Skip to main content

Table 6 Discrimination performance of the three risk scores

From: Early predictors of poor outcome after out-of-hospital cardiac arrest

Risk score

Scores

TTM risk score

>10

>13

>16

  n (%)

612–615 (66–66)

410–417 (44–45)

217–221 (23–24)

 Sensitivity, %

86–87

69–70

40–41

 Specificity, %

57–58

83–84

95–96

 PPV, %

69–70

82–83

91–91

 NPV, %

79–80

71–71

59–59

CAHP risk score

>150

>200

 

  n (%)

711–716 (76–77)

312–318 (33–34)

 

 Sensitivity, %

91–91

48–49

 

 Specificity, %

39–40

82–83

 

 PPV, %

63–63

76–76

 

 NPV, %

79–79

59–59

 

OHCA risk score

>2.0

>17.4

>32.5

  n (%)

829–833 (89–89)

565–574 (62–62)

266–271 (29–29)

 Sensitivity, %

94–94

77–78

44–45

 Specificity, %

16–17

56–58

89–89

 PPV, %

56–56

66–67

82–82

 NPV, %

71–73

69–70

59–59

  1. Abbreviations: CAHP Cardiac Arrest Hospital Prognosis, NPV Negative predictive value, PPV Positive predictive value, ROSC Return of spontaneous circulation, TTM Target Temperature Management trial
  2. Discrimination performance of the three different risk scores in our TTM trial cohort with minimum-maximum values of the first five imputations. The TTM risk score is divided into quartiles where the Youden’s J statistic cut-off (>13 points) coincides with second quartile upper limit (i.e., the median). The CAHP and OHCA risk scores are calculated and divided as described in their respective original publications [11, 12]. They were defined as high risk if >200 points and >32.5 points for the CAHP and OHCA risk scores, respectively