Reviewing methodological approaches to dose-response modelling in complex interventions: insights and perspectives

Area of Science:

• Psychotherapy research
• Clinical trial methodology
• Statistical modeling

Background:

• Dose-response relationships are vital for optimizing clinical outcomes, especially in complex interventions like psychotherapy.
• Research on dose-response in psychotherapy is underdeveloped compared to pharmaceutical contexts.
• This review focuses on statistical methods for modeling dose-response in complex interventions, specifically psychotherapy.

Purpose of the Study:

• To review existing statistical methods for modeling dose-response relationships in complex interventions, with a focus on psychotherapy.
• To identify gaps and limitations in current methodologies for psychotherapy dose-response analysis.
• To understand clinical researchers' perspectives on dosing decisions in psychotherapy trials.

Main Methods:

• Systematic literature search following PRISMA guidelines.
• Identification and categorization of seven eligible studies into multilevel/longitudinal modeling, non-parametric regression, and causal inference with instrumental variables.
• Survey of clinical researchers on psychotherapy dosing decisions.

Main Results:

• Multilevel/longitudinal modeling requires sessional data, limiting causal interpretation.
• Non-parametric regression offers causal inference but relies on assumptions.
• Causal inference with instrumental variables shows promise but needs a priori dose-response function assumptions.
• Researchers lack sufficient information for empirical dosing decisions in psychotherapy.

Conclusions:

• Robust statistical methods for psychotherapy dose-response evaluation are scarce.
• Current dose-response methodology in randomized controlled trials (RCTs) is underdeveloped, limiting causal inference or requiring strong assumptions.
• There is a critical need for advanced methodologies, clearer guidelines, and patient involvement in psychotherapy dosing decisions to optimize patient care.