EXTERNAL VALIDATION OF SAFES 6-WEEK MORTALITY-RISK INDEX ON AN AFRO-CARIBBEAN OLDER PATIENTS COHORT
Geriatric guidelines recommend considering prognosis in the context of clinical decision making. Several mortality-risk indexes have been developed. While external validation is a recognized crucial step before implementation in clinical practice, few of these indexes have actually been externally v...
Saved in:
| Published in: | Innovation in aging Vol. 1; no. suppl_1; pp. 1099 - 1100 |
|---|---|
| Main Authors: | , , , , |
| Format: | Journal Article |
| Language: | English |
| Published: |
US
Oxford University Press
01-07-2017
|
| Subjects: | |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Geriatric guidelines recommend considering prognosis in the context of clinical decision making. Several mortality-risk indexes have been developed. While external validation is a recognized crucial step before implementation in clinical practice, few of these indexes have actually been externally validated. Our aim was to test the external validity of the SAFEs (
Sujets Ages Fragiles: Evaluation et suivi
) 6-week mortality-risk index developed from a multicentre prospective French cohort on an Afro-Caribbean cohort of older patients. This cohort was collected through a prospective study of 287 patients from the University Hospital of Martinique (French West Indies) from January to June 2012. Patients 75+ hospitalized for an acute condition were eligible. The SAFEs 6-week mortality-risk index of each patient was collected. It included assessments of delirium, risk of malnutrition and functional impairment. Mean age was 86 years. Six-week mortality rate was 19.9%, 52.1% were severely dependant for activities of daily living, 96.1% were at risk of malnutrition, 16.0% had a delirium. The external validity of the SAFEs 6-week mortality-risk index was poor in terms of calibration and discrimination: observed and predicted probability of mortality differed of more than 10% for two out of three risk levels, and the area under the receiver operating curve was 0.58 [0.49–0.66]. Our results corroborate a recent literature review on external validation of risk prediction models showing that external validation is rarely done and often leads to poor results when performed. Clinicians should be aware of these limitations and exercise caution before implementing prognostic indexes in their practice. |
|---|---|
| ISSN: | 2399-5300 2399-5300 |
| DOI: | 10.1093/geroni/igx004.4030 |