Validation of the ICEBERG emergency room screening tool for early identification of older patients with geriatric consultation needs

Abstract

BackgroundThe growing number of older and oldest-old patients often present in the emergency room (ER) with undiagnosed geriatric syndromes posing them at high risk for complications in acute care.ObjectiveTo develop and validate an ER screening tool (ICEBERG) to capture 9 geriatric domains of risk in older patients.Design, setting, and participantsFor construct validity we performed a chart-based study in 129 ER patients age 70 years and older admitted to acute geriatric care (pilot 1). For criterion validity we performed a prospective study in 288 ER patients age 70 years and older admitted to acute care (pilot 2).ExposureIn both validation steps, the exposure was ICEBERG test performance below and above the median score (10, range 0–30).Outcome measures and analysisIn pilot 1, we compared the exposure with results of nine tests of the Comprehensive Geriatric Assessment (CGA). In pilot 2, we compared the exposure assessed in the ER to following length of hospital stay (LOS), one-on-one nursing care needs, in-hospital mortality, 30-day re-admission rate, and discharge to a nursing home.Main resultsMean age was 82.9 years (SD 6.7; n = 129) in pilot 1, and 81.5 years (SD 7.0; n = 288) in pilot 2. In pilot 1, scoring ≥10 was associated with significantly worse performance in 8 of 9 of the individual CGA tests. In pilot 2, scoring ≥10 resulted in longer average LOS (median 7 days, IQR 4, 11 vs. 6 days, IQR 3, 8) and higher nursing care needs (median 1,838 min, IQR 901, 4,267 vs. median 1,393 min, IQR 743, 2,390). Scoring ≥10 also increased the odds of one-on-one nursing care 2.9-fold (OR 2.86, 95%CI 1.17–6.98), and the odds of discharge to a nursing home 3.7-fold (OR 3.70, 95%CI 1.74–7.85). Further, scoring ≥10 was associated with higher in-hospital mortality and re-hospitalization rates, however not reaching statistical significance. Average time to complete the ICEBERG tool was 4.3 min (SD 1.3).ConclusionOur validation studies support construct validity of the ICEBERG tool with the CGA, and criterion validity with several clinical indicators in acute care.