An update on a systematic review of the use of geriatric assessment for older adults in oncology.

BACKGROUND: Our previous systematic review of geriatric assessment (GA) in oncology included a literature search up to November 2010. However, the quickly evolving field warranted an update. Aims of this review: (i) provide an overview of all GA instruments developed and/or in use in the oncology setting; (ii) evaluate effectiveness of GA in predicting/modifying outcomes (e.g. treatment decision impact, treatment toxicity, mortality, use of care).

MATERIALS AND METHODS: Systematic review of literature published between November 2010 and 10 August 2012. English, Dutch, French and German-language articles reporting cross-sectional or longitudinal, intervention or observational studies of GA instruments were included.

DATA SOURCES: MEDLINE, EMBASE, PsycINFO, CINAHL and Cochrane Library. Two researchers independently reviewed abstracts, abstracted data and assessed the quality using standardized forms. A meta-analysis method of combining proportions was used for the outcome impact of GA on treatment modification with studies included in this update combined with those included in our previous systematic review on the use of GA.

RESULTS: Thirty-five manuscripts reporting 34 studies were identified. Quality of most studies was moderate to good. Eighteen studies were prospective, 11 cross-sectional and 5 retrospective. Three studies examined treatment decision-making impact and found decisions changed for fewer than half of assessed patients (weighted percent modification is 23.2% with 95% confidence interval (20.3% to 26.1%). Seven studies reported conflicting findings regarding predictive ability of GA for treatment toxicity/complications. Eleven studies examined GA predictions of mortality, and reported that instrumental activities of daily living, poor performance status and more numerous GA deficits were associated with increased mortality risk. Other outcomes could not be meta-analyzed.

CONCLUSION: Consistent with our previous review, several domains of GA are associated with adverse outcomes. However, further research examining effectiveness of GA on treatment decisions and oncologic outcomes is needed.