Related Experiment Video
Updated: Feb 13, 2026
In Silico Clinical Trials for Cardiovascular Disease
Published on: May 27, 2022
Development of algorithms for identifying fatal cardiovascular disease in Medicare claims
Fenglong Xie1,2, Lisandro D Colantonio1, Jeffrey R Curtis1,2
1Department of Epidemiology, School of Public Health, University of Alabama at Birmingham, Birmingham, AL, USA.
Insights
New algorithms accurately identify deaths from cardiovascular disease (CVD) using Medicare claims data. These claims-based tools improve upon traditional methods for cause of death determination in research.
Area of Science:
- Biostatistics
- Epidemiology
- Health Informatics
Background:
- Administrative claims data often lack precise cause of death information.
- Accurate identification of mortality is crucial for epidemiological research and public health surveillance.
Purpose of the Study:
- To develop and validate algorithms using Medicare claims data to identify deaths due to fatal cardiovascular disease (CVD), including coronary heart disease (CHD) and stroke.
- To compare the performance of these claims-based algorithms against existing methods.
Main Methods:
- Linked Reasons for Geographic and Racial Differences in Stroke (REGARDS) study data with Medicare claims for algorithm development.
- Utilized adjudicated events from the REGARDS study as the gold standard for validation.
- Employed stepwise selection to identify key predictors from Medicare data and assessed performance using C-index, sensitivity, specificity, PPV, and NPV.
Main Results:
- The developed algorithms demonstrated strong performance in discriminating fatal CVD (C-index: 0.87).
- Achieved a sensitivity of 0.64, specificity of 0.90, PPV of 0.65, and NPV of 0.90 for fatal CVD.
- Claims-based algorithms showed improved reclassification of fatal events compared to using a 28-day post-hospitalization window.
Conclusions:
- Claims-based algorithms provide a more accurate method for identifying fatal cardiovascular events compared to relying solely on hospital discharge diagnosis codes.
- These algorithms enhance the utility of administrative claims data for epidemiological studies on CVD mortality.
Background:
Cause of death is often not available in administrative claims data.
Objective:
To develop claims-based algorithms to identify deaths due to fatal cardiovascular disease (CVD; ie, fatal coronary heart disease [CHD] or stroke), CHD, and stroke.
Methods:
Reasons for Geographic and Racial Differences in Stroke (REGARDS) study data were linked with Medicare claims to develop the algorithms. Events adjudicated by REGARDS study investigators were used as the gold standard. Stepwise selection was used to choose predictors from Medicare data for inclusion in the algorithms. C-index, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were used to assess algorithm performance. Net reclassification index (NRI) was used to compare the algorithms with an approach of classifying all deaths within 28 days following hospitalization for myocardial infarction and stroke to be fatal CVD.
Results:
Data from 2,685 REGARDS participants with linkage to Medicare, who died between 2003 and 2013, were analyzed. The C-index for discriminating fatal CVD from other causes of death was 0.87. Using a cut-point that provided the closest observed-to-predicted number of fatal CVD events, the sensitivity was 0.64, specificity 0.90, PPV 0.65, and NPV 0.90. The algorithms resulted in positive NRIs compared with using deaths within 28 days following hospitalization for myocardial infarction and stroke. Claims-based algorithms for discriminating fatal CHD and fatal stroke performed similarly to fatal CVD.
Conclusion:
The claims-based algorithms developed to discriminate fatal CVD events from other causes of death performed better than the method of using hospital discharge diagnosis codes.
Related Concept Videos
01:27Psychoneuroimmunology: Cardiovascular Disease
A key area of focus in PNI is the relationship between stress and coronary...
01:11Testing a Claim about Mean: Known Population SD
Estimating a population mean requires the samples to be distributed normally. The data should be collected from the randomly selected samples having no sampling bias. The sample size needed to be higher than 30, and most importantly, the population standard deviation should be already known.
In most realistic situations, the population standard deviation is often unknown, but in rare circumstances, when it...
01:24Testing a Claim about Population Proportion
There are two methods of testing a claim about a population proportion: (1) Using the sample proportion from the data where a binomial distribution is approximated to the normal distribution and (2) Using the binomial probabilities calculated from the data.
The first method uses normal distribution as an approximation to the binomial distribution. The requirements are as follows: sample size is large...
01:19Testing a Claim about Standard Deviation
The hypothesis testing for the claim of population standard deviation (or variance) requires the data and samples to be random and unbiased. The population distribution also must be normal. There is no specific requirement on the sample size as the estimation is based on the chi-square distribution.
As a first step, the hypothesis (null and alternative) concerning the claim about...
01:21Testing a Claim about Mean: Unknown Population SD
Estimating a population mean requires the samples to be approximately normally distributed. The data should be collected from the randomly selected samples having no sampling bias. There is no specific requirement for sample size. But if the sample size is less than 30, and we don't know the population standard deviation, a different approach is used;...
01:12Trial and Error and Algorithm