Editor’s note: In a related piece, former US Assistant Secretary for Aging and administrator of the Administration for Community Living Kathy Greenlee also reflects on the elimination of the sexual orientation and gender identity measure from the National Survey of Older Americans Act Participants.

Over the past two decades, the field of Lesbian, Gay, Bisexual, and Transgender (LGBT) health research has grown exponentially, thanks, in part, to dedicated researchers who developed strategies for identifying LGBT people in federal data and to a series of initiatives aimed at increasing sponsorship of LGBT health studies by the National Institutes of Health and the National Institute on Minority Health and Health Disparities. Under the Obama administration, measures of sexual orientation were added to 12 federal surveys and studies and gender identity measures were added to seven. Federal programs, like Medicare, now actively incentivize collection of information on sexual orientation and gender identity (SOGI). States have also made key contributions here, with more than half of states including a measure of sexual orientation and/or gender identity in their version of the Behavioral Risk Factor Surveillance System (BRFSS) questionnaire.

LGBT Data Collection Rollbacks

Recently, the Trump administration has begun to roll back gains made in the inclusion of LGBT populations in federal data collection by canceling consideration of SOGI measures in the 2020 Census and eliminating SOGI measures or abandoning proposals to include them in important federal surveys, including the National Survey of Older Americans Act Participants and the Annual Program Performance Report for Centers for Independent Living, both administered by the Administration for Community Living (ACL). Dropping the sexual orientation measure from the National Survey of Older Americans Act Participants is particularly problematic for our understanding of the challenges older LGBT adults are facing since the survey collects information on support, service utilization, and program gaps for older Americans.

According to the ACL, the SOGI questions were dropped due to the small sample size of older adults identifying as LGBT. Unfortunately, such claims are not warranted. Previous studies on older adults—such as the Health and Retirement Study and the Medicare Current Beneficiary Survey—have added questions on sexual orientation despite small sample sizes. Furthermore, even with small sample sizes, researchers can combine multiple years of data to sufficiently analyze sexual and gender minorities. In the context of this rollback, we reiterate Randall Sell’s recent call for all federal surveys to include SOGI measures and for LGBT-inclusive research to remain on the federal agenda.

Two Things We Can Do Now to Advance LGBT Research

At the same time, it is clear that the current administration is unlikely to honor such requests. To this end, we contend that, now more than ever, it is important to:

1) make the data we do have count, and

2) create the data that others count on by conducting LGBT-inclusive research.

First, in the absence of a mandate to include SOGI measures in all federal data, we must make the best possible use of the data we do have. State-level surveys like the BRFSS that have already added SOGI measures are valuable resources that can be examined individually or pooled across states and years to examine the effects of state initiatives on LGBT health, disparities, and resiliencies. Other surveys such as the National Health Interview Survey, the National Survey of Family Growth, the National Health and Nutrition Examination Survey, and National Longitudinal Study of Adolescent to Adult Health also offer a rich set of sexual behavior and sexual identity questions. In addition, researchers can continue to advance LGBT research using data in which gay and lesbian adults and couples are identifiable through deduction, such as the Decennial Census, the General Social Survey, the National Health and Social Life Survey, the American Community Survey, and the Current Population Survey. Although these data sources have some specific limitations because they do not directly measure an individual’s sexual orientation or gender identity, they have enabled ground breaking studies on various health, social, and economic outcomes for same-sex couples and their families.

Second, we must create the data that others count on. All researchers who engage in original data collection have immense power to include SOGI measures in their work as standard practice. This includes adding SOGI measures where they are absent and expanding the population that responds to these measures. The National Health and Nutrition Examination Survey, for example, includes measures of recent sexual behavior and sexual identity but does not collect them from participants over age 59.

Widespread changes in attitudes and the law as it relates to LGBT Americans also provide a unique opportunity to justify LGBT-inclusive expansions of current studies in areas such as health, aging, social networks, insurance, family formation, child bearing and raising, and employment. In order for these inclusions to be meaningful, investigators might consider applying for administrative supplements to existing awards, foundation funding, and other small grant mechanisms to add an oversample of LGBT populations to current research. Oversamples are often employed to provide additional data on minority and hard-to-reach populations to ensure that researchers have adequate sample sizes for statistical analyses. Oversampling is also an efficient and cost-effective way to incorporate LGBT populations into existing research because investigators are able to reuse instruments and trained personnel. While these mechanisms may not be available or attainable for all studies, they are a useful way to advance LGBT-inclusive research in some cases.

As an example, the inclusion of an oversample of older LGBT adults through an administrative supplement to a larger, representative study of older adults in the NIH-funded UC Berkeley Social Networks Study (UCNets) has provided unique opportunities to examine how access to marriage affects network structure, availability of support, and health outcomes among older LGBT adults. The UCNets sample of older adults includes approximately 700 randomly sampled 50-70-year-olds living in one of six counties in the San Francisco Bay Area in 2015. To adequately address small sample size issues for older LGBT adults, the UCNets team added an oversample of about 300 LGBT adults aged 50-70 who were recruited via snowball, venue-based, and targeted online sampling methods. The oversample was recruited approximately four months after completion of recruitment for the random sample in the same geographical area and used the same survey instrument. Together, these data attend to the various empirical and methodological limitations of previous research on the effects of social networks on health in older LGBT populations.

The Next Step In LGBT Data Collection

While we must continue to make high-level demands that the federal government recognize LGBT populations in national surveys, there are several other data sources that we can make widespread use of right now to advance LGBT health research. Additionally, it is imperative that we practice what we preach at other levels of data collection by making our own studies inclusive, and expanding data collection where we can through the use of oversamples. Oversampling LGBT populations in representative surveys should be the next step of LGBT data collection, not the retraction of existing questions that ascertain sexual orientation and gender identity.