Measurement Invariant but Non-Normal Treatment Responses in Guided Internet Psychotherapies for Depressive and Generalized Anxiety Disorders

Direction of dependence analysis for pre-post assessments using non-Gaussian methods: a tutorial

OBJECTIVE

We introduced methods for solving causal direction of dependence between variables observed in pre- and post-psychotherapy assessments, showing how to apply them and investigate their properties via simulations. In addition, we investigated whether changes in depressive symptoms drive changes in social and occupational functioning as suggested by the phase model of psychotherapy or vice versa, or neither.

METHOD

As a Gaussian (normal-distribution) model is unidentifiable here, we used an identifiable linear non-Gaussian structural vector autoregression model, conceptualizing instantaneous effects as during-psychotherapy causation and lagged effects as pre-treatment predictors of change. We tested six alternative estimators in six simulation settings that captured different real-world scenarios, and used real psychotherapy data from 1428 adult patients (Finnish Psychotherapy Quality Registry; assessments on Patient Health Questionnaire-9 and Social and Occupational Functioning Assessment Schedule).

RESULTS

The methodology was successful in identifying causal directions in simulated data. The real-data results provided no evidence for single direction of dependence, suggesting shared or reciprocal causation.

CONCLUSIONS

A powerful new tool was presented to investigate the process of psychotherapy using observational data. Application to patient data suggested that depression symptoms and functioning may reciprocate or reflect third variables instead of one predominantly driving the other during psychotherapy.

Finnish Psychotherapy Quality Register: rationale, development, and baseline results

BACKGROUND

The push to systematically follow treatment outcomes in psychotherapies to improve health care is increasing worldwide. To manage psychotherapeutic services and facilitate tailoring of therapy according to feedback a comprehensive and feasible data system is needed.

AIMS

To describe the Finnish Psychotherapy Quality Register (FPQR), a comprehensive database on availability, quality, and outcomes of psychotherapies.

METHODS

We describe the development of the FPQR and outcome for outsourced psychotherapies for adults in Helsinki and Uusimaa hospital district (HUS). Symptom severity and functioning are measured with validated measures (e.g. CORE-OM, PHQ-9, OASIS, AUDIT, and SOFAS). Questionnaires on therapeutic alliance, risks, methods, and goals are gathered from patients and psychotherapist.

RESULTS

During 2018-2021, the FPQR included baseline data for 7274 unique patients and 336 psychotherapists. Response rate of measures was 85-98%. The use of the register was mandatory for the outsourced therapist of the hospital districts, and the patients were strongly recommended to fulfill the questionnaires. We report outcome for three groups of patients (n = 1844) with final/midterm data. The effect sizes for long psychotherapy (Hedge's g = 0.65 of SOFAS) were smaller than those for short psychotherapy (g = 0.75-0.91). Within three months of referral, 26-60% entered treatment depending on short- or long-term therapy.

CONCLUSION

The FPQR forms a novel rich database with commensurate data on availability and outcomes of outsourced psychotherapies. It may serve as a basis for a national comprehensive follow-up system of psychosocial treatments. The Finnish system seems to refer patients with milder symptoms to more intensive treatments and achieve poorer results compared to the IAPT model in UK, Norway, or Australia.

Text Topics and Treatment Response in Internet-Delivered Cognitive Behavioral Therapy for Generalized Anxiety Disorder: Text Mining Study

Background

Text mining methods such as topic modeling can offer valuable information on how and to whom internet-delivered cognitive behavioral therapies (iCBT) work. Although iCBT treatments provide convenient data for topic modeling, it has rarely been used in this context.

Objective

Our aims were to apply topic modeling to written assignment texts from iCBT for generalized anxiety disorder and explore the resulting topics’ associations with treatment response. As predetermining the number of topics presents a considerable challenge in topic modeling, we also aimed to explore a novel method for topic number selection.

Methods

We defined 2 latent Dirichlet allocation (LDA) topic models using a novel data-driven and a more commonly used interpretability-based topic number selection approaches. We used multilevel models to associate the topics with continuous-valued treatment response, defined as the rate of per-session change in GAD-7 sum scores throughout the treatment.

Results

Our analyses included 1686 patients. We observed 2 topics that were associated with better than average treatment response: “well-being of family, pets, and loved ones” from the data-driven LDA model (B=–0.10 SD/session/∆topic; 95% CI –016 to –0.03) and “children, family issues” from the interpretability-based model (B=–0.18 SD/session/∆topic; 95% CI –0.31 to –0.05). Two topics were associated with worse treatment response: “monitoring of thoughts and worries” from the data-driven model (B=0.06 SD/session/∆topic; 95% CI 0.01 to 0.11) and “internet therapy” from the interpretability-based model (B=0.27 SD/session/∆topic; 95% CI 0.07 to 0.46).

Conclusions

The 2 LDA models were different in terms of their interpretability and broadness of topics but both contained topics that were associated with treatment response in an interpretable manner. Our work demonstrates that topic modeling is well suited for iCBT research and has potential to expose clinically relevant information in vast text data.