Johnson, Matthew
(2016)
Analysis of racial differences in ICD shock burden using Cox proportional hazards regression.
Master's Thesis, University of Pittsburgh.
(Unpublished)
Abstract
Background/Research Aims: Sudden cardiac arrest (SCA) and sudden cardiac death (SCD) are major public health issues that most often result in death and affect a significant number of adults. Implantable cardiac defibrillators (ICDs) are an important tool used in SCA/SCD prevention. Racial differences with respect to ICD firings have not been very well explored with only limited previous research on the topic. This study seeks to analyze a large dataset of patients with defibrillators by race and also build a risk model for ICD shock using Cox proportional hazards regression.
Methods: Using data from the GRADE study, N= 1770 patients (1449 Whites and 321 African Americans) were initially compared by race for almost 80 baseline variables. N= 1524 patients (1275 Whites and 249 African Americans) had information on ICD shocks and were compared by race via Kaplan-Meier survival curves. A Cox proportional hazard regression analysis was performed to produce hazard ratios and evaluate each baseline characteristic with respect to defibrillator shock over time individually, followed by a multivariable model building process to examine race in the context of other significant covariates. Missing data was also examined in the terms comprising the final model and multiple imputations were performed for any variables deemed to have excessive missingness.
Results: Overall, African Americans were younger, had more nonischemic cardiomyopathy, had a higher prevalence of hypertension, smoked more, and had a lower ejection fraction than whites in the study. In comparing Kaplan-Meier survival curves, African Americans had a higher burden of shocks over time than whites by the end of the 5 year time period (p<0.01). Over 20 variables were individually related to shock at 60 months in a Cox regression setting (including race). The final multivariable model consisted of seven terms: race (not statistically significant but forced into the model), ejection fraction, history of NSVT, antiarrhythmic medication, diagnosis, age, and BUN. One variable was found to have excessive missingness (BUN) and after performing multiple imputations, results were overall similar in the second analysis except race became statistically significant (p= 0.04, HR= 1.385) while BUN went from statistically significant to not.
Conclusions: The public health impact of this study is in trying to build on the limited previous research and paint a more complete portrait of the relationship between race and appropriate defibrillator firings. After accounting for data missingness, race was found to be statistically significant in its relationship to ICD shock over time when adjusting for several additional covariates in the final multivariable model.
Share
| Citation/Export: |
|
| Social Networking: |
|
Details
| Item Type: |
University of Pittsburgh ETD
|
| Status: |
Unpublished |
| Creators/Authors: |
|
| ETD Committee: |
|
| Date: |
9 September 2016 |
| Date Type: |
Publication |
| Defense Date: |
2 June 2016 |
| Approval Date: |
9 September 2016 |
| Submission Date: |
26 May 2016 |
| Access Restriction: |
2 year -- Restrict access to University of Pittsburgh for a period of 2 years. |
| Number of Pages: |
48 |
| Institution: |
University of Pittsburgh |
| Schools and Programs: |
School of Public Health > Biostatistics |
| Degree: |
MS - Master of Science |
| Thesis Type: |
Master's Thesis |
| Refereed: |
Yes |
| Uncontrolled Keywords: |
ICD, defibrillator, race, Cox |
| Date Deposited: |
09 Sep 2016 17:33 |
| Last Modified: |
01 Jul 2018 05:15 |
| URI: |
http://d-scholarship.pitt.edu/id/eprint/28090 |
Metrics
Monthly Views for the past 3 years
Plum Analytics
Actions (login required)
 |
View Item |
![[img]](http://d-scholarship.pitt.edu/style/images/fileicons/application_pdf.png)
![[feed]](/images/feed-icon-32x32.png){kind=icon}