Defining an Abnormal Geriatric Assessment: Which Deficits Matter Most?

Area of Science:

• Geriatric Oncology
• Clinical Assessment
• Treatment Planning

Background:

• Geriatric assessment (GA) is vital in geriatric oncology, but a clear definition of 'abnormal' is lacking.
• The association between abnormal GA domains and treatment plan modification (TPM) requires further investigation.
• Identifying key GA domains predictive of TPM is essential for optimizing cancer care in older adults.

Purpose of the Study:

• To define what constitutes an abnormal geriatric assessment in geriatric oncology.
• To determine the association between specific geriatric assessment domains and treatment plan modification.
• To identify the optimal numerical threshold of abnormal domains for predicting treatment plan modification.

Main Methods:

• Retrospective review of geriatric oncology clinic database (May 2015 - June 2022).
• Logistic regression modelling to assess the association between GA domains/thresholds and TPM.
• Analysis of 736 new patients undergoing treatment decision-making.

Main Results:

• The best univariable threshold for abnormal GA domains predicting TPM was 4 (AUC 0.628).
• A multivariable model including cognition, comorbidities, and falls risk showed the highest predictive accuracy (AUC 0.704).
• A threshold of 4 abnormal domains in a multivariable model yielded an AUC of 0.689.

Conclusions:

• An abnormal geriatric assessment in oncology may be best defined by abnormalities in cognition, comorbidities, and falls risk.
• The optimal numerical threshold for predicting treatment plan modification is 4 abnormal domains.
• These findings can refine the definition of abnormal GA and improve treatment decisions for older cancer patients.