Developing and Implementing Questions for Collecting Data on Sexual Orientation and Gender Identity1
Key Points Raised by the Individual Speakers
- To address limitations in the way gender identity is represented in EHRs, a two-question approach may provide more accurate information while affording more respect to transgender people.
- Experience from surveys shows that it is feasible to collect accurate information about sexual identity and behavior in a diverse population.
- EHR users and vendors are working together to address issues relating to data collection on sexual orientation and gender identity in order to develop standardized questions that best serve the needs of patients.
“In a perfect world,” said panel moderator Nancy Bates, of the Census Bureau, “we wish we had some tested, generally accepted, embedded set of standard questions for gender identity and sexual orientation that worked in all different settings and for all populations. We aren’t there yet.” However, researchers are working toward that perfect world, and
1 This section is based on the presentations of JoAnne Keatley, Director, Center for Excellence for Transgender Health, University of California, San Francisco; Kristen Miller, Director, Question Design Research Laboratory, National Center for Health Statistics; David Grant, Director, California Health Interview Survey; Edward Callahan, Professor of Family and Community Medicine, University of California, Davis, Health System; and Denise Rasmussen, Epic Systems Corporation.
workshop participants heard from several of those who have expertise in this area and who conveyed some of the lessons they have learned in survey settings and other settings. In this chapter, five speakers discuss efforts to design questions for collecting data on sexual orientation and gender identity.
GENDER IDENTITY DATA COLLECTION
A working definition of the word transgender, said JoAnne Keatley, is a person whose gender identity or gender expression differs from the sex assigned to them at birth. She reiterated a comment made earlier in the workshop that many transgender people do not like to self-identify as transgender, instead wanting to have their affirmed gender recognized without it being connected to their transgender history. She also pointed out the distinction between gender identity—an internal sense of self regarding gender—and sexual orientation, which is a feeling of sexual or romantic attraction to gender or gender expression of other individuals and is not necessarily linked to anatomy. This distinction can cause confusion in health surveys, an example being the case of a man who is in a relationship with a transgender woman often being categorized as a man who has sex with a man when in fact that may not be the perception of that particular individual. In fact, there is a history of the CDC capturing transgender individuals and their partners under the men who have sex with men category.
In Keatley’s opinion, EHRs currently have limitations in the way they represent gender identity. If a transgender person’s sex is listed in terms of anatomy and the only mention of that person’s transgender status is in a provider note, other providers may not go to the notes section of the EHR and the result can be a disconnect that impacts the patient. Another limitation concerns denial of care from insurance coding that has to be gender-specific—if a transgender female is identified as female on an EHR, they may be denied prostate screening. To address these limitations, the Center of Excellence for Transgender Health recommends that transgender-inclusive data be collected using two questions. First, ask about the person’s current gender identity, and then ask for the person’s assigned sex at birth. Asking about current gender identity first honors the patient’s gender expression by allowing the patient to self-identify in their current gender identity. It may be necessary to ask additional questions to specify current anatomy, but that can be done through drop-down menus or through provider questions.
Regarding implementation, the first step is to update the health care system’s data collection forms and databases to include the two-question system. The CDC, Keatley noted, implemented the two-question system
in 2011 in its adult HIV case report form and its Enhanced HIV/AIDS Reporting System. She added that CDC data collected using the two-question system registered a 64 percent increase in transgender people compared to just asking current gender identity without asking about sex assigned at birth.
The next step in implementation is to train all staff and health care professionals to ask questions about gender identity correctly and consistently. Then, it is important to educate funders, health departments, and government agencies about the need to aggregate and disseminate the collected data. Only through data aggregation and dissemination will it be possible to determine the size of the transgender population and identify specific health risks and health care needs of transgender people. The last step is to use the disseminated data to improve services and programs for transgender people.
DEVELOPING A “PERFECT” SEXUAL IDENTITY MEASURE
Developing the sexual identity question for the National Health Interview Survey (NHIS) involved multiple steps that Kristen Miller described, starting with settling on what exactly the survey needed to measure. Sexual orientation, she said, is not a measurable concept, but rather is a generic catch-all term that describes an essence of how an individual feels. Instead, sexual orientation is an amalgam of three other concepts: sexual behavior, sexual attraction, and sexual identity.
Sexual behavior looks at actual same-sex versus opposite-sex sexual behavior and is not necessarily consistent with self-conception or presentation of self. One problem in designing a question to quantify sexual behavior is that definitions of what counts as “sex” vary across sub-groups. Sexual attraction refers to same sex versus opposite sex or gender desire, and it, too, is not necessarily consistent with self-conception or presentation of self. Sexual attraction is a latent and elusive phenomenon that is neither tangible nor observable. Miller said that the problem with designing a question to quantify sexual attraction is that there are different variations in what people think of as sexual attraction. Sexual identity describes an individual’s conscious identification of an understanding of self, and in that respect, is similar to racial identity because it represents the individual’s relationship to their social world. Sexual identity is a complex phenomenon that may fluctuate.
The key to designing questions around these three concepts is to not conflate them, Miller stated, adding that the term “sexual orientation” should be eliminated when used with scientific data because it does not reflect the realities of people’s lives. This is particularly true, she added, for women, and in her mind, the use of the term “sexual orientation” does
a disservice to the LGBT community. She also noted that in designing questions, it is important to consider whether the questions will be asked in a provider’s office or by a field survey interviewer.
Over the past decade, the National Center for Health Statistics (NCHS) has used a number of different questions to collect data on sexual identity in a variety of surveys, and the data are inconsistent across these surveys. The result, said Miller, is a distorted picture of what is happening within the lesbian, gay, and bisexual community. In a 2002 survey that used the terms heterosexual, homosexual, and bisexual to refer to sexual orientation, lesbians were identified as the most obese group of sexual minorities. In contrast, a 2006 survey that used the terms “straight or heterosexual,” “gay or lesbian or homosexual,” and “bisexual” found that lesbians were the least obese group.
To develop the sexual identity questions for the NHIS, Miller and her colleagues conducted seven cognitive tests involving 377 cognitive interviews. They also conducted a number of quantitative studies and examined a variety of different surveys that used some variation of a sexual identity question. These studies showed that LGBT people answer questions from a different perspective than non–sexual minority individuals, and that this difference stems in large part from the fact that non–sexual minority people are unfamiliar with the terms used in the questions. Based on these findings, the NCHS revised the questions designed for the NHIS based on three principles. First, use labels that respondents use to refer to themselves. Second, do not use labels that respondents do not understand. Third, use follow-up questions to meaningfully categorize “something else” and “don’t know” answers. Following these principles, the NCHS developed the sexual identity questions listed in Box 5-1.
She noted that though the second question includes a transgender option, this is not meant to be a transgender question. Rather, it is meant to address the complex relationship between gender and sexual identity. In closing, Miller said that this set of questions successfully categorized all but a few respondents, and that these questions will be included in the next NHIS.
ASSESSING SEXUAL IDENTITY AND BEHAVIOR IN CALIFORNIA’S DIVERSE POPULATION
The California Health Information Survey (CHIS) is an omnibus, multilingual, public health telephone survey of 42,000 to 56,000 California households conducted biannually since 2001 and continuously since 2011. CHIS has included sexual identity questions since 2001 and gender of sexual partner questions since 2003. The questionnaire development process, explained David Grant, involves multiple stakeholders organized into an advisory board, technical advisory committees, and both ad hoc and topic-specific workgroups. The general rule with CHIS is to never write new questions but to use questions from other well-established surveys. One caveat is that questions designed for data collection using registration forms or patient portals can be difficult to use in a phone survey, so questions are often adapted for that particularly medium, as well as to account for the linguistic diversity in California. Extensive testing and training and ongoing monthly monitoring of live interviews completes the process.
BOX 5-1 NCHS Sexual Identity Questions
Do you think of yourself as…?
Lesbian or gay
Straight, that is, not gay
By Something Else, do you mean that…?
You are not straight, but identify with another label such as queer, trisexual, omnisexual, or pan-sexual
You are transgender or transsexual
You have not or are in the process of figuring out your sexuality
You do not think of yourself as having a sexuality
You personally reject all labels of yourself
You made a mistake and did not mean to pick this answer
You mean something else What do you mean by something else? __________________________
By Don’t Know, do you mean that…?
You don’t understand the words
You understand the words, but you have not or are in the process of figuring out your sexuality
You mean something else What do you mean by something else? __________________________
SOURCE: Miller presentation.
In the first CHIS cycle in 2001, the survey asked two sexual identity questions of adults under age 65. The first was a yes-or-no question: Are you gay, lesbian, or bisexual? The word lesbian was omitted if the respondent was male. If yes, the second question clarified the answer to the first
question. There was a concern that there would be a significant hang-up rate with these questions, but, in fact, these questions posed no more of a problem than other questions in the survey. An issue that did arise was that respondents would often answer the first question by stating their sexual identity, which created problems for the interviewers and slowed down the interview process. An analysis of the 2001 survey data high-lighted the need for sexual behavior questions, and these were added in 2003. Beginning in 2003, they have asked a single sexual identity question that reads, “Do you think of yourself as straight or heterosexual; as gay, lesbian, or homosexual; or bisexual?” There is also the option of opening a text box and inputting an answer verbatim. In addition, there are two questions on sexual behavior. In cycles since 2003, these questions have been only asked of adults under the age of 70 because of a high rate of misclassification concerning the word “bisexual” in adults over age 70.
Discussing data from CHIS 2009, Grant noted that about 3 percent of men reported that they had exclusively male sex partners in the preceding 12 months, and about 1.6 percent of women reported having exclusively female sex partners in the preceding 12 months. Less than 0.5 percent of both men and women reported having sex partners of both sexes. From the sexual identity questions, about 2.5 percent of men identified as gay or homosexual, and 1.1 percent of women identified as gay, lesbian, or homosexual. One percent of men and 1.7 percent of women identified as being bisexual. Non-response to these questions was low, about 2-3 percent. In contrast, 23 percent of the sample refused to answer questions on household income. Non-response, which combines “don’t know” and refusals, was highest among Cantonese and Spanish speakers. Women who spoke Asian languages—Cantonese, Korean, Mandarin, and Vietnamese—were two to six times less likely to respond than men, which Grant said was unexpected.
In summarizing the lessons learned from over a decade of surveys, Grant said that it is clearly important to collect data on sexual identity in EHRs. “Electronic health records have great potential, especially for trans-gender populations because even in a large survey like CHIS, we’re likely to pick up very few or too few to make much statistical inference about that population,” he explained. He added that the CHIS results show that it is feasible to collect relatively accurate information about sexual identity and behavior in a diverse population, but that the wording of questions is important, as is the setting in which the questions are asked.
SEXUAL ORIENTATION AND GENDER IDENTITY IN THE ELECTRONIC HEALTH RECORD: A USER AND VENDOR PERSPECTIVE
The University of California, Davis (UC Davis) Health System began collecting demographic data in its EHR in 2009. At the time it was felt by the Health System’s EHR task force that it could not include sexual orientation or gender identity information in the EHR. A subsequent decision to appoint a new task force on inclusion of sexual orientation and gender identity in the UC Davis Health System elicited opposition from some administrators who felt that clinicians would not be comfortable discussing this issue or that the topic was too sensitive to be discussed in the context of an EHR. And even though this resistance was eventually UC Davis was set to roll out its sexual orientation and gender identity questions in November 2012—Callahan said this experience convinced him that the only way that this information will be collected in EHRs on a widespread basis is if the federal government requires it to happen.
The taskforce saw this opposition as an opportunity to make a compelling case to both the clinical and LGBT communities for the need to collect these data. In essence, said Callahan, the task force developed a business case based on the premise that there are significant, even devastating, health disparities that LGBT people experience because of the de facto “Don’t Ask, Don’t Tell” attitude that pervades the medical community. “If we can desensitize our providers to be able to talk about sexual orientation and gender identity,” said Callahan, “they might learn that LGBT people are part of us, not some group of ‘other.’” In fact, the task force’s efforts, he said, are changing attitudes and the culture at UC Davis. One tangible result of this attitude change is the development of a new medical school curriculum that incorporates LGBT health issues across the entire 4-year process.
One of the most compelling pieces of evidence for including sexual orientation and gender identity data in EHRs, said Callahan, comes from Caitlin Ryan and the Family Acceptance Project at the University of California, San Francisco. In a 2009 study, Ryan focused on young adult LGBT people and looked at their self-reports of how much acceptance and support they received from their families as they came out and the current status of their health. For those who got little support, the adverse effects were large: 8 times the amount of suicidal ideation, 8.4 times the amount of suicide attempts, almost 6 times the amount of serious depression, and significant increases in illegal drug use and unprotected sex. These data point to the value of being able to identify these young adults and provide them with necessary counseling and other health services.
For the past 3 years, Callahan and his collaborators at UC Davis have been conducting Grand Rounds about LGBT health disparities and are
in the process of rolling out some written educational materials for all of the system’s providers. They are also surveying these providers to identify those who are willing to be listed as LGBT friendly. The UC Davis Health System is going to use the MyChart patient portal feature of the Epic2 EHR to provide a questionnaire that patients can answer in the privacy of their own home that includes questions about sexual orientation and gender identity. It is also working with Epic to systematize where and how questions on sexual orientation and gender identity are located in the EHR. One idea is to tie the EHR to systematic prompts in response to certain answers to these questions. For example, when someone identifies as being transgender, the prompts would alert the care provider about what organ systems need to be monitored, what kind of preventive care needs to be presented, and how best they can be a good provider for that person. Callahan noted that Epic appears willing to make these changes to its EHR system.
In fact, said Denise Rasmussen, Epic is eager to receive input from users as to how to collect information on sexual orientation and gender identity in its EHR systems. Epic already has a field to capture information on sexual partners and runs reports and triggers alerts based on the sex of that partner. She noted, too, that at least some of the 260 organizations that use Epic have extended the sex category list to capture items related to gender identity and sexual orientation.
What would be helpful to Epic as an EHR vendor, she said, is a more standardized way to document information about sexual orientation and gender identity and more direction on where each birth sex, legal gender, and identified gender should be used in terms of billing, health information exchanges, clinical decision support, and preparing materials designed for patients. Epic is also concerned about who should document this information and who should be allowed to see it. For example, while it is obvious that the primary care physician should see this information, should the patient’s chiropractor?
Regarding technical issues, Rasmussen said that Epic knows that it needs to create a new field or fields to represent the sex of a patient in a way that better captures birth sex, legal gender, and identified gender and then display that information in appropriate places in the EHR. The company is also interested in making alerts more meaningful, that it is alerting the primary care physician to the issues that are appropriate to the patient in front of them: telling a physician to order a prostate exam for a female-to-male transgender patient is not helpful.
2 Epic is a company that makes software, including EHRs, for health providers and facilities. They are one of several large commercial vendors.
When asked by Jessica Xavier from HRSA why the NHIS includes a question about gender identity rather than behavior, Miller answered that the NHIS is a vehicle for collecting basic health data and including questions about sexual behavior would not be appropriate in the survey. She added that the National Survey of Family Growth (NSFG) and the National Health and Nutrition Examination Survey (NHANES) are more appropriate venues for asking sexual behavior questions. Clinton Anderson of the American Psychological Association asked Miller about the significance of the placement of the sexual identity and the sexual behavior questions. Miller explained that the placement of questions is important because respondents get cued in to certain topics. For example, if a question focusing on reproduction is asked after a number of behavioral questions, a respondent may be cued up to provide a sexual behavior answer. In contrast, asking the question as part of a demographic section may cue respondents to indicate sexual identity rather than behavior. Aaron Tax asked Grant if he had been able to identify questions that worked better with older adults who do not always understand terms like “bisexual” and Grant indicated that he hopes to add a question developed by Miller and her colleagues to the CHIS in 2013 that would be easier for older adults to understand. He added that moving away from the terms heterosexual and homosexual may help better identify bisexual older adults.
There was some discussion around the term “sexual orientation.” While Miller asserted during her talk that the term is not a measurable concept, Harvey Makadon pointed out that the term is commonplace and any attempt to substitute “sexual identity” or another term for it would be challenging. Miller responded that “sexual orientation” is a catchall phrase and does not provide constructs around which scientific questions can be developed. Makadon, however, said that he thought of orientation and identity as the same thing and questioned the need to abandon “sexual orientation.”
In response to questions from Bates and John Knudsen, Keatley discussed the two-question approach. She explained that the CDC’s one-question approach provided the following response options:
- transgender male to female
- transgender female to male
The two-question model asks about current gender identity separate from sex assigned at birth. She further explained that for the gender identity question, the suggested drop-down options are
- transgender woman
- transgender man
- gender queer
For the open-ended option, they ask people to specify so that they capture those data. Gal Mayer, medical director at Callen-Lorde Community Health Center in New York, who was participating by phone, e-mailed that at their community health center they do ask the sex at birth and current gender identity questions, the two-step process that Keatley discussed. However, they also ask sex listed on health insurance for billing purposes and that becomes the field for that particular purpose.
Leslie Calman noted that she was struck by the ambivalence that continually is expressed about who needs to know these data. On the one hand, she explained, we know that being a self-identified lesbian or gay man creates all kinds of health issues from childhood that relates to stress and discrimination. On the other hand, it is not clear who needs to know this information. For example, does a chiropractor need to know? Perhaps it is irrelevant. But maybe the patient’s back is out of whack because she is a lesbian; not because she has sex with women, but because of the stresses of being a lesbian. She noted that it has been a thread throughout the entire day, and it presents a dilemma for patients. Rasmussen responded that there was no clear answer. She theorized that it is a generational divide, with younger people more open to sharing this information with a wider range of care givers and older people being more hesitant to make these data widely available. Callahan added that as the culture shifts, he expects this will be less of a problem. Shane Snowdon responded to this by saying that in her opinion it was not only older people who were hesitant to disclose their sexual orientation, but younger people as well. She noted that San Francisco is an LGBT-friendly city, but that her experience working in a medical center there highlights that discrimination for sexual and gender minorities exists.
In response to a question from Harvey Makadon about how Epic is working with clients to coordinate efforts to develop a common approach to collecting sexual orientation and gender identity data in EHRs, Rasmussen explained that Epic had an LGBT breakout group at its user group meeting in September 2012. This user group had 10 to 15
organizations participating and there was a good exchange of information about what each organization was doing and how Epic might approach this problem until standards are developed. This is a work in progress, she added, and this group plans to meet frequently to provide advice to Epic. “As a vendor, we hate to push something out that is then going to change. If we’ve pushed that out, all of the different organizations will be collecting this information, and then potentially have to collect it again,” she said.
Ignatius Bau asked if sexual orientation was a field that users have customized and added to Epic products. Rasmussen said requests for adding a field on sexual orientation had come up and that it was relatively easy to do. She noted that the workflow around transgender documentation was more complicated. In response to a question about vendors working together around the issue of standards for EHRs, Rasmussen replied that while a standard that all vendors could follow would be ideal, what will likely happen is that each vendor will develop their own standards and then work together to map fields from each other’s EHRs so that they can exchange data accurately. This is the approach vendors have used in the past.