Key result
Six 1-year mortality risk scores demonstrated only moderate prognostic performance in hospitalized older adults with multimorbidity, with none outperforming the others.
Why the study?
Mortality risk scores developed to predict 1-year mortality in multimorbid older adults often lack external validation in large samples, and head-to-head comparisons in prospective contemporary cohorts are missing.
How do 6 mortality risk scores compare in predicting 1-year mortality risk in hospitalized older adults with multimorbidity?
Cohort (n=1,879)
Yes
How do 6 mortality risk scores compare in predicting 1-year mortality risk in hospitalized older adults with multimorbidity?
Six commonly used 1-year mortality risk scores showed only moderate prognostic performance in hospitalized older adults with multimorbidity, suggesting none can be recommended for routine clinical use.
Answers are grounded in this paper and cite their sources.
Prospective head-to-head validation of 6 mortality scores in multimorbid older adults leaves open optimal tool selection; further testing required before routine use.
Importance: The most appropriate therapy for older adults with multimorbidity may depend on life expectancy (ie, mortality risk), and several scores have been developed to predict 1-year mortality risk. However, often, these mortality risk scores have not been externally validated in large sample sizes, and a head-to-head comparison in a prospective contemporary cohort is lacking. Objective: To prospectively compare the performance of 6 scores in predicting the 1-year mortality risk in hospitalized older adults with multimorbidity. Design, Setting, and Participants: This prognostic study analyzed data of participants in the OPERAM (Optimising Therapy to Prevent Avoidable Hospital Admissions in Multimorbid Older People) trial, which was conducted between December 1, 2016, and October 31, 2018, in surgical and nonsurgical departments of 4 university-based hospitals in Louvain, Belgium; Utrecht, the Netherlands; Cork, Republic of Ireland; and Bern, Switzerland. Eligible participants in the OPERAM trial had multimorbidity (≥3 coexisting chronic diseases), were aged 70 years or older, had polypharmacy (≥5 long-term medications), and were admitted to a participating ward. Data were analyzed from April 1 to September 30, 2020. Main Outcomes and Measures: The outcome of interest was any-cause death occurring in the first year of inclusion in the OPERAM trial. Overall performance, discrimination, and calibration of the following 6 scores were assessed: Burden of Illness Score for Elderly Persons, CARING (Cancer, Admissions ≥2, Residence in a nursing home, Intensive care unit admit with multiorgan failure, ≥2 Noncancer hospice guidelines) Criteria, Charlson Comorbidity Index, Gagné Index, Levine Index, and Walter Index. These scores were assessed using the following measures: Brier score (0 indicates perfect overall performance and 0.25 indicates a noninformative model); C-statistic and 95% CI; Hosmer-Lemeshow goodness-of-fit test and calibration plots; and sensitivity, specificity, and positive and negative predictive values. Results: The 1879 patients in the study had a median (IQR) age of 79 (74-84) years and 835 were women (44.4%). The median (IQR) number of chronic diseases was 11 (8-16). Within 1 year, 375 participants (20.0%) died. Brier scores ranged from 0.16 (Gagné Index) to 0.24 (Burden of Illness Score for Elderly Persons). C-statistic values ranged from 0.62 (95% CI, 0.59-0.65) for Charlson Comorbidity Index to 0.69 (95% CI, 0.66-0.72) for the Walter Index. Calibration was good for the Gagné Index and moderate for other mortality risk scores. Conclusions and Relevance: Results of this prognostic study suggest that all 6 of the 1-year mortality risk scores examined had moderate prognostic performance, discriminatory power, and calibration in a large cohort of hospitalized older adults with multimorbidity. Overall, none of these mortality risk scores outperformed the others, and thus none could be recommended for use in daily clinical practice.
No takes yet. Share an insight, caveat, or question.
Schneider et al. (2022) conducted a cohort in Multimorbidity (n=1,879). 6 mortality risk scores was evaluated on any-cause death occurring in the first year (C-statistic 0.62 to 0.69). Six 1-year mortality risk scores demonstrated only moderate prognostic performance in hospitalized older adults with multimorbidity, with none outperforming the others.
Synapse has enriched 10 related papers. Consider them for comparative context: