eTable. Diagnostic characteristics of the individual instrument in further selected studies with no immediate relevance for German intensive care units*.
| Instrument(s) | Study | Number/age of patients | Design | Diagnostic accuracy[95% CI] | Conclusion | Utility for German intensive care |
| HFRS | Redfern et al. (22) | 31 812/≥ 75 | Two-center retrospective observational study (including intensive care patients) | No association between HFRS and hospital mortality in intensive care patients (p = 0.178). | The HFRS has no additional mortality benefit in intensive care patients. | None |
| FRAIL | Lopez et al. (15) | 132/≥ 65 | Single-center prospective observational study(intensive care patients) | Frailty associated with 1-month mortality (OR: 3.5; [1.22; 10.03]; p <0.05) 6-month mortality (OR: 2.62; [1.04; 6.56]; p <0.05) |
FRAIL is a predictor of mortality in intensive care patients. | Limited, because no very old patients were investigated |
| mFI with 52 criteria | Zeng et al. (e13) | 155/≥ 65 | Single-center prospective observational study(intensive care patients) | Thirty-day mortality with mFI cut-off of > 0.46 and AUC = 0.89 ± 0.03 with FI <0.22 = 0% mortality | The mFI is very complex to measure, but can predict mortality. | Limited |
| mFI | Kizilarslanoglu et al. (e14) | 122/mean age 71 | Single-center prospective observational study(intensive care patients) | In the multivariate analysis the FI—as used by these authors—was an independent predictor of mortality in the intensive care unit (HR: 39.1; [1.2; 1232.5]). | The mFI is very complex to measure, but can predict mortality. | Limited |
| CFS FP |
Ritt et al. (32) | 307/≥ 65 | Single-center prospective observational study(no intensive care patients) | Six-month survival:CFS: AUC 0.867; [0.807; 0.926]; p <0.001 FP: AUC: 0.754; [0.688; 0.821]; p <0.001) Secondary endpoints, such as unplanned hospital admission or falls: CFS superior to FP |
The CFS is superior to FP in some cases. | Limited (no intensive care patients) |
| CFS PRISMA-7 ISAR |
O’Caoihm et al. (35) | 280/≥ 70 | Single-center prospective observational study(no intensive care patients) | CFS had good diagnostic accuracy (AUC 0.83; 95% CI 0.77–0.88). This was lower than PRISMA-7 (AUC: 0.88; [0.83; 0.93], but the difference between the two was not significant (p = 0.15). Youden’s index for the CFS with c ut-off of ≥ 4 at 0.5 (sensitivity 78%, specificity 72%, positive predictive value 80%, negative predictive value 69%, false positive 20%, false negative 31%) Cut-off of ≥ 5 at 0.45 (sensitivity 51%, specificity 94%, positive predictive value 57%, negative predictive value 7%, false positive 7%, false negative 43%) „Gold standard = geriatric consultant’s assessment |
The CFS has good diagnostic accuracy. | Limited (no intensive care patients) |
| FRAIL FI CHS SOF |
Malstrom et al. (e9) | 998 | Single-center prospective observational study(exclusively Afro-American cohort) | Longitudinal cohort study over 9 years On ROC analyses, the FI and the FRAIL scale were superior to the other scores for prediction of disability and mortality |
FRAIL is a valid parameter. | Limited (no intensive care patients, only Afro-Americans) |
| CFS FRAIL |
Chong et al. (36, 37) | 210/ mean age 89.4) | Single-center prospective observational study | Hospital mortality:FRAIL (aOR: 3.31; [1.43; 7.67]; p = 0.005) CFS (aOR: 2.57; [1.14; 5.83]; p = 0.023) Twelve-month mortality: CFS (aOR: 5.78; [3.19; 10.48]; p <.001) |
The CFS is superior for 12-month mortality, FRAIL for hospital mortality. | Limited (no intensive care patients) |
| CFS | Shears et al. (30) | 150/≥ 18 | Single-center prospective observational study(intensive care patients) | Assessments of various operators (researcher, intensivist, geriatric consultant) not significantly different (p > 0.05) | The CFS possesses high inter-rater reliability. | Limited (young patients also included) |
| CFS | Kaeppeli et al. (31) | 2393/≥ 65 | Single-center prospective observational study(emergency patients) | Thirty-day mortality and hospitalization (AUC 0.81; [0.77; 0.85] Good inter-rater reliability (weighted Cohen‘s κ: 0.74; [0.64; 0.85] |
The CFS possesses high inter-rater reliability and is validated for emergency patients. | Limited, because emergency patients were investigated |
| FP | Baldwin et al. (38) | 22/65 | Single-center prospective observational study(intensive care patients) | For every FP point the rate of disability after 1 month rose by 90% (rate ratio: 1.9; [0.7; 4.9]) after 3 months, threefold (rate ratio: 3.0; [1.4; 6.3]) |
FP helps to assess the long-term care need. | Limited owing to the very low number of patients |
*To determine whether the instruments examined in the studies are suitable for use in German intensive care units, the studies were analyzed with regard to the following questions: Were intensive care patients investigated? Were a sufficient number of patients analyzed? Did the validation take place in German/European intensive care units? Was the intended age group investigated? Do the results justify the conclusion that the instrument is suitable?
APACHE II, Acute Physiology and Chronic Health Score II; AUC, area under the curve; aOR, adjusted odds ratio; CFS, Clinical Frailty Scale; CHS, Cardiovascular Health Study Scale;
eFI, electronic frailty index; FRAIL, FRAIL scale; FP, Fried’s phenotype; FI, frailty index; HR, hazard ratio; HFRS, Hospital Frailty Risk Score; mFI, modified frailty index; OR, odds ratio; SAPS II/3, Simplified Acute Physiology Score II/3; SOFA, Sequential Organ Failure Assessment Score; SOF, Study of Osteoporotic Fractures scale; ROC, receiver operating characteristic; 95% CI, 95% confidence interval