Comparing Non-Laboratory-Based and Laboratory-Based Cardiovascular Disease Risk Scores
Author Information
Author(s): Ankur Pandya, Milton C. Weinstein, Thomas A. Gaziano, Murielle Bochud
Primary Institution: Harvard School of Public Health
Hypothesis
Can a non-laboratory-based cardiovascular disease risk score effectively assess risk compared to commonly used laboratory-based scores?
Conclusion
The non-laboratory-based score can serve as a useful proxy for laboratory-based scores in assessing cardiovascular disease risk.
Supporting Evidence
- Over 91% of men and 94% of women were equivalently characterized as high or low risk by both non-laboratory-based and laboratory-based scores.
- The non-laboratory-based score had a c-statistic of 0.782 for men and 0.809 for women, indicating good discrimination of CVD risk.
- Statistical analysis showed no significant differences in risk prediction for men, while women showed a significant difference in c-statistics between the non-laboratory-based and Framingham scores.
Takeaway
This study shows that a simple risk score that doesn't need lab tests can help doctors figure out who is at risk for heart problems, just like more complicated tests.
Methodology
The study analyzed individual-level risk scores from the NHANES III dataset, comparing a non-laboratory-based score to four laboratory-based scores using statistical methods including Spearman correlation and ROC curve analysis.
Potential Biases
Potential bias due to the reliance on imputed data for missing values.
Limitations
The study was limited by the use of cause-specific mortality data, which restricted the ability to assess calibration of the risk scores.
Participant Demographics
Participants were adults aged 25-74 years, with a sample of 5,999 individuals without a history of cardiovascular disease or cancer.
Statistical Information
P-Value
0.04
Confidence Interval
(0.739–0.825)
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website