Highlights from the GenderSci Lab’s US Gender/Sex Covid-19 Data Tracker
AUTHORS: Ann Caroline Danielsen, Tamara Rushovich, and Mimi Tarrant
As a part of its COVID Project, the GenderSci Lab began collecting weekly data on cases and deaths for the fifty U.S. States, Washington, DC, Puerto Rico, and the US Virgin Islands on April 13, 2020. For the US Gender/Sex COVID 19 Data Tracker published today on our website, the GenderSci Lab calculated and analyzed percentages, crude, and age-adjusted rates for each state with available data, and provided longitudinal data showing changes in the sex ratio for each state across time.
Here, we offer key take-aways from this first data roll-out. Stay tuned in future weeks for more insights, including, we hope, from others who use the Data Tracker as a research tool.
How complete is the GenderSci Lab’s US Gender/Sex Covid-19 Data Tracker?
By and large, we found extensive reporting of sex-disaggregated data at the state level, and our Data Tracker captures this. The number of states reporting outcomes by gender/sex has increased over the tracked period. Initially, in the week of April 13, only 19 states were reporting fatalities by sex. Today, all states except for two (New Jersey and Hawaii) are reporting COVID-19 case counts disaggregated by sex.
Fatality data is presently more limited. Thirty-nine states report COVID-19 death counts disaggregated by sex. Occasionally, some state data is missing, so that male and female percentages do not add up to 100%. The “Unknown” classification for sex makes up to 5% of fatalities for some states. For more on our methods, click here.
Which states have the largest and smallest gender/sex disparities?
The simplest way to understand gender/sex differences (for more on why we say “gender/sex” rather than “sex” to describe these disparities click here) in COVID-19 cases and deaths is to look at the percentage of total number of cases or deaths that were women and the percentage that were men. See: COVID-19 Cases and Deaths Disaggregated by Sex
Of the states reporting cases disaggregated by sex, the percentage of cases that were women ranged from 43.8% to 58.3%.
Of the states reporting deaths disaggregated by sex, the percentage of cases that were women ranged from 40.0%-55.6%.
Table 1 shows the five states with the highest percentage of cases and deaths that are women and men, as of June 22, 2020.
Rate of cases and deaths per 100,000 women or men
Simply comparing percentages does not account for different underlying population distributions. For example, a state with more women dying of COVID-19 might just have more women living in the state overall. For this reason, it is useful to calculate the rate of cases or deaths per population. We did this by dividing the sex-specific number of cases by the sex-specific population of the state. See: COVID-19 Case and Mortality Rates Disaggregated by Sex
Of states that reported sex-disaggregated data, there were 33 states where the rate of cases per 100,000 people was higher among women than among men.
Of states that reported sex-disaggregated data, there were 10 states where the rate of death per 100,000 people was higher among women than among men. These states were: Rhode Island, Massachusetts, Connecticut, Delaware, South Dakota, Kentucky, Pennsylvania, Minnesota, Idaho, and Alaska.
Age-adjusted rates of mortality
Because COVID-19 has been shown to disproportionately impact individuals of older ages, and on average women have longer life expectancies than men in the United States, it is important to account for the age distribution of a population when comparing rates.
Because not all states report case or death counts disaggregated by age and sex, it was not possible to do direct age adjustment (1). However, indirect age adjustment was possible, using the age-specific overall U.S. rate of COVID-19 fatalities as the standard. Indirect age-adjustment is a statistical method that allows valid comparisons of rates across populations that have different underlying age structures. Using this method, rates are adjusted to have the same age structure as a standard population (in this case the U.S. population) so that they can be compared. See: COVID-19 Age-Adjusted Mortality Rates Disaggregated by Sex
When the GenderSci Lab calculated indirect age-adjusted mortality rates, every state had a higher rate among men. However, there is considerable variation in the magnitude of the difference between the states, with some states, (e.g., New York and Texas) having rates among men that were almost double those among women, while other states had rates that were statistically the same (e.g., Alaska, South Dakota, and Idaho).
Changes over time in gender/sex ratios
One feature of our Data Tracker is that we record weekly historical data reported on public health department websites each week from April to the present. This allows us to observe that of the 39 states with reported data, 31 had sex-specific age-adjusted fatality rates that became more equal to each other from the first week of available data to the most recent week. See: Time Series Rate Ratio of COVID-19 Age-Adjusted Mortality Rate
Understanding the reason for this trend is difficult as it could be due to changes in tracking and reporting of COVID-19 deaths (for example, there have been changes in the requirement of nursing homes to report COVID-19 deaths), or actual changes in the sex ratio of mortality rates. We might expect to see changes in the sex ratio of mortalty rates due to changes in exposure as States enact and then phase out different types of restrictions, which could impact women and men’s exposure to COVID-19 differentially.
The GenderSci Lab is working hard to explore these hypotheses, but limited data reporting from States on how variables such as occupation, comorbidity, age, and prison or nursing home confinement interact with gender/sex currently makes this challenging.
What matters is relative, not absolute, increases in mortality rates
Accounting for the underlying age structure is still only one step in really understanding what may be driving gender/sex differences in case and mortality rates of COVID-19. Other analyses have shown the importance of accounting for comorbidities and underlying differences in baseline risk of death.
Almost everywhere, men have a higher baseline mortality rate than women, especially at younger ages. For example, in a recent Lancet article, Dr. Nancy Krieger and colleagues from Harvard T.H. Chan School of Public Health analyzed all-cause mortality data for Massachusetts and compared the mortality rate during the COVID-19 surge in April 2020 to the mortality rate during the same time period in prior years. They found that the relative increase in excess deaths from COVID-19 in 2020 among men compared to women was identical.
The GenderSci Lab plans to conduct similarly nuanced analyses to better understand sex/gender disparities in COVID-19 outcomes, relative to baseline rates and across U.S. States and Territories.
Global variations of gender/sex disparities in COVID-19 outcomes
Just like at the U.S. state level, the gender/sex distribution of COVID-19 cases and deaths varies tremendously across the globe.
There are localities in which men make up a significant share of total COVID-19 deaths: for instance, men account for 80% of reported COVID-19 deaths in Haiti and for 77% of deaths in Bangladesh.
There are also, however, localities where the share of male COVID-19 deaths is close to 50% (Belgium, 51%; Northern Ireland, 52%; Portugal, 50%; Scotland, 50%; South Africa, 52%; South Korea, 53%), or under 50% (Canada, 46%; Estonia, 46%; Finland, 48%; Ireland, 49%; Slovenia, 41%), indicating that COVID-19 outcomes in relation to gender/sex can vary significantly depending on context.
Key take-homes
There is great variability in the gender/sex disparity in COVID-19 case and mortality rates nationally, globally, and over time.
In US States, the gender/sex disparity has narrowed over the 10-week time period captured by the US Gender/Sex COVID-19 Data Tracker.
Most, but not all, US States are reporting sex-disaggregated COVID-19 cases. Fewer are reporting sex-disaggregated COVID-19 deaths.
Exclusively reporting “female” and “male” counts and percentages of COVID-19 cases and deaths can lead to inaccurate conclusions.
Sex-disaggregated case and mortality rates should be analyzed in relation to the underlying population age distribution and sex ratio within each age group.
Sex-disaggregated mortality rates need to be analyzed in relation to baseline mortality rates for women and men.
Lack of intersectional data pairing sex-disaggregated data with variables such as comorbidities, occupation, race, age, institutional living environment, and other variables hinders efforts to analyze disparities in gender/sex outcomes.
Recommended Citation:
Rushovich, T. Danielsen, AC. Tarrant, M. “Highlights from the GenderSci Lab’s US Gender/Sex Covid-19 Data Tracker, ” GenderSci Blog, June, 24, 2020. Retrieved from: https://www.genderscilab.org/blog/covid-data-highlights
Statement of Intellectual Labor:
Rushovich, Danielsen, and Tarrant gathered and analyzed the data and drew up the initial draft. Rushovich integrated lab member comments and edits during the revision process. All authors provided substantive contributions to the ideas expressed in this blog post and participated in the preparation of the post.
Contact:
Questions, interested in collaborating with the GenderSci Lab, or media inquiry? Email us at: genderlab@fas.harvard.edu.
ENDNOTES:
As of May 1, 2020, the Centers for Disease Control and Prevention (CDC) provide provisional weekly updates of COVID-19 deaths broken down by age and gender/sex in all states. However, our analysis is based on deaths as reported by states themselves because the lag in time between when the death occurred and when the death certificate is completed, submitted to the National Center for Health Statistics and processed for reporting purposes by the CDC can range from 1 week to 8 weeks, or more.