRs treatment concept was tested in the treatment of test anxiety on a student sample in the summer semester of 2016. The six treatment sessions of 50 minutes each took place over a period of three weeks, twice a week with an interval of two working days. The sessions were video recorded, and audio was also collected separately using directional microphones, as well as electrodermal activity (EDA). Data collection covered a 14-week period and was oriented to the semester schedule. Subjects were divided into three groups. The groups differed only in treatment onset. Group 1 began treatment sessions two weeks after the baseline measurement, Group 2 four weeks, and Group 3 six weeks thereafter. Depending on the grouping, the subject completed a questionnaire once a week for a period of three to seven weeks before treatment began and a questionnaire once a week for a period of three weeks after treatment was completed.
During the period between early May 2015 and late January 2016, a group therapy program for the treatment of chronic depression was offered at the PALF outpatient psychotherapy clinic. The group therapies were conducted as part of the study "Investigating the Effectiveness of Mindfulness-Based Loving-Kindness Meditation in Patients with Chronic Depression" (Schilling et al., 2018). The program included various meditation practices and mindfulness-based techniques to reduce rumination and depressive mood. In addition, participants received suggestions for emotion regulation, managing worry and rumination, and learned a more benevolent, less critical approach to themselves. Compared with the start of treatment, significant improvements were achieved in terms of the patients' general distress. However, there was initially no significant improvement at the endpoint of the therapy with regard to depressive symptoms. A continuation of the program is currently not planned. Effective elements of the program (psychoeducational concepts, mindfulness-based exercises, meditation instructions, etc.) were integrated into the feedback portal used by the therapists at the PALF psychotherapy outpatient clinic. The materials can be used by the therapists of the Psychotherapieambulanz PALF for individual therapy.
On a micro or basic level at the intersection of psychotherapy research and neuroscience, the background and specific conditions of change processes in psychotherapy are analyzed. In this research focus, the working group is concerned with the effectiveness and, in particular, the neural correlates of individual specific psychotherapeutic intervention methods (e.g., cognitive restructuring, relaxation procedures, etc.). The question arises which therapeutic techniques are effective under which conditions and which are the basic mechanisms of action. The aim is to develop concepts and models for a specific scientific support of differential indication decisions. Several third-party funded studies (TransCoop program of the Alexander von Humboldt Foundation, Research Professorship of the Swiss National Science Foundation, Research Fund of the University of Trier, Research Initiative Rhineland-Palatinate) provide first promising results showing the influence of therapeutic microinterventions on brain physiological activity (e.g., Zaunmüller & Lutz, 2012; Zaunmueller, Lutz, & Strauman, 2012; Strauman, et al., 2013).