Psychiatrist and data scientist uses ‘big data’ to serve minority populations

If health services researchers do not ask about distinct subpopulations, these populations will be invisible.

This is one of the important lessons from a study published last year in the journal LGBT health by psychiatrist Adrienne Grzenda, MD, Ph.D., former APA Health Services Research Fellow and current bioinformatics and “big data” consultant with the APA Research Division.

“It is unacceptable that minority segments of the population known to receive disparate care remain hidden in health care data and health surveys.” – Adrienne Grzenda, MD, Ph.D.

Grzenda’s study showed a dramatic increase in mental health distress among people from sexual and gender minorities associated with Donald Trump’s election in 2016.

The finding has implications for how mainstream political, social and cultural trends can sometimes directly impact mental health and distress – a finding that has been picked up by major news outlets including USA today and television network news.

But Grzenda believes the study also demonstrated the importance of asking the right questions and including populations in research that have been previously overlooked – in this case, people from sexual and gender minorities. Grzenda’s co-authors were Haiyong Xu, Ph.D., Jeanne Miranda, Ph.D., and Susan Ettner, Ph.D., all of the University of California, Los Angeles (UCLA).

“It’s only recently that large health surveys like the CDC’s Behavioral Risk Factor Surveillance System [BRFSS] even started asking about sexual orientation and gender identity,” Grzenda told Psychiatric News. “I find it gratifying that one of the things this study has revealed is the importance of representation in data. It is unacceptable that minority segments of the population known to receive disparate care remain hidden in health care data and health surveys.

She began the research that culminated in the August 2021 study during her two years as the inaugural APA Health Services Research Fellow during her third and fourth years of residency at UCLA. She is now on the faculty of UCLA’s David Geffen School of Medicine and UCLA-Olive View Medical Center. The two-year fellowship began in 2017.

Grzenda brought a remarkable resume to this investigation – she earned her MD and Ph.D. (in Biochemistry and Molecular Biology, studying Epigenomics) from the Mayo Clinic and a Master of Science in Bioinformatics and Computational Biology from the University of Minnesota. She completed her residency in the research stream at UCLA-Semel Institute and is certified in Psychiatry and Clinical Informatics.

She has applied this tremendous background to studying the mental health of LGBT populations. “When I applied for the APA Fellowship, my proposal was to investigate the mental health of gender minorities using underexplored types of data, such as insurance claims and social media data, through example, looking at accounts of suicide by transgender and non-binary individuals in social media posts,” Grzenda said.

“We know that there is a gap in mental health services for LGBT populations and that the lifetime rate of suicide attempts among transgender and non-binary people has persisted at around 40%,” he said. she declared. “How can we better capture information about the mental health and distress of gender minorities to provide better service? »

In the study, she and her colleagues used cross-sectional data from the 17 states that administered the sexual orientation and gender identity module in the 2015 and 2018 BRFSS surveys. The sample included 268,851 adult respondents: 12,006 sexual and gender minority adults and 256,845 cisgender and heterosexual adults.

The primary outcome was physical distress, mental distress, limited activity, and/or fair or poor general health for 14 or more days in the previous month.

Grzenda noted that in the run-up to the election, Trump and his running mate Mike Pence have spoken prominently about rolling back some of the LGBT protections enacted during Obama’s presidency. After the election, she and her colleagues noted in the study, 129 bills were proposed during the 2017 legislative season in 30 states, targeting LGBT and gender minority protections ranging from housing to health care. health through adoption.

“Our assumption was that we should be able to use this large representative sample [the BRFSS] to empirically test whether there has been an increase in LGBT mental distress in relation to the election,” she said.

She pointed out that a survey asking people if the election had caused distress would elicit many biased responses. But the BRFSS is a “cold” phone survey that includes general health and mental health questions.

Simply comparing pre-election and post-election changes in mental distress could be misleading as factors other than the election could be responsible. Grzenda and his colleagues therefore used a quasi-experimental approach known as “difference-in-differences” to calculate the impact of an event (in this case, the election of Donald Trump in 2016).

By comparing changes in outcome (mental distress) in a “treatment” group affected by the event to a “control” group, unaffected by the event, the design better controls for unseen confounders.

“That way, difference in differences allows you to make causal inferences,” she explained. Data analysis also controlled for sociodemographic variables, health care coverage, and chronic medical conditions.

They found that the LGBT population as a whole experienced a statistically significant increase in frequent mental distress compared to that of the cisgender and heterosexual sample, but the gender minority group (transgender and non-binary people) experienced a dramatic increase.

Compared to the cisgender and heterosexual population, the cisgender and lesbian and gay populations experienced an increase in primary outcome of 1.5%, a relative increase of 11.6% from 2015, although this is not statistically significant. Cisgender and bisexual or other respondents experienced a 4.8% increase over the cisgender and heterosexual population, an increase of 27.6% from 2015.

However, the transgender or gender non-conforming population saw a 14.8% increase over the cisgender and heterosexual population, a surprising relative increase of 117.5% from 2015.

“Previous studies indicate that gender-minority and bisexual people are often more vulnerable to mental health and health problems,” Grzenda and colleagues wrote. “Existing disparities can then sensitize these subgroups to the effects of acute or compounded stress. The overwhelming majority of proposed anti-LGBT laws during the period under review specifically targeted the rights of gender minority people. … The effect on subgroups may then be dose-dependent, with the increase in mental distress being proportional to the extent to which the individual perceives the impact.

The study was presented in USA today and on NBC News. It also drew criticism from conservative outlets such as Breitbart News.

Today, Grzenda uses her expertise in data science to help APA’s Research Division explore how large demographic datasets can be used to improve the mental health of individuals and communities.

“Big data is the next evolution of health services research,” she said. “How can we leverage ever-proliferating health data with machine learning and other artificial intelligence techniques to improve patient outcomes?” ■

“Impact of the 2016 Election on the Quality of Life of Sexual and Gender Minority Adults: A Differences-in-Differences Analysis” is posted here.

Sean N. Ayres