Abstract
Background: Heart disease has many predisposing factors. Genetics, lifestyle, socio-economic status have all been shown to play a role. The National Health and Nutrition Examination Survey (NHANES) combines data from interviews and physical examinations from approximately 5000 people each year in the United States. It is an excellent source for acquiring nationally representative data on known cardiovascular risk factors. By its nature, survey data, such as from NHANES, frequently has missing entries. Multiple imputation provides a statistically robust way to handle missingness. Rather than discarding partially complete entries in a listwise manner, multiple imputation uses a Bayesian model to produce multiple datasets that include uncertainty on the missing data. The datasets are then recombined to provide a complete dataset with more accurate standard errors than would be obtained by other imputation methods.
Methods: We used the R statistical programming language to download and process anonymized NHANES data from the 2017-2018 data acquisition cycle. Several parameters known to have a bearing on cardiac health were analyzed. Multiple imputation was used to handle missingness in the data. Survey weighting was also used to account for under/over-represented demographic groups in the data. Logistic regression was carried out the parameters using the presence of heart disease as the dependent variable.
Results: Preliminary results in this study show that the strongest predictors for heart disease were having a first-degree relative suffering from a myocardial infarction before the age of 50, followed by higher Hgb A1c values. The greatest “protectors” against heart disease were having a greater number of family members in the house, followed by more weekend nightly sleep hours.
Conclusion: It is no surprise that family history of early myocardial infarction and high A1c values are strong risk factors to acquiring heart disease. However, it may be less obvious that sleep acquired during the weekend and household family size would have much of a bearing. It could be the case that weekend sleep compensates for any sleep deficit acquired during the workweek and thereby reduces physiologic stress from sleep deprivation. Regarding household family size, perhaps having a greater number of dependents fosters more responsible lifestyle behaviors.
Methods: We used the R statistical programming language to download and process anonymized NHANES data from the 2017-2018 data acquisition cycle. Several parameters known to have a bearing on cardiac health were analyzed. Multiple imputation was used to handle missingness in the data. Survey weighting was also used to account for under/over-represented demographic groups in the data. Logistic regression was carried out the parameters using the presence of heart disease as the dependent variable.
Results: Preliminary results in this study show that the strongest predictors for heart disease were having a first-degree relative suffering from a myocardial infarction before the age of 50, followed by higher Hgb A1c values. The greatest “protectors” against heart disease were having a greater number of family members in the house, followed by more weekend nightly sleep hours.
Conclusion: It is no surprise that family history of early myocardial infarction and high A1c values are strong risk factors to acquiring heart disease. However, it may be less obvious that sleep acquired during the weekend and household family size would have much of a bearing. It could be the case that weekend sleep compensates for any sleep deficit acquired during the workweek and thereby reduces physiologic stress from sleep deprivation. Regarding household family size, perhaps having a greater number of dependents fosters more responsible lifestyle behaviors.
Original language | American English |
---|---|
Pages | 38 |
State | Published - 17 Feb 2023 |
Event | Oklahoma State University Center for Health Sciences Research Week 2023 - Oklahoma State University Center for Health Sciences, 1111 W. 17th street, Tulsa, United States Duration: 13 Feb 2023 → 17 Feb 2023 https://medicine.okstate.edu/events/index.html?trumbaEmbed=view%3Devent%26eventid%3D160681489 |
Conference
Conference | Oklahoma State University Center for Health Sciences Research Week 2023 |
---|---|
Country/Territory | United States |
City | Tulsa |
Period | 13/02/23 → 17/02/23 |
Internet address |
Keywords
- heart disease
- multiple imputation
- logistic regression
- NHANES
- R programming language