Table 4.
GRADE evidence profile/summary of findings table: neuroprognostication: msTBI
| Disease | Outcome | Predictor or model | Quality of evidence | Summary of findings (narrative of effect size) | ||||
|---|---|---|---|---|---|---|---|---|
| RoB | Inconsistency | Indirectness | Imprecision | QoE: summary (high/moderate/low/very low) | ||||
| Models | ||||||||
| TBI | Mortality | CRASH-basic | ↓ ROB: applicability [some mild TBI], self-fulfilling prophecy | Moderate | 14-day mortality: AUC 0.8–0.82 (95% CI 0.77–0.75, 0.85–0.87) good calibration; caveat, original model includes mild TBI patients | |||
| TBI | Mortality | CRASH CT | ↓ ROB: applicability [some mild TBI], self-fulfilling prophecy | Moderate | 14-day mortality: AUC 0.83 (0.78–0.87); good calibration; caveat, original model includes mild TBI patients | |||
| TBI | Mortality | IMPACT Core | ↓ of 7 papers high ROB due to small sample sizes, poor report on methods, variable outcome time points, self-fulfilling prophecy; remaining papers high quality | Moderate | 6-month mortality: AUC 0.66–0.9 (95% CI 0.8–0.86, 0.87–0.94), most papers report calibration | |||
| TBI | Mortality | IMPACT extended | ↓ of 7 papers high ROB due to small sample sizes, poor report on methods, variable outcome time points, self-fulfilling prophecy; remaining papers high quality | Moderate | 6-month mortality: AUC 0.71–0.85 (95% CI 0.76–0.81, 0.86–0.89) | |||
| TBI | Mortality | IMPACT lab model | ↓ two larger paper, two smaller papers, two with high ROB for participation, predictors, outcome, self-fulfilling prophecy) | Moderate | 6-month mortality: AUC 0.72–0.80 (95% CI 0.75–0.86) | |||
| TBI | Mortality | Marshall CT | ↓ high ROB (applicability, lack of calibration, outcome ROB, self-fulfilling prophecy) | Moderate |
6-month mortality: AUC 0.635- 0.819 (95% CI 0.769–0.869); calibration reported for one study 12-month mortality: AUC 0.61 (95% CI 0.57–0.65), calibration not reported |
|||
| TBI | Mortality | Rotterdam CT | ↓ 4 papers, three with high ROB for applicability, sample size, analysis, 1/3 mild TBI, self-fulfilling prophecy) | ↓ | Low | 6-month mortality: AUC 0.699–0.875 (95% CI not reported-0.814, not reported-0.936) | ||
| TBI | Mortality | Helsinki CT | ↓ both papers with high ROB for applicability, analysis, outcome, 1/3 mild TBI, self-fulfilling prophecy lack of calibration) | Moderate |
6-month mortality: AUC 0.744 (95% CI not reported); calibration not reported 12- month mortality: AUC 0.74 (95% CI 0.7–0.77); calibration not reported |
|||
| TBI | Mortality | Stockholm CT | ↓ single paper, high ROB (applicability, lack of calibration, outcome, self-fulfilling prophecy ROB) | ↓ | Low | 12- month mortality: AUC 0.77 (95% CI 0.73–0.8); calibration not reported | ||
| TBI | Unfavorable functional outcome | CRASH basic | ↓ ROB (two of four studies with ROB for participation, applicability, self-fulfilling prophecy calibration not reported) | Moderate | 6-month AUC 0.8–0.86 (95% CI 0.75–0.88, 0.84–0.9) | |||
| TBI | Unfavorable functional outcome | CRASH CT | ↓ ROB (two studies, one with ROB for participation, applicability, self-fulfilling prophecy calibration not reported) | Moderate | 6-month AUC 0.86–0.89 95% CI (0.82–0.84, 0.89–0.93) | |||
| TBI | Unfavorable functional outcome | IMPACT Core | ↓ ROB (1 of 8 studies with ROB for participation, outcome, self-fulfilling prophecy, 2 for analysis, 2 without calibration, 1 with variable outcome time point assessment) | Moderate | 6-month AUC 0.71–0.87, 95% CI (0.76–0.81, 0.84–0.91) | |||
| TBI | Unfavorable functional outcome | IMPACT extended | ↓ ROB (1 of 6 studies with ROB for analysis, self-fulfilling prophecy, no calibration, 1 with variable outcome time point assessment) | Moderate | 6-month AUC 0.73–0.88 (95% CI 0.77–0.83, 0.86–0.93) | |||
| TBI | Unfavorable functional outcome | IMPACT lab model | ↓ ROB (self-fulfilling prophecy, 1 with variable outcome time point assessment, 1 of 4 studies with not enough information to judge ROB) | Moderate | 6-month AUC 0.75–0.87 (95% CI 0.82–0.92) | |||
| TBI | Unfavorable functional outcome | Marshall CT | ↓ ROB (participation [1/3–40% mild TBI], applicability, analysis, self-fulfilling prophecy) | ↓ | Low |
6-month AUC 0.635–0.87 (95% CI 0.82–0.92) 12-month unadjusted AUC 0.54 (95% CI 0.45 – 0.63), calibration not reported |
||
| TBI | Unfavorable functional outcome | Rotterdam CT | ↓ ROB (participation [1/3–40% mild TBI], applicability, analysis, self-fulfilling prophecy) | ↓ | Low |
6-month AUC 0.682 12-month unadjusted AUC 0.61 (0.53–0.7), calibration not reported |
||
| TBI | Unfavorable functional outcome | Helsinki CT | ↓ ROB (participation [1/3–40% mild TBI], applicability, analysis, self-fulfilling prophecy) | ↓ | Low |
6-month AUC 0.72–0.75 (95% CI 0.69–0.75) 12-month unadjusted AUC 0.63 (0.54–0.71), calibration not reported |
||
| TBI | Unfavorable functional outcome | Stockholm CT | ↓ ROB (applicability, lack of calibration, outcome, self-fulfilling prophecy ROB) | ↓ | Low | 12-month AUC 0.7–77 (95% CI 0.61–0.79); calibration not reported | ||