# How trustworthy and applicable is the evidence from systematic reviews of depression treatments: Protocol for systematic examination

## TL;DR

This study aims to assess how reliable and useful evidence from depression treatment reviews is, using a detailed method to evaluate their quality and impact.

## Contribution

A novel protocol is introduced to evaluate the trustworthiness and applicability of systematic reviews on depression treatments using multiple methodological tools.

## Key findings

- The protocol outlines a systematic search and evaluation of depression treatment reviews.
- It integrates both qualitative and quantitative methods to assess review quality and bias.
- The study aims to improve evidence-based decision-making for depression treatments.

## Abstract

Depression is a common mental disorder significantly impacting daily functioning. Standard treatments include drugs, psychotherapies, or a combination of both. Treatment selection relies on scientific evidence, though the trustworthiness and applicability of this evidence can vary.

This protocol presents a method to evaluate evidence from systematic reviews for pharmacological and psychological treatments for depression, focusing on trustworthiness and applicability structured into five components: quality of conduct and reporting, risk of bias, spin in abstract conclusions, robustness of meta-analytical results, heterogeneity and clinical diversity.

We will conduct a systematic search of systematic reviews in MEDLINE, Embase, PsycInfo, and Cochrane Database of Systematic Reviews. Our focus will be on systematic reviews of first-line treatments for depression in adults, including antidepressants, psychotherapy, or combined treatments, compared to either active or inactive comparators. We will extract information needed for a comprehensive methodological evaluation using qualitative tools, including AMSTAR 2, ROBIS, Conflict-of-Interest assessment, Referencing Framework for SRs, Spin Measure, and heterogeneity exploration assessment. For quantitative analyses, such as Fragility Index, Ellipse of Insignificance, Region of Attainable Redaction, GRIM test, Leave-N-Out analysis, and prediction intervals, we will select and recalculate two meta-analyses per review. We define a set of outcomes to enable practical and intuitive interpretation of these analyses’ results. Descriptive statistics, non-parametric statistical tests, and narrative summaries will be used to synthesize and compare outcomes across several pre-specified subgroups.

We expect these analyses to provide an enhanced perspective on the practice of evidence synthesis in the field of mental health, offer methodological guidance for future systematic reviews and meta-analyses, and contribute to improved informed decision-making by clinicians and patients.

osf.io/7f9cj and osf.io/ynejs