Changes to reclassification across three cardiovascular prediction models, in the validation cohort
From model 1 (original SCORE) to model 2 (recalibrated SCORE) . | ||||||||
---|---|---|---|---|---|---|---|---|
Status after follow-up . | Predicted 10-year risk (original SCORE) . | Predicted 10-year risk (recalibrated SCORE) . | Reclassified . | Net correctly reclassified . | ||||
<5% . | >5% . | Increased risk . | Decreased risk . | |||||
Died from CVD (n = 91) | ||||||||
<5% | 27 | 11 | 12% | 11% | ||||
>5% | 1 | 52 | 1% | |||||
Did not die from CVD (n = 4541) | ||||||||
<5% | 3572 | 204 | 4% | −2.3% | ||||
>5% | 98 | 667 | 2% | |||||
Categorical net reclassification improvement (95% CI) | 0.09 | (0.02 to 0.16) | P = 0.02 | |||||
Continuous net reclassification improvement (95% CI) | −0.09 | (−0.28 to 0.10) | P = 0.34 |
From model 1 (original SCORE) to model 2 (recalibrated SCORE) . | ||||||||
---|---|---|---|---|---|---|---|---|
Status after follow-up . | Predicted 10-year risk (original SCORE) . | Predicted 10-year risk (recalibrated SCORE) . | Reclassified . | Net correctly reclassified . | ||||
<5% . | >5% . | Increased risk . | Decreased risk . | |||||
Died from CVD (n = 91) | ||||||||
<5% | 27 | 11 | 12% | 11% | ||||
>5% | 1 | 52 | 1% | |||||
Did not die from CVD (n = 4541) | ||||||||
<5% | 3572 | 204 | 4% | −2.3% | ||||
>5% | 98 | 667 | 2% | |||||
Categorical net reclassification improvement (95% CI) | 0.09 | (0.02 to 0.16) | P = 0.02 | |||||
Continuous net reclassification improvement (95% CI) | −0.09 | (−0.28 to 0.10) | P = 0.34 |
From model 2 (recalibrated SCORE) to model 3 (HAPIEE SCORE) . | ||||||||
---|---|---|---|---|---|---|---|---|
Status after follow-up . | Predicted 10-year risk (recalibrated SCORE) . | Predicted 10-year risk (HAPIEE SCORE) . | Reclassified . | Net correctly reclassified . | ||||
<5% . | >5% . | Increased risk . | Decreased risk . | |||||
Died from CVD (n = 91) | ||||||||
<5% | 15 | 13 | 14% | 12% | ||||
>5% | 2 | 61 | 2% | |||||
Did not die from CVD (n = 4541) | ||||||||
<5% | 3309 | 361 | 8% | −6.4% | ||||
>5% | 70 | 801 | 2% | |||||
Categorical net reclassification improvement (95% CI) | 0.06 | (−0.02 to 0.14) | P = 0.16 | |||||
Continuous net reclassification improvement (95% CI) | 0.53 | (0.37 to 0.68) | P < 0.0001 |
From model 2 (recalibrated SCORE) to model 3 (HAPIEE SCORE) . | ||||||||
---|---|---|---|---|---|---|---|---|
Status after follow-up . | Predicted 10-year risk (recalibrated SCORE) . | Predicted 10-year risk (HAPIEE SCORE) . | Reclassified . | Net correctly reclassified . | ||||
<5% . | >5% . | Increased risk . | Decreased risk . | |||||
Died from CVD (n = 91) | ||||||||
<5% | 15 | 13 | 14% | 12% | ||||
>5% | 2 | 61 | 2% | |||||
Did not die from CVD (n = 4541) | ||||||||
<5% | 3309 | 361 | 8% | −6.4% | ||||
>5% | 70 | 801 | 2% | |||||
Categorical net reclassification improvement (95% CI) | 0.06 | (−0.02 to 0.14) | P = 0.16 | |||||
Continuous net reclassification improvement (95% CI) | 0.53 | (0.37 to 0.68) | P < 0.0001 |
From model 1 (original SCORE) to model 3 (HAPIEE SCORE) . | ||||||||
---|---|---|---|---|---|---|---|---|
Status after follow-up . | Predicted 10-year risk (original SCORE) . | Predicted 10-year risk (HAPIEE SCORE) . | Reclassified . | Net correctly reclassified . | ||||
<5% . | >5% . | Increased risk . | Decreased risk . | |||||
Died from CVD (n = 91) | ||||||||
<5% | 14 | 24 | 26% | 23% | ||||
>5% | 3 | 50 | 3% | |||||
Did not die from CVD (n = 4541) | ||||||||
<5% | 3296 | 480 | 11% | −8.7% | ||||
>5% | 83 | 682 | 2% | |||||
Categorical net reclassification improvement (95% CI) | 0.14 | (0.04 to 0.25) | P = 0.006 | |||||
Continuous net reclassification improvement (95% CI) | 0.13 | (−0.03 to 0.28) | P = 0.12 |
From model 1 (original SCORE) to model 3 (HAPIEE SCORE) . | ||||||||
---|---|---|---|---|---|---|---|---|
Status after follow-up . | Predicted 10-year risk (original SCORE) . | Predicted 10-year risk (HAPIEE SCORE) . | Reclassified . | Net correctly reclassified . | ||||
<5% . | >5% . | Increased risk . | Decreased risk . | |||||
Died from CVD (n = 91) | ||||||||
<5% | 14 | 24 | 26% | 23% | ||||
>5% | 3 | 50 | 3% | |||||
Did not die from CVD (n = 4541) | ||||||||
<5% | 3296 | 480 | 11% | −8.7% | ||||
>5% | 83 | 682 | 2% | |||||
Categorical net reclassification improvement (95% CI) | 0.14 | (0.04 to 0.25) | P = 0.006 | |||||
Continuous net reclassification improvement (95% CI) | 0.13 | (−0.03 to 0.28) | P = 0.12 |
CI, confidence interval; CVD, cardiovascular disease.
Changes to reclassification across three cardiovascular prediction models, in the validation cohort
From model 1 (original SCORE) to model 2 (recalibrated SCORE) . | ||||||||
---|---|---|---|---|---|---|---|---|
Status after follow-up . | Predicted 10-year risk (original SCORE) . | Predicted 10-year risk (recalibrated SCORE) . | Reclassified . | Net correctly reclassified . | ||||
<5% . | >5% . | Increased risk . | Decreased risk . | |||||
Died from CVD (n = 91) | ||||||||
<5% | 27 | 11 | 12% | 11% | ||||
>5% | 1 | 52 | 1% | |||||
Did not die from CVD (n = 4541) | ||||||||
<5% | 3572 | 204 | 4% | −2.3% | ||||
>5% | 98 | 667 | 2% | |||||
Categorical net reclassification improvement (95% CI) | 0.09 | (0.02 to 0.16) | P = 0.02 | |||||
Continuous net reclassification improvement (95% CI) | −0.09 | (−0.28 to 0.10) | P = 0.34 |
From model 1 (original SCORE) to model 2 (recalibrated SCORE) . | ||||||||
---|---|---|---|---|---|---|---|---|
Status after follow-up . | Predicted 10-year risk (original SCORE) . | Predicted 10-year risk (recalibrated SCORE) . | Reclassified . | Net correctly reclassified . | ||||
<5% . | >5% . | Increased risk . | Decreased risk . | |||||
Died from CVD (n = 91) | ||||||||
<5% | 27 | 11 | 12% | 11% | ||||
>5% | 1 | 52 | 1% | |||||
Did not die from CVD (n = 4541) | ||||||||
<5% | 3572 | 204 | 4% | −2.3% | ||||
>5% | 98 | 667 | 2% | |||||
Categorical net reclassification improvement (95% CI) | 0.09 | (0.02 to 0.16) | P = 0.02 | |||||
Continuous net reclassification improvement (95% CI) | −0.09 | (−0.28 to 0.10) | P = 0.34 |
From model 2 (recalibrated SCORE) to model 3 (HAPIEE SCORE) . | ||||||||
---|---|---|---|---|---|---|---|---|
Status after follow-up . | Predicted 10-year risk (recalibrated SCORE) . | Predicted 10-year risk (HAPIEE SCORE) . | Reclassified . | Net correctly reclassified . | ||||
<5% . | >5% . | Increased risk . | Decreased risk . | |||||
Died from CVD (n = 91) | ||||||||
<5% | 15 | 13 | 14% | 12% | ||||
>5% | 2 | 61 | 2% | |||||
Did not die from CVD (n = 4541) | ||||||||
<5% | 3309 | 361 | 8% | −6.4% | ||||
>5% | 70 | 801 | 2% | |||||
Categorical net reclassification improvement (95% CI) | 0.06 | (−0.02 to 0.14) | P = 0.16 | |||||
Continuous net reclassification improvement (95% CI) | 0.53 | (0.37 to 0.68) | P < 0.0001 |
From model 2 (recalibrated SCORE) to model 3 (HAPIEE SCORE) . | ||||||||
---|---|---|---|---|---|---|---|---|
Status after follow-up . | Predicted 10-year risk (recalibrated SCORE) . | Predicted 10-year risk (HAPIEE SCORE) . | Reclassified . | Net correctly reclassified . | ||||
<5% . | >5% . | Increased risk . | Decreased risk . | |||||
Died from CVD (n = 91) | ||||||||
<5% | 15 | 13 | 14% | 12% | ||||
>5% | 2 | 61 | 2% | |||||
Did not die from CVD (n = 4541) | ||||||||
<5% | 3309 | 361 | 8% | −6.4% | ||||
>5% | 70 | 801 | 2% | |||||
Categorical net reclassification improvement (95% CI) | 0.06 | (−0.02 to 0.14) | P = 0.16 | |||||
Continuous net reclassification improvement (95% CI) | 0.53 | (0.37 to 0.68) | P < 0.0001 |
From model 1 (original SCORE) to model 3 (HAPIEE SCORE) . | ||||||||
---|---|---|---|---|---|---|---|---|
Status after follow-up . | Predicted 10-year risk (original SCORE) . | Predicted 10-year risk (HAPIEE SCORE) . | Reclassified . | Net correctly reclassified . | ||||
<5% . | >5% . | Increased risk . | Decreased risk . | |||||
Died from CVD (n = 91) | ||||||||
<5% | 14 | 24 | 26% | 23% | ||||
>5% | 3 | 50 | 3% | |||||
Did not die from CVD (n = 4541) | ||||||||
<5% | 3296 | 480 | 11% | −8.7% | ||||
>5% | 83 | 682 | 2% | |||||
Categorical net reclassification improvement (95% CI) | 0.14 | (0.04 to 0.25) | P = 0.006 | |||||
Continuous net reclassification improvement (95% CI) | 0.13 | (−0.03 to 0.28) | P = 0.12 |
From model 1 (original SCORE) to model 3 (HAPIEE SCORE) . | ||||||||
---|---|---|---|---|---|---|---|---|
Status after follow-up . | Predicted 10-year risk (original SCORE) . | Predicted 10-year risk (HAPIEE SCORE) . | Reclassified . | Net correctly reclassified . | ||||
<5% . | >5% . | Increased risk . | Decreased risk . | |||||
Died from CVD (n = 91) | ||||||||
<5% | 14 | 24 | 26% | 23% | ||||
>5% | 3 | 50 | 3% | |||||
Did not die from CVD (n = 4541) | ||||||||
<5% | 3296 | 480 | 11% | −8.7% | ||||
>5% | 83 | 682 | 2% | |||||
Categorical net reclassification improvement (95% CI) | 0.14 | (0.04 to 0.25) | P = 0.006 | |||||
Continuous net reclassification improvement (95% CI) | 0.13 | (−0.03 to 0.28) | P = 0.12 |
CI, confidence interval; CVD, cardiovascular disease.
This PDF is available to Subscribers Only
View Article Abstract & Purchase OptionsFor full access to this pdf, sign in to an existing account, or purchase an annual subscription.