Correlates of Missing Information on Race and Ethnicity in the Georgia COVID-19 Surveillance Database Public
Bonuedie, Hillary (Spring 2021)
Published
Abstract
Objectives: We quantified the amount and pattern of missing race and ethnicity information among COVID-19 cases and deaths in Georgia at the individual and county levels.
Methods: Among confirmed COVID-19 cases and deaths recorded between April 2020 to March 2021 in the Georgia Department of Public Health surveillance database, the percentage of cases and deaths missing race and ethnicity data was calculated. This was compared with the percentage of missing age, sex, zip code, and county of residence among COVID-19 cases and deaths. At the individual level, correlates of missing race and ethnicity information among COVID-19 cases were identified using logistic regression. At the county-level, linear regression was used to identify correlates of differences in the percentage of cases with missing race and ethnicity.
Results: Confirmed COVID-19 cases were missing race and ethnicity information more often than confirmed COVID-19 deaths. The difference in missingness between cases and deaths, respectively, was more pronounced for information on race (18.6% vs 0.8%) and ethnicity (27.6% vs 1.0%) than for age (0.7% vs 0.01%), sex (1.1% vs 0.1%), zip code (2.3% vs 0.8%), or county of residence (2.0% vs 0.26%). At the individual level, in a logistic regression model of COVID-19 cases, males ages 0-17 (OR = 2.50; 95% CI: 2.05, 3.04) and males ages 18-64 (OR = 1.69; 95% CI: 1.49, 1.92) had higher relative odds of missing race information when compared with women ages 65+; age and sex patterns for missing ethnicity information were similar. At the county level, in adjusted linear regression models, the main correlates of the percentage of cases missing race and ethnicity, respectively were case rate (=0.99; 95% CI: 0.27, 1.72 for race and =1.47; 95% CI: 0.55, 2.38 for ethnicity) and the percent of the county population reporting multiple races (=0.76; 95% CI: 0.026, 1.49 for race and =1.18; 95% CI: 0.25, 2.10 for ethnicity).
Conclusions: The Georgia COVID-19 surveillance system was successful at collecting age and sex information, yet race data were missing for nearly 1 in 5 cases. Both individual demographics and county-level characteristics were informative in predicting missing race and ethnicity information among cases.
Table of Contents
INTRODUCTION 1
LITERATURE REVIEW 3
METHODS 8
RESULTS 14
DISCUSSION, CONCLUSIONS, AND PUBLIC HEALTH RECOMMENDATIONS 17
REFERENCES 23
TABLES 26
FIGURES 33
SUPPLEMENTARY TABLES 37
SUPPLEMENTARY FIGURES 42
About this Master's Thesis
School | |
---|---|
Department | |
Subfield / Discipline | |
Degree | |
Submission | |
Language |
|
Research Field | |
Mot-clé | |
Committee Chair / Thesis Advisor | |
Committee Members | |
Partnering Agencies |
Primary PDF
Thumbnail | Title | Date Uploaded | Actions |
---|---|---|---|
|
Correlates of Missing Information on Race and Ethnicity in the Georgia COVID-19 Surveillance Database () | 2021-05-07 14:11:21 -0400 |
|
Supplemental Files
Thumbnail | Title | Date Uploaded | Actions |
---|