Background: Frailty is an indicator of physiological reserve in older people. In non-cancer settings, frailty indices are reliable predictors of adverse health outcomes. The aims of this study were to 1) derive and validate a frailty index (FI) from comprehensive geriatric assessment (CGA) data obtained in the solid tumour chemotherapy setting, and 2) to explore whether the FI-CGA could predict chemotherapy decisions and survival in older cancer patients with solid tumours. Methods: Prospective cohort study of a consecutive series sample of 175 cancer patients aged 65 and older with solid tumours. A frailty index was calculated using an accumulated deficits model, coding items from the comprehensive geriatric assessment tool administered prior to chemotherapy decision-making. The domains of physical and cognitive functioning, nutrition, mood, basic and instrumental activities of daily living, and comorbidities were incorporated as deficits into the model. Results: The FI-CGA had a right-skewed distribution, with median (interquartile range) of 0.27 (0.21-0.39). The 99% limit to deficit accumulation was below the theoretical maximum of 1.0, at 0.75. The FI-CGA was significantly related (p < 0.001) to vulnerability as assessed by the Vulnerable Elders Survey-13 and to medical oncologists' assessments of fitness or vulnerability to treatment. Baseline frailty as determined by the FI-CGA was also associated with treatment decisions (Treatment Terminated, Treatment Completed, No Planned Treatment) (p < 0.001), with the No Planned Treatment group significantly frailer than the other two groups. Conclusion: The FI-CGA is a potentially useful adjunct to cancer clinical decision-making that could predict chemotherapy outcomes in older patients with solid tumours.