Geriatric assessment domains as predictors for clinical endpoints in older adults with cancer: Protocol for an updated systematic review

TL;DR

This study aims to evaluate which geriatric assessment domains predict outcomes in older cancer patients, to help guide treatment decisions when full assessments are not possible.

Contribution

The study updates the understanding of which geriatric assessment domains are most predictive of clinical outcomes in older cancer patients.

Findings

The review will identify which GA domains are most strongly associated with mortality and treatment outcomes.

It will assess variability in GA tools and cutoffs used across studies.

Findings may help optimize geriatric assessments for older cancer patients.

Abstract

Geriatric assessments (GA) are increasingly used to inform treatment decision making and tailoring supportive care for older adults with cancer. Identifying which domains predict clinically relevant outcomes might be particularly useful for risk stratification in settings where a GA is not available and/or feasible. The objective of this updated systematic review is to evaluate individual GA domains as predictors for mortality and treatment-related outcomes. Eligible studies will be identified using a predefined search strategy developed in collaboration with an expert librarian in electronic databases (Medline, Cochrane, Embase, CINAHL) and comprise peer-reviewed papers published in any language from July 2017 and reporting on the prospective association between individual GA domains and mortality as well as surgical- or systemic treatment-related outcomes in older adults with cancer. All title/abstract screening, full-text screening, and data extraction will be performed independently by at least 2 authors. Information on cut-offs of GA domains will also be extracted to assess for variability across studies. A decision on performing a meta-analysis versus a narrative summary will be made based on predetermined criteria, which will include heterogeneity among studies and variability in GA tools and cutoff used for each individual domain, as well as level of risk of bias. If a meta-analysis is indicated, a random effects meta-analysis will be conducted using the Comprehensive Meta-Analysis software. The review will be reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines. This protocol has been registered with PROSPERO (ID: CRD42024580404). This review seeks to investigate individual GA domains as predictors for patient- and treatment-related outcomes. Findings may inform efforts on optimizing GA for this population.