The “Unknown” Side of State COVID-19 Gender/Sex Reporting
Authors: Kai Jillson and Heather Shattuck-Heidorn
In a forthcoming Health Affairs Blog, the GenderSci Lab, in collaboration with Harvard’s SOGIE Health Equity Research Collaborative, makes the case for the urgency of more complete reporting of COVID-19 data among trans and nonbinary people. We argue that it is critical to understand how gender-identity is represented in the COVID data in order to understand the effect of the pandemic on trans people.
Over the past year, the GenderSci Lab has been tracking state-level data on sex disparities in COVID-19 cases and mortality at our US State Gender/Sex Data Tracker. Using our State Report Card, we have also tracked and graded states on the comprehensiveness of their reporting of socially-relevant factors. From the beginning, a consistent question of lab members and those engaging with our work has been: how are individuals identified as female or male across states, and how do trans people fit into these data?
This piece takes a more in-depth look at the current reporting of individuals who identify as non-binary or who are unclassified by sex. The take-home, summarized in Tables 1 and 2 below, is that there are few states explictly collecting data on trans and nonbinary people, and that the ways in which states report “unknown” gender/sex lack transparency and are highly discordant across states.
The Third Sex Category
Since December 2020, lab member Kai Jillson carried out a series of interviews with state public health departments to answer the question of how trans people are represented in the data and how gender/sex is identified. Here, we document state health department responses about their gender/sex reporting and shed light on how that correlates with the categories available in their online COVID-19 dashboards.
Image Credit: United Nations on Unsplash
All 50 states report COVID-19 cases and fatalities through online dashboards. Since April 2020, the GenderSci Lab has collated these data on a weekly basis into an open-access dataset tracking gender/sex differences in COVID-19 outcomes. In our survey of state dashboards, we noted that states clearly disaggregate their data by gender/sex with labels like “female” and “male.” Whether creators and users of the database understand these labels as representing biological sex or gender identity is usually not clear.
Most states also list at least a third sex category, usually titled “unknown.” It is generally unclear what this category represents. For instance, it could indicate missing data, or could indicate that the sex categories given were insufficient. This “unknown” category can account for as much as 20% of the total data (Wyoming, March 8, 2021). The GenderSci Lab contacted 43 state public health departments in total (all the states with “unknown” categories) to gain information about sex reporting and the “unknown” category, and conducted 20 total interviews.
How States Collect Gender/Sex
In order to understand how gender/sex is reported, we need to understand how states collect information about COVID-19 cases and deaths. Gender/sex reporting methods vary in important ways across states, and can be different for COVID-19 cases vs. deaths. Some states, such as Oregon and Virginia, collect the gender/sex information for COVID cases at the testing location and through follow-up interviews. Virginia described their method of collecting gender/sex information as follows:
“We may get information on positive COVID-19 cases via testing data or from health providers. Labs or health providers may send over testing or case data that have missing demographic information, such as sex, age, or race/ethnicity. We may gain this information later during a case investigation; however, there is always a possibility of missing demographic data on COVID-19 cases.” (Virginia Department of Health, 2020, December 8)
Other states, like Tennessee, match patients to pre-existing electronic health record systems that contain gender/sex identifiers. If follow-up interviews with patients are performed, the public health department may confirm pertinent details such as gender/sex, age, employment, etc. The confirmed gender is then uploaded to the dashboard. Oregon and Wyoming will change the listed gender to that from the patient interview, if it does not match the pre-existing medical record. However, this data could still be listed as “unknown” if they are not able to reach the patient.
“Whether a state considers their ‘sex’ categories to represent gender identity or birth sex is rarely immediately clear”
States also vary in what demographic information they collect concerning COVID-19 cases and deaths. Some, such as Alaska, use the federal CDC’s reporting form for their own data collection method. Other states create their own form. These forms can differ significantly from the federal form. For instance, North Carolina has created their own form that requires local health department staff to choose between the gender options “M”, “F”, “M to F”, and “F to M”, for patients.
Along with North Carolina, several other states collect data on sex and gender identity. California’s morbidity report allows health departments to select “male”, “female”, “trans man/transman”, “trans woman/transwoman”, “genderqueer or non-binary”, and “declined to answer.” Also listed is the option to select “gender not listed (specify):______”, which gives state health departments the ability to be even more specific for gender reporting. Rhode Island’s reporting is extremely thorough and includes a demographic spreadsheet that records COVID-19 cases not only by gender identity but also by LGBT+ identity as well. Rhode Island’s table is broken down into the "sex assigned at birth” with “female”, “male”, and “other” categories listed. Gender identity is collected separately, with “cisgender man” and “cisgender woman” included, as well as a category that groups together people who identify as gender non-conforming, gender non-binary, and other transgender. Oregon is the only state that reports “non-binary” as an identity in their COVID-19 reporting, and Massachusetts is the only state to currently include “transgender” in their gender/sex reporting. This variable is only found in the mortality data and is not further disaggregated by male/female labels.
The Unknown
Generally, states reported that their “unknown” category refers to instances where the data was missing or left blank, the patient refused to answer, or the epidemiologist was not able to contact the patient. The data could be missing due to not asking the patient during initial collection, or if the electronic health records system was unable to match the patient to a pre-existing record. In other words, the “unknown” category results from a variety of fairly different causes. States indicated that both existing health providers and testing sites could be responsible for an “unknown” categorization, and emphasized that the sheer amount of information that is sent in to public health departments means that errors can occur.
“43 of 50 US states are not specifically reporting the effects of the pandemic on transgender and gender diverse individuals”
Our interviews indicated that the most frequent cause for the “unknown” category is probably missing data. There seems to be a general protocol that states follow to collect gender/sex on COVID cases/mortalities. First, the data is collected at the point of diagnosis. If the data is missing after that point, states first respond by checking to see if the patient is already in their online electronic health record (EHR) surveillance system. Such databases are state-specific and describe patient health information over time, and include demographic data on the patients in the system. However, not all patients will have a pre-existing EHR. Also, some states (Idaho, Vermont) report trying to confirm missing sex data in an interview, and using the “unknown” category if that is not successful (relatedly, Tennessee uses the category “pending” to indicate the same status).
Not every state reports using the EHR to fill in missing data. In Arkansas, which relies on physicians to document a patient’s gender/sex, the “unknown” refers to either an individual who preferred to be identified as gender-neutral or to a case in which the data was not included when submitted to the state. Tennessee explained that despite health department pleas for data on gender, they often do not receive it from reporting locations, and asserted that testing locations like pharmacies or drive-thru testing centers tend to make mistakes during reporting, which results in inaccurate information. Washington expressed something similar, stating that:
“While labs are legally required to submit birth sex of the patient with test results, they don’t always have that information. This also happens with things like address, county of residence, and race/ethnicity. If a lab doesn’t submit that information to the Dept. of Health, it can be very challenging for us to find it from other sources. As a result, some people with COVID-19 are shown as having “unknown” sex or “unknown” race/ethnicity.” (Washington State Department of Health, 2020, December 15)
In other localities, the “unknown” category may be operating as a marker for trans identities, either as assessed by the individual themselves, or possibly when gender is unclear to medical providers. For instance, South Carolina reports that their “unknown” refers to individuals who do not know their gender or do not wish to provide it.
Case versus Mortality Data
While 43 states have at least the third category (unknown or similar) for gender/sex in case data, only 25 states have categories beyond female and male for mortality data (Table 2). Health departments reported several reasons for this difference in reporting cases and deaths. First, more time is spent investigating COVID-19 deaths, which means fewer instances of data missing due to errors. This also explains why, in states that do report the “unknown” or “other” category for both cases and mortalities, there are very few deaths reported as “unknown.” For example, Alabama, which currently reports an “unknown” category for both cases and deaths, lists the percentage of “unknown” gender/sex as 1.6% for cases and 0.2% for deaths, meaning that approximately 6,100 COVID cases are gender/sex-unknown compared to about 14 COVID deaths.
Most states rely on the death certificate as well as an epidemiologist interview for attributions of gender/sex in the case of a death, though information may also be identified using the online health system. Again, there is variation in how data are collected from state to state. For instance, in Texas, case demographics are based on case interviews only and therefore represent only a fraction of the total number of cases, whereas fatality demographics are not based on interview but rather by the death certificate. States may also not report “unknown” categories because of the assumption that the sex of a deceased person is obvious. For instance, Delaware explained that “the reason why ‘unknown’ is not included is because for all deaths at this time, we know whether they are male or female, and none are ‘unknown’ so this option is not displayed on the page” (Delaware Public Health Department, 2020, December 8)
Best Practices for Gender/Sex Collection and COVID-19
Trans rights organizations recommend a two-step question process for distinguishing between sex and gender for collecting relevant data. Few states explicitly follow this practice. While several collect transgender identity, only Rhode Island appears to be asking for an individual’s sex at birth followed by their gender identity. This method, if put into practice more often, could increase the accuracy of gender/sex reporting.
“Disaggregating data with only female/male labels disregards those who identify outside of the cisgender binary.”
Currently, 24 states disaggregate a binary gender/sex category into “sex” categories, and 26 label the binary as “gender.” Outside of explicitly asking for sex at birth and gender identity, it becomes unclear which is being reported by these categories of “sex” or “gender”. States may implicitly consider the binary as indicating sex at birth. However, at least Tennessee and Vermont considered the “male” and “female” categories to implicitly represent gender identity, reporting that “Gender/sex is what is reported to us from the patient, regardless of what their birth sex was” (Tenessee Department of Health, 2021, January 6). States that utilize the patient interview process, like Oregon and Wyoming, appear to prioritize the patient’s described gender identity over their assigned sex at birth. However, whether a state considers their “sex” categories to represent gender identity or birth sex is rarely immediately clear.
The Future of Inclusive Data
43 of 50 US states are not specifically reporting the effects of the pandemic on transgender and gender diverse individuals, hindering our ability to understand the effect that the pandemic has had on these groups. Public health data that leans on the binary model of gender/sex neglects the COVID-19 health outcomes for trans and gender expansive people, ignoring the social and environmental factors that drive health disparities. Disaggregating data with only female/male labels disregards those who identify outside of the cisgender binary.
We recommend that health providers, testing locations, and public health departments streamline their communication of demographic data to prevent errors. State online dashboards also need to increase the options in their gender/sex reporting to include transgender and non-binary individuals who are currently either ignored, lumped into inaccurate categories, or possibly presented as “unknown.”
Clear data are critical for public health recommendations. The ambiguity, inconsistency, and lack of transparency in gender/sex reporting of COVID-19 outcomes is an obstacle to sound epidemiological interpretation of data on gender and sex disparities in COVID-19 outcomes, as well as public understanding of the effect of the pandemic on gender-diverse communities.
Recommended Citation
Jillson, K. & Shattuck-Heidorn H. “The ‘Unknown’ Side of COVID-19 State Gender/Sex Reporting,” GenderSci Blog, 2021 April 28. genderscilab.org/blog/unknown-covid19-gendersex-reporting
Statement of Intellectual Labor
Kai Jillson compiled the state-level data and authored the post, advised and edited by Heather Shattuck-Heidorn. A. C. Danielsen and S. S. Richardson provided valuable editorial feedback.
Image (From Unsplash):
https://unsplash.com/photos/fmhZecom1b8