180 Comparative performance of a new 3-item health index in predicting mortality: a seven-year follow-up study

Abstract

Abstract Background The effectiveness of health measures in predicting mortality remains a critical aspect of clinical practice. This study aimed to evaluate the performance of a novel 3-item Health Index (HI) and compare it with existing frailty measures, the Frailty Index (FI) and Frailty Phenotype (FP), for predicting all-cause mortality in a cohort of older, community-dwelling adults. Methods Mortality data spanning seven years, encompassing multiple confounding variables, were analysed using Cox proportional regression. The 3-item HI was developed based on resting-state systolic blood pressure sample entropy (sBP SampEn), Sustained Attention Reaction Time (SART) performance, and gait speed. Hazard Ratios (HR) and Area Under the Curve (AUC) were derived for comparative analysis. Results In total, there were 4,265 participants (178 deceased) in the study (mean age = 61.6 ± 8.2 years; 54% female) from The Irish Longitudinal Study on Ageing (TILDA). The new 3-item HI demonstrated superior predictive ability for all-cause mortality compared to the FI and FP, with an AUC of 0.71, surpassing the AUC values of 0.68 and 0.62 for the FI and FP, respectively. Categorising individuals into ‘Low-Risk’, ‘Medium-Risk’, and ‘High-Risk’ groups, the High-Risk group of the HI exhibited a HR of 9.14 (95% CIs: 4.21, 19.86), while the Frail group of the FI and FP displayed HRs of 4.12 (95% CIs: 2.77, 6.12) and 6.87 (95% CIs: 3.58, 13.19), respectively. Furthermore, the 3-item HI remained a significant predictor of mortality even after comprehensively adjusting for confounding variables. Conclusion The objectively measured new 3-item HI demonstrated superior predictive ability for all-cause mortality compared to (mostly self-reported) count-based frailty measures currently employed in clinical practice. To facilitate wider utilisation, we have developed a user-friendly MATLAB App, freely available to researchers and clinicians, allowing easy calculation of the 3-item HI in various healthcare settings. The implementation of this innovative health assessment tool could advance risk prediction and inform decision-making in patient care.