4

Existing Data Collection Practices in Clinical Settings1

Key Points Raised by the Individual Speakers

  • Transgender people are more likely to want to self-identify as such in an EHR than are lesbian, gay, or bisexual people. It is important, then, to consider these differing sensitivities when designing and implementing questions for use in a patient portal or EHR.
  • The primary concerns about collecting data on sexual orientation and gender identity among LGBT people relate to privacy and use: who is going to ask these questions, who will be able to see this information, and how will it be used to benefit the patient?
  • Soliciting input from patients and clinicians, and gaining buy-in from health care system administrators, is critical to developing appropriate and useful data-collection tools for deployment across a health care system.
  • Training health care system staff about working with LGBT people and the sensitivities involved in collecting data on sexual orientation and gender identity is essential for successful implementation of these tools.

____________

1 This section is based on the presentations of Shane Snowdon, Director, LGBT Health & Aging Program, Human Rights Campaign Foundation, and Founding Director, Center for LGBT Health and Equity, University of California, San Francisco; John Knudsen, Chief, LGBTI Practice Task Force, Mayo Clinic; Ronald Copeland, President and Executive Medical Director, Ohio Permanente Medical Group; Alex Gonzalez, Medical Director, Fenway Health; Jesse Ehrenfeld, Assistant Professor of Biomedical Informatics, Vanderbilt University; and Robin Weinick, Associate Director, RAND Health.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 25
4 Existing Data Collection 1 Practices in Clinical Settings Key Points Raised by the Individual Speakers • Transgender people are more likely to want to self-identify as such in an EHR than are lesbian, gay, or bisexual people. It is important, then, to consider these differing sensitivities when designing and implementing questions for use in a patient portal or EHR. • The primary concerns about collecting data on sexual orientation and gender identity among LGBT people relate to privacy and use: who is going to ask these questions, who will be able to see this information, and how will it be used to benefit the patient? • Soliciting input from patients and clinicians, and gaining buy-in from health care system administrators, is critical to developing appropriate and useful data- collection tools for deployment across a health care system. • Training health care system staff about working with LGBT people and the sensitivities involved in collecting data on sexual orientation and gender identity is essential for successful implementation of these tools. 1 This section is based on the presentations of Shane Snowdon, Director, LGBT Health & Aging Program, Human Rights Campaign Foundation, and Founding Director, Center for LGBT Health and Equity, University of California, San Francisco; John Knudsen, Chief, LGBTI Practice Task Force, Mayo Clinic; Ronald Copeland, President and Executive Medi- cal Director, Ohio Permanente Medical Group; Alex Gonzalez, Medical Director, Fenway Health; Jesse Ehrenfeld, Assistant Professor of Biomedical Informatics, Vanderbilt Univer- sity; and Robin Weinick, Associate Director, RAND Health. 25

OCR for page 25
26 COLLECTING SEXUAL ORIENTATION AND GENDER IDENTITY DATA Though the use of EHRs to collect data on sexual orientation and gender identity is still evolving, a number of health care systems around the nation are in the process of developing and implementing questions to gather that data. These health care systems include both single-city organizations that largely serve an LGBT population to huge, multistate providers whose treatment population includes LGBT people. This chap- ter relates some of these health care systems’ experiences and includes lessons that they have learned and suggestions for others who want to collect data on sexual orientation and gender identity in their EHRs. PATIENT SELF-IDENTIFICATION IN ELECTRONIC HEALTH RECORDS Each year, hundreds of health care facilities use the online national Healthcare Equality Index (HEI) survey, managed by the Human Rights Campaign, to evaluate themselves against established criteria for LGBT patient-centered care. Shane Snowdon explained that this is the only sur- vey in the country that invites hospitals and outpatient clinics to examine their practices and policies from an LGBT perspective, report those prac- tices and policies, and receive education and feedback about them. In the latest survey, completed in March 2012, 122 responses were received; 12 responses came from LGBT clinics, with the remaining 110 coming from general health facilities. Most of the latter facilities are in urban locations and consider themselves progressive or on the leading edge of health care. There are three types of LGBT patient data: same-sex partnership, transgender status, and sexual orientation. Patients, said Snowdon, have very different attitudes toward each of these types of data and about whether and when they want to volunteer or be asked about them. Same- sex partner­ hip, she added, is often overlooked in these discussions about s LGBT-relevant data, but patients consider it extremely important. “Patients with same sex partners really want to at least be offered the option to say that they have a same sex partner,” said Snowdon. “They do not want to look at or hear the options single, married, widowed, divorced.” Reflecting this preference, 58 percent of the facilities that participated in the HEI survey said that they do offer a partner or significant other option, either on a paper form or an EHR, and while this question is likely aimed at different-sex partners, same-sex partner couples appreciate the option. Of the facilities included in the survey, 37 percent are in so-called registration states that offer same-sex couples the opportunity to register their relationship status and receive various health care rights and respon- sibilities, and some 75 percent of the facilities in a registration state offer a registered domestic partner drop-down option. Snowdon pointed out that patients often want to volunteer their same-sex partnership during

OCR for page 25
EXISTING DATA COLLECTION PRACTICES 27 the admitting process and to have their partnership recognized in the health care setting. An increasing number of transgender people, Snowdon continued, want to self-indentify as such despite the fear and anxiety they may have about doing so. They do, however, want people to know how to commu- nicate with them respectfully, appropriately, and accurately. They want the entire health care system—the labs running clinical assays, the phar- macies filling prescriptions, and their physicians—aware of their sex at birth because transgender people know that information is important for them to receive the best care. Having said that, the most recent HEI survey found that only three clinics—two that specifically serve an LGBT popula- tion, the other serving a general population—offer a drop-down option for transgender status on their EHR. However, 78 facilities reported that they are actively considering a drop-down option on transgender status. The one general health facility that has a drop-down option has retained “male” and “female” as the options in the existing gender field, in order to match legal identification and insurance information, but also indicates “transgender MTF [male to female]” or “transgender FTM [female to male]” in a prominent drop-down list that a patient can access at their choosing. Snowdon noted that the facility made this modification to the standard EHR without involving the vendor. In contrast to the situation with transgender people, who increas- ingly want to self-identify, lesbian, gay, and bisexual people seem more reluctant to self-identify in the context of an EHR, Snowdon said. Provid- ing information in an EHR allows providers to know more about their patients, but many lesbian, gay, and bisexual people do not necessarily want their providers to know about their sexual orientation. EHRs are also envisioned to allow information to be shared more widely, and, in Snowdon’s opinion, many lesbian, gay, and bisexual patients do not want that information shared widely, if at all. While EHRs can be used to empower patients, many patients feel that being confronted with routine sexual orientation questions is disempowering. Summarizing what she and her colleagues have learned over the years, Snowdon said that lesbian, gay, and bisexual patients have many concerns regarding volunteering information about sexual orientation in the health care setting. They are unsure whether to respond in terms of attraction, behavior, or identity. In her opinion, women in particular are unsure about whether the question refers to today, the past, or the future because they tend to view themselves as more sexually fluid. Reflecting this challenge, medical students are now taught that sexual orientation occurs on a spectrum that is not necessarily fixed over a lifetime. They are also taught to not ask about sexual orientation, but rather to ask their patients about actual behavior.

OCR for page 25
28 COLLECTING SEXUAL ORIENTATION AND GENDER IDENTITY DATA Lesbian, gay, and bisexual patients worry, often with reason she said, that they will be treated differently if their sexual orientation is known. They may even delay and avoid care because of how they feel they will be treated, not just by physicians, but by the nurses, the allied health professions, and other facility employees. “So it isn’t just a question of, ‘Are my data going to be held confidentially in the research phase or in the EHR?’” said Snowdon. “It is, ‘I don’t even want you to ask me unless I volunteer it.’” This concern is especially prevalent, she added, in the 29 states in which lesbian, gay, and bisexual people do not have legal protections. Only 26 percent of facilities record sexual orientation in print or in an EHR, and no general health facilities record sexual orientation in an EHR using a drop-down option. Instead, the patient will note their sexual orientation on a print form or the provider will enter the patient’s sexual orientation in print or electronic notes. In contrast to same sex partnership and transgender status, which patients often want to volunteer, no general health facility reported patient requests to indicate sexual orientation at admitting or registration. On the basis of these findings, Snowdon made the following recommendations: • Explore not only the phrasing of sexual orientation questions and confidentiality of data in research but also who will be asking those questions, what training will those asking the questions receive, in what context will they ask the questions, and to what extent will the information be available within the health care system. • Convene diverse groups of patients to discuss their concerns because they are likely to differ substantially by gender, age, income, race and ethnicity, and health condition. • If routine sexual orientation questioning is incentivized or other- wise encouraged, it will be essential to provide widespread patient education and empowerment activities, to require facilities to have meaningful nondiscrimination policies and staff training in place, and to teach health care providers in a systematic way how to dis- cuss sexual orientation sensitively with patients. Concerning this last point, Snowdon said it would be interesting to see an initiative from CMS that would require LGBT nondiscrimination policies to go hand-in-hand with incentivizing data collection. As a final note, Snowdon said that EHRs need to be modified, both by vendors and facilities, so that LGBT patients can always, if they wish, ask that infor- mation on same-sex partnership, transgender status, and sexual orienta- tion be recorded. In doing so, however, it is essential to remember that a patient’s desire to volunteer that data can vary tremendously.

OCR for page 25
EXISTING DATA COLLECTION PRACTICES 29 THE MAYO CLINIC EXPERIENCE The Mayo Clinic, said John Knudsen, has only recently begun an effort to gather information on sexual orientation and gender identity. He noted that the Mayo Clinic started on its path thanks to what he called a union of forces that included four reports—HHS’s Healthy People 2020 report, the IOM’s 2011 report on the health of LGBT people, the Joint Commission report Advancing Effective Communication, Cultural Compe- tence, and Patient- and Family-Centered Care for the Lesbian, Gay, Bisexual, and Transgender (LGBT) Community, and the HEI survey—as well as a receptive leadership that includes clinicians and the Mayo Clinic LGBT Employee Resource Group. This resource group, explained Knudsen, started the process by requesting that the Mayo Clinic identify a group of providers in its primary care practice who were “LGBT-friendly,” and this led to an effort to collect data on some of the difficulties experienced by both LGBT patients and employees. This data collection effort is being led by the Mayo Clinic’s clinical prac- tice committee, which heard a wealth of personal and painful ­ tories that s patients and employees shared about suboptimal care they had received from the organization. Mayo Clinic’s leadership responded by creating the LGBT Practice Task Force, which was charged with studying the situation with regard to LGBT people and to develop recommendations that would facilitate and sustain an inclusive and welcoming environment for the Mayo Clinic’s LGBT patients and families. This effort also included intersex people, those born with an ambiguity about their sex assignment. One of the factors complicating efforts to collect data on LGBT and intersex people at the Mayo Clinic is that it is not just a hospital—it includes group practices and multispecialty practices based in Arizona, Florida, and Minnesota, and a regional health care system that includes 71 communities, 900 physicians, and 12,000 allied health staff spread across Iowa, Minnesota, and Wisconsin. In addition, the Mayo Clinic system uses three different EHRs. To address this last issue, the Mayo Clinic system has developed an enterprise-wide data governance and health informa- tion coordination policy that is creating uniformity and standardization across the entire system. Another complication arises from the fact that the Mayo Clinic system uses patient portals to input information, but this information is managed separately from the EHR. Knudsen explained that there is currently no mechanism or interface for transmitting data from the patient-provided information portals into the EHR. He noted, too, that changes are likely to be slow in being enacted, giving the example of the more than 4 years it took to make a change in the collection of race and ethnicity data based on recommendations from a research team. In terms of LGBT data, a stewardship group is developing a recommended list of data elements

OCR for page 25
30 COLLECTING SEXUAL ORIENTATION AND GENDER IDENTITY DATA and business rules around these data elements and codes. An external standards team will develop language and standards that will then need to be endorsed and approved by multiple stakeholders and Mayo Clinic leadership, all of which will be coordinated by the enterprise project management office at the Mayo Clinic. As a final note, Knudsen said that the goal of all of this activity is to collect data that will be able to trigger alerts for screening and preventive health services, track quality metrics, and create research opportunities. THE KAISER PERMANENTE EXPERIENCE Like the Mayo Clinic, Kaiser Permanente is also in the early stages of collecting data relevant to sexual orientation and gender identity. ­ aiser, K as the nation’s largest not-for-profit health plan, has long served a diverse population that today comprises more than 230 granular ethnicities speaking more than 170 different languages, said Ronald Copeland. As a result, Kaiser has had a great deal of experience managing and optimizing diversity to improve care for its patients and enhance the performance of its workforce. When Kaiser first introduced its EHR and began collecting data on its membership, its internal KP Pride Caucus, which consists of LGBT employees, questioned why the EHR collected data on race, ethnicity, and language, but not sexual orientation and gender identity. The answer was that the organization did not have the capacity to collect these data and it had concerns about the sensitive nature of asking questions about sexual orientation and gender identity. Instead of jumping right into what could have been a troublesome area, the leadership at Kaiser intended to phase these questions in based on internal research and in conjunction with an education and awareness campaign. To drive this effort forward, Kaiser created two study groups, one focused on data collection and implementation, the other on referral processes to physicians who were knowledgeable about and sensitive to issues specific to LGBT populations. The data collection study group oper- ated from the assumptions that future legislation will require and regulate the collection of LGBT data and that data will be collected in ­ aiser’s K HealthConnect EHR. It also assumed that technical issues about data structure and definition of terms could be overcome based on ­ aiser’s K experience with race, ethnicity, and language data and that while the final implementation may look different in each of the organization’s regions, the data format would be identical. Finally, the study group assumed that individuals will have the right to refuse to answer questions about sexual orientation or gender identity. The study group recommendations are presented in Box 4-1.

OCR for page 25
EXISTING DATA COLLECTION PRACTICES 31 BOX 4-1 Recommendations from Kaiser’s Study Group for Sexual Orientation and Gender Identity Questions The Kaiser study group recommended the collection of self-identified data at the Patient level, and recorded in Kaiser’s EHR system: 1. Sexual Orientation • How do you identify yourself sexually? 2. Gender Identity • What gender (sex) do you consider yourself? • What gender was recorded on your original birth certificate? 3. Sexual Behavior • In the last 24 months, with whom have you had sex? SOURCE: Copeland presentation. In the second phase of Kaiser’s efforts, the organization has created 11 centers of excellence that are research units within its clinical space. These units are designed to take ideas, create models, pilot the models, work out the details, and then present them for further evaluation across the entire organization. Action teams are now looking at issues involving technol- ogy and systems, operations and workflow, education and training of staff, and brand strategy. The technology and systems action team, for example, intends to develop and implement a valid and reliable system for storing, collecting, and using LGBT and intersex data by the end of the second quarter of 2013. Operating with the same deadline, the opera- tions and workflow action team will embed in the Kaiser infrastructure a clinician referral system while the education and training action group develops education and training programs that advance different levels of core competency in caring for LGBT and intersex people. The final phase will implement the process. THE FENWAY HEALTH EXPERIENCE The mission of Fenway Health, an independent, federally qualified health center in Boston, is to enhance the health of the LGBT commu- nity and other people in its neighborhoods and beyond through access to the highest-quality health care, education, research, and advocacy, explained Alex Gonzalez. In 2011, Fenway Health served some 18,000 unique patients, and Gonzalez estimated that maybe half of the patients are lesbian, gay, or bisexual, and another 800 are transgender. He based

OCR for page 25
32 COLLECTING SEXUAL ORIENTATION AND GENDER IDENTITY DATA these estimates on his experience taking care of patients over the past 7 years. While Fenway Health has been using an EHR system since 1997, it has captured sexual orientation and gender identity data only in an unstructured format as part of a careful social history or behavioral health intake. That kind of data, said Gonzalez, is trackable only by doing a chart review, which is unrealistic for 18,000 patients. Beginning in December 2011, Fenway Health began collecting sexual orientation data on the reg- istration form along with other demographic data, and it added informa- tion on gender identity in August 2012 (see Box 4-2). Text on the form indicates that the information is for demographic purposes only and will not impact patient care. The registration form also comes with a pam- phlet on patient rights that includes an explanation of confidentiality and articulates a philosophy of respect for each patient as an individual with unique health needs. In addition, Fenway Health now prints out a patient profile for patients who are already registered and are returning to care in order to have the patient double-check his or her own information. The patient profile also helps capture sexual orientation and gender identify data on people who have been long-time patients. BOX 4-2 Fenway Health Registration Form • Do you think of yourself as…? — Lesbian, gay, or homosexual — Straight or heterosexual — Bisexual — Something else — Don’t know • What is your gender? — Female — Male — Genderqueer or not exclusively male or female • What was your sex at birth? — Female — Male • Do you identify as transgender or transsexual? — Yes — No — Don’t know SOURCE: Gonzalez presentation.

OCR for page 25
EXISTING DATA COLLECTION PRACTICES 33 The registration form has one question designed to capture sexual orientation data in a structured format and three questions to capture data on gender identity, which are currently being pilot-tested. For sexual orientation identity, two versions—one with a write-in option and one without—were pilot-tested. For gender identity, Fenway Health convened a patient advisory committee and surveyed transgender patients on their preferences. Gonzalez said that since implementing these questions in its informa- tion packets and patient profiles, there has been no discernible pattern of non-response to the sexual orientation and gender identity ­ uestions, nor q have there been reports from front desk staff or providers of patients ask- ing why they are being asked these questions. It is his opinion that LGBT patients seem to appreciate being asked the questions, while non- LGBT patients seem to be unaffected by the questions. For trans­ enderg patients, the data collected during the test phase suggest that there are several groups of transgender persons. There were those who identify with their affirmed gender and do not identify as being transgender as well as those who do identify as being transgender. There were also those who identify along a non-binary spectrum, or as genderqueer—neither male nor female. So far, Fenway Health has been able to capture struc- tured data on 11-12 percent of all of its patients. Since the addition of the sexual orientation question in December 2011, roughly one-third of all patients have answered it. The questions on gender identity were only added in August, making it too soon to have acquired meaningful data. In his final remarks, Gonzalez discussed some additional consider- ation for transgender patients. Third-party payment systems require that name and sex be recorded in the EHR as they are listed with the insurer. This can be an alienating experience as insurance information often does not match identity. However, transgender patients report that they feel empowered when their preferred name and pronoun are recorded in a way that is honored by all staff at Fenway Health. THE VANDERBILT UNIVERSITY EXPERIENCE It is important to remember, said Jesse Ehrenfeld, that the vast major- ity of LGBT patients across the United States receive their health care from non-LGBT affiliated medical facilities such as Vanderbilt and that facilities such as Vanderbilt also have a large number of LGBT ­ mployees, stu- e dents, and patients. He estimated that at his institution, around 5 percent of the patients are LGBT and intersex people and added that in 2009, the Vanderbilt University Medical School initiated an extensive curriculum reform effort to bring LGBT health education to the forefront. Vanderbilt has revised its clinical intake forms to ensure that they are inclusive and

OCR for page 25
34 COLLECTING SEXUAL ORIENTATION AND GENDER IDENTITY DATA that they provide the necessary information to maximize the value of its EHR for providing care for its patients. It also launched a new program for LGBT health and created the nation’s second LGBT office in an aca- demic medical center. (The University of California, San Francisco, had the first.) With that institutional climate as background, Ehrenfeld discussed the work that he and his colleagues are carrying out on identifying bar- riers to disclosure of LGBT and intersex status by patients and how they have been using informatics tools to identify LGBT patients in their efforts to evaluate and reduce health care disparities. In a study conducted by colleague Sarah Fogel, participants recruited from gay and lesbian com- munity centers were asked questions about their willingness to disclose their sexual orientation to a provider. The study had a diverse group of subjects in terms of race, ethnicity, age, and education level, and 84 per- cent of the participants had disclosed their sexual identity to a health care provider. This study identified a long list of facilitators to disclosure, such as the health care provider asking about the patient’s sexual identity or making clear that this information was important for the patient’s care. It also identified a long list of barriers to disclosure, such as the sexual iden- tity information is written into the patient’s medical record and that staff other than the health care provider would have access to this information. Ehrenfeld then described some of his work using natural language processing to first identify LGBT and intersex patients from electronic health records that already exist and then determine how LGBT and intersex status affects treatment, diagnosis, and health outcomes. ­ atural N language processing, he explained, uses advanced algorithms and the rules of the English language to analyze free-text data, such as that in unstructured text notes in medical records, for specific information through trial and error. Using this approach on its entire enterprise data warehouse, Ehrenfeld and his collaborators were able to identify several thousand unique patients using search terms such as homosexual, LGBT, gay, lesbian, receptive partner, and others. They have since refined their search technique using concepts such as negation logic, semantic search, and other natural language processing tools to reduce a 22 percent false positive rate to three percent. Additional work is under way to further reduce the false positive rate to close to zero. An initial finding from an analysis of the identified records was that it takes an average of 30 months between the time of a patient’s first visit and when LGBT status appears in the medical record, and that patients had an average of 17 clinical encounters in which LGBT status was identified. Ehrenfeld described Vanderbilt’s approach to modifying its clinical intake forms and EHRs. This effort started with an extensive consensus- building process that involved Vanderbilt leadership, its patient and fam-

OCR for page 25
EXISTING DATA COLLECTION PRACTICES 35 ily council, a growing list of clinical departments, and its cultural and linguistic council. This process resulted in four recommendations that Ehrenfeld said are worth adopting at all medical centers: 1. Patients should have the opportunity to self-identify their sexual orientation and gender identity. 2. Patients should have the opportunity to define their relationship status. 3. Patients should have the opportunity to identify their family. 4. Patients should be asked about power of attorney/health care proxy and advanced health directives. Vanderbilt has now developed a three-stage implementation plan for modifying the 210 unique clinical intake forms used by its 95 physician practices. The recommended text for use on the intake forms includes two parts. The gender identification section asks for patients’ legal name; their preferred name, if different; their sex at birth; and their gender identity—male, female, transgender, and prefer not to answer. The sexual orientation section asks if the patient is heterosexual, straight, lesbian, gay, bisexual, queer, something else, or prefer not to answer. In developing these questions, the Vanderbilt team considered the same issues that the previous speakers had identified: Who is asking the questions? Where is the information recorded in the EHR? How often is the patient asked for this information? What to ask in terms of identity versus behavior? In summary, Ehrenfeld made the following three recommendations: 1. Self-identified orientation, gender identity, and sexual behavior should be collected during routine clinical encounters in order to improve health outcomes. 2. When collecting sexual orientation and gender identity data on health care intake forms, patients should have the right to opt out. 3. Prior to collecting this information, health care institutions should develop and widely distribute educational materials, not just for staff but also for patients, that explain clearly why developing and collecting this information is important and how it remains confidential. SUPPORTING PROVIDERS IN THE COLLECTION OF DATA Given the importance of training providers and health care system staff prior to introducing questions about sexual orientation and gender identity, it is critical to begin thinking about how to accomplish that train- ing early in the process, said Robin Weinick. For most health care systems,

OCR for page 25
36 COLLECTING SEXUAL ORIENTATION AND GENDER IDENTITY DATA training will likely need to be done using some kind of video tool given that the people who need training are going to be dispersed, both in terms of location and time. Training methods must also take into account that three-quarters of all office visits nationally are to practices with five or fewer clinicians. And most of those practices do not have the resources to conduct extensive training programs. Though there are important differences between collecting race and ethnicity data and sexual orientation and gender identity data, there are lessons to be learned from the experience of introducing race and ethnicity data into medical records. The first lesson is that the language provided to guide those who ask these questions must be well defined, that is, there should not be much left to the imagination of the person who is asking the question. A corollary to this lesson is that the information learned when testing questions in a research or survey context are not automatically transferrable to a clinical context, said Weinick. A second lesson learned from collecting race and ethnicity data is that it is extremely helpful to prepare an introductory statement for the data collectors to use that explains why it is important to gather the informa- tion. That introductory statement serves the dual purpose of explaining to the patient and the staff who are going to ask these questions why the information is important to collect. Another important lesson is that it is important to keep staff calm by identifying potential patient and staff objections in advance and build those into the training. Scripting responses so that staff can start to get comfortable with various situations is helpful, as are role-playing exer- cises. Another way to reassure staff is to provide them information from pilot data collection studies that show that if questions are asked in the right way, patients do not get angry. In addition, repetition is critical to making staff feel at ease. Weinick explained that the goal of training should be to drive home two points: (1) do not make assumptions about a patient based on how they look, act, or sound, and (2) patient self- identification is the only thing that matters. Three additional factors go into successful training of the people who will be asking these questions. First, they need to understand as con- cretely as possible why the issue is important. Second, they need to know clearly and very specifically what they are being asked to do. And, finally, they need to have skills and scripts for addressing patient concerns. DISCUSSION Aaron Tax, from Services and Advocacy for GLBT Elders, asked the four speakers about their experiences in collecting sexual orientation and gender identity data for elderly persons who may be more reluctant

OCR for page 25
EXISTING DATA COLLECTION PRACTICES 37 because of personal experience to divulge this information. Ehrenfeld said that his team certainly observed a generation gap in the willingness to self-identify on intake forms. Gonzalez said that Fenway Health’s approach to this problem has been to retain the ability to ask about sexual orientation and gender identity in the clinical setting rather than during registration because it is easier to put people at ease and educate them about why this information is important. Snowdon noted that this is not just a problem at the older end of the spectrum—a large longitudi- nal study of youth conducted by the Human Rights Campaign found a similar reluctance among youth between 16 and 24 years of age. Harvey Makadon noted, though, that in his experience, older LGBT people feel more comfortable when they talk to their physicians about their sexual orientation or gender identity because it makes them feel that their health care professional is listening to them. George Brown of the Department of Veterans Affairs asked how inter- sex people are supposed to identify themselves when the sex-at-birth question offers just male or female options. Snowdon replied that these choices are dictated by insurance company systems that only accept male or female. In that respect, that field only reflects the legal identification or insurance record of the patient. Intersex status is collected in the gender identity drop-down list. Barbara Warren, director for LGBT Health Services at Continuum Health Partners in New York, relayed her experiences in working with the New York State Office of Mental Health to add questions on sexual orientation and gender identity to the standard patient admissions form. The challenges she faced in adequately training personnel was reflected in the low rate of practices in New York State including those questions and the resulting small amount of data that was obtained. She added that they were still working to improve the training component. Weinick responded to this comment by discussing the need to approach repetitive training creatively. She used sexual harassment training as an example of training that uses role playing and vignettes to reinforce the training components. The issue for some LGBT patients, explained Snowdon, is that just being asked these questions will make them very anxious and may make them less likely to seek care. She asked if there was an equivalent of that in race and ethnicity. Rebecca Fox, of Wellspring Advisors, pointed out that collecting ethnicity data in a community clinic in Arizona, and ask- ing about a patient’s Latino background, may make some patients very a ­ nxious. It could lead them to not seek care because they are afraid of being asked what about their racial and ethnic background and that may bring up fears around immigration status. Weinick agreed that while bias against race and ethnicity differs somewhat from bias against sexual ori-

OCR for page 25
38 COLLECTING SEXUAL ORIENTATION AND GENDER IDENTITY DATA entation or gender identity, there is some overlap in terms of fear possibly driving patients away from health settings. Bau commented that there were similarities with language. For example, in some cases when patients are asked if they need language assistance, they refuse because they mis- takenly believe it will make them vulnerable to discrimination. In fact, the right to language assistance is guaranteed and could benefit them. In response to a question from Eva Powell, from the National Partner- ship for Women and Families, about whether natural language process- ing could be used in combination with information gathered directly from the patient as a means of improving the quality of the resulting data, ­ hrenfeld said that the natural language processing methods were a E research tool, not an operational tool. He does not expect natural language processing to be used in an operational setting, though it could be used in a quality improvement setting. He added that natural language process- ing will not provide useful information if the physician has never asked the patient about sexual orientation or gender identity or if the patient has never volunteered that information. When asked by Susan Queen from the Office of the Assistant Secre- tary for Planning and Evaluation at HHS about whether information from a patient portal can be transferred directly to an EHR, Knudsen said that yes it can, and that at the Mayo Clinic, patient-provided information is incorporated into the EHR. Ehrenfeld added that at Vanderbilt, informa- tion from the patient portal is first verified by the provider before it is added to the patient’s EHR. Queen asked Copeland if the response categories for Kaiser’s sexual orientation question was a set of fixed categories or did it prompt the individual to write in a response. Copeland replied that at the present time, the patient writes in the response. He added that in all of their data collection of this type Kaiser tries to have some predetermined responses, but they also want to have the patient self-identify whenever possible. They have found that, in terms of engaging patients, using the terms the patient uses is probably the most effective way to move forward with con- versations between patients and clinicians and care teams. In his opinion, this approach is preferable to forcing them to make a choice and fit it into a category that has already been pre-determined. A participant attending by phone asked about patient autonomy and whether the data collected could be used for research without patients’ consent. Knudsen said that in Minnesota patients sign a form that can remove their data from any future research. This means that even if a researcher has approval from an institutional review board, certain patients can opt out of the research and they are given the opportunity to make that decision when they are seen as patients.