An Evaluation of Open Source Tools to Estimate the Reproduction Number of the 2009 Influenza A/H1N1 Pandemic in the USA Open Access
Kumbale, Carla (Spring 2018)
Abstract
Introduction: The reproduction number (R), which is a key epidemiological parameter to be estimated during emerging outbreaks such as the 2009 influenza pandemic, is computed using the R0 R-package. In addition, another R-package, EpiEstim, is also utilized to compute the instantaneous reproduction numbers (R(t)). Estimates are compared with the literature in order to further validate the use of these R-packages so that institutions such as the Centers for Disease Control and Prevention can implement these tools in order to rapidly compute these parameters during an emergency.
Methods: In the R0 package we use the Maximum Likelihood (ML), Exponential Growth (EG), Time-Dependent (TD), and Sequential Bayesian (SB) methods to compute four different estimates of R. In the EpiEstim package, we use the Bayesian Statistical technique implemented in this R-package to estimate R(t). A serial interval of 3.6 days with a standard deviation of 1.6 days is assumed for all of the estimates. These reproduction numbers are then compared to the literature.
Results: Several estimates are computed through the use of the R0 package. For the ML method, we estimate R values that vary between 1.41 to 3.54 depending on the selected time period of incidence cases. For the TD, EG, and SB methods, an R value of 1.88, 1.91, and 1.24 are computed respectively. Finally, the R(t) values, computed through the use of the EpiEstim package vary between 1.04-3.37 (weekly time window) and 1.24 – 3.33 (10-day time window). These values are subsequently compared with the results found in the literature.
Conclusion: This study has both demonstrated the use of and further validated the two computer packages that are now available for general use by non-modelers. Although the use of these packages still requires a certain minimum knowledge of statistical methods, the availability of these packages vastly improves the tools now at the disposal of public health practitioners during an epidemic/pandemic.
Table of Contents
INTRODUCTION ......................................................................................................... 1
Purpose ................................................................................................................... 5
Specific Aims ......................................................................................................... 6
METHODS .................................................................................................................... 7
Data ....................................................................................................................... 7
R0 Package ........................................................................................................... 9
EpiEstim Package ................................................................................................. 9
Description of Mathematical Methods ............................................................... 10
R0 - Maximum Likelihood Estimator(ML) ........................................... 10
R0 - Exponential Growth Method (EG) ................................................ 13
R0 - Sequential Bayesian Method (SB) ................................................ 16
R0 – Time Dependent Method (TD) ..................................................... 17
EpiEstim - Bayesian Statistical Inference.............................................. 18
Parameter Inputs: R0 Package .......................................................................... 22
Parameter Inputs: EpiEstim Package ................................................................ 23
RESULTS ................................................................................................................... 24
Aim 1 Results ....................................................................................................... 25
Aim 2 Results ....................................................................................................... 25
Aim 3 Results ....................................................................................................... 26
DISCUSSION ........................................................................................................... 28
Limitations and Observation ................................................................................ 33
Conclusion ............................................................................................................ 34
WORKS CITED ....................................................................................................... 35
FIGURES ............................................................................................................... 37
TABLES .................................................................................................................. 40
APPENDICES ......................................................................................................... 47
APPENDIX A ..................................................................................................... 47
APPENDIX B ..................................................................................................... 48
About this Master's Thesis
| School | |
|---|---|
| Department | |
| Subfield / Discipline | |
| Degree | |
| Submission | |
| Language |
|
| Research Field | |
| Keyword | |
| Committee Chair / Thesis Advisor | |
| Partnering Agencies |
Primary PDF
| Thumbnail | Title | Date Uploaded | Actions |
|---|---|---|---|
|
|
An Evaluation of Open Source Tools to Estimate the Reproduction Number of the 2009 Influenza A/H1N1 Pandemic in the USA () | 2018-04-24 18:25:35 -0400 |
|
Supplemental Files
| Thumbnail | Title | Date Uploaded | Actions |
|---|