Validity of a self-administered G8 screening test for older patients with cancer

Abstract

IntroductionThe Geriatric 8 (G8) is a brief cancer-specific tool which screens for patients who require a comprehensive geriatric assessment (CGA). The G8 test assesses patients on eight domains such as mobility, polypharmacy, age, and self-rated health. However, the current G8 requires a healthcare professional (nurse or physician) present to conduct the test, which limits its usefulness. The Self-G8 questionnaire (S-G8) is an adaptation of the original G8 test, assessing all the same domains, with questions modified to be appropriate for patients to self-complete. Our objective was to evaluate the performance of S-G8 compared to the G8 and CGA. Materials and MethodsThe initial S-G8 was designed by our team through review of the literature and questionnaire design principles, and was optimized through feedback from patients over the age of 70. The questionnaire subsequently underwent further refinement after undergoing pilot testing (N = 14). The diagnostic accuracy of the final iteration of the S-G8 was evaluated along with the standard G8 in a prospective cohort study (N = 52) in an academic geriatric oncology clinic at the Princess Margaret Cancer Centre, Toronto, Canada. Psychometric characteristics were evaluated including internal consistency, sensitivity, and specificity compared to the G8 and to the CGA. ResultsThere was strong correlation between the G8 and S-G8 scores, with a Spearman correlation co-efficient of 0.76 (p < 0.001). Internal consistency was acceptable at 0.60. The frequency of abnormality (<14 score) for the G8 and S-G8 was 82.7% and 61.5%, respectively. The mean score for the original G8 and S-G8 was 11.9 and 13.5, respectively. The cut-off of 14 for the S-G8 yielded the best combination of sensitivity of 0.70 ± 0.07 and specificity of 0.78 ± 0.14 when compared to the G8. When compared to two or more abnormal domains on the CGA, the S-G8 performed at least as well as the G8 with a sensitivity of 0.77, specificity of 0.85, and a Youden's index of 0.62. DiscussionThe S-G8 questionnaire appears to be an acceptable alternative to the original G8 in identifying older adults with cancer who will benefit from a CGA. Large scale testing is warranted.