A Quality Improvement Approach to Enhance LGBTQ+ Inclusivity in Pediatric Primary Care

This QI project was conducted from 2017 to 2018 at a single outpatient pediatric primary care clinic in Chapel Hill, North Carolina, that serves as a training site for >60 pediatric residents. The clinic’s faculty, residents, and staff provide care to ∼8000 patients per year, about 70% of whom are insured by Medicaid. The hospital system uses a version of the Epic EHR (Epic Systems Corporation, Verona, WI) that permits documentation of SO/GI in discrete fields; this capability was enabled shortly after project onset.

A QI approach exemplified by the Institute for Healthcare Improvement's Model for Improvement was used to perform small tests of change.²³  A key driver diagram helped establish objectives. The 14-person project team included front desk staff, nurses, residents, faculty, clinic leadership, EHR leadership, department leadership, content experts, and a transgender patient and his parent. In addition, QI team members were of diverse gender identities, sexualities, races and ethnicities, ages, and religious backgrounds. The team met at least monthly to develop Plan-Do-Study-Act (PDSA) cycles and review progress.

The QI team strongly agreed with national recommendations for training personnel in the language of LGBTQ+ youth¹⁷  and believed this should occur before changes in SO/GI collection methods. Clinicians and staff completed the National LGBTQIA+ Health Education Center’s BASE readiness assessment, which has been used to assess multiple other health care organizations.²⁴  This determined baseline knowledge and comfort, as well as areas on which to focus subsequent trainings.

This study used multiple instructional methods to accomplish this goal while accounting for complex personnel schedules. Clinicians were invited but not required to attend staff trainings.

• First, the National LGBTQIA+ Health Education Center delivered a 4-hour train-the-trainer event on inclusive care. Motivated staff and clinicians in various clinic roles (nursing, front desk, clinicians) attended.

  • ^ Participating clinic staff and nursing then delivered a 1-hour training for all clinic staff.

  • ^ Participating residents gave a separate 1-hour lecture to clinicians.

• Several weeks later, clinic and nursing staff attended a question-and-answer session with a transgender teen and his father.

• In a subsequent training, all clinic and nursing staff participated in a role-play exercise practicing different patient scenarios, including a clinic walk-through to identify welcoming or alienating visual cues, areas of inadequate privacy, gendered signage, and LGBTQ+-specific resources.

• A pediatric endocrinology fellow delivered three 1-hour lectures on gender-affirming care to residents.

• As a resident clinic with predictable, significant physician turnover annually, ongoing education is necessary. Incoming residents attend a 4-hour LGBTQ+ inclusivity training led by the University of North Carolina’s LGBTQ+ Center and are instructed on SO/GI documentation during clinic orientation. New staff members are instructed on SO/GI documentation during employee orientation.

Staff and clinicians were asked to repeat the National LGBTQIA+ Health Education Center’s BASE survey at 6 months to assess change over time. This survey contains multiple domains, including 2 questions about preparedness, 1 asking about “lesbian, gay, and bisexual” patients and the other about “transgender” patients; other sexual and gender identities are not explicitly addressed. Four-point Likert scales were dichotomized (strongly agree and agree versus disagree and strongly disagree); baseline and 6-month follow-up were compared using t test.

We next sought to display visual cues of a welcoming environment. Three large, rainbow-colored posters reading “You Are Safe Here” were displayed in common spaces and “Do Ask, Do Tell” signs developed by the National LGBTQIA+ Health Education Center were placed in all examination rooms.²⁵  Clinic personnel were offered wearable symbols, such as rainbow ribbons or pronoun pins, which have been suggested by national organizations to indicate that clinicians will affirm identities and that disclosure is safe and important.¹⁹ 

Three months after project initiation, Epic’s SO/GI function was enabled at the system level, allowing SO/GI and pronoun documentation in discrete fields. This 3-month phase was deemed the “baseline,” with electronic facilitation of SO/GI documentation beginning “week 1” (Fig 1). Physician-patient partnership facilitated by the QI team was important for respectful implementation of this function; for instance, following team feedback, chosen name was displayed more prominently than alternate identifiers.

A number of PDSA cycles sought to identify a sustainable method to systematically, privately collect SO/GI and pronouns early in teen acute visits (defined as all non-well visits because screening already occurred in well encounters). These occurred in the baseline phase (Fig 1) before implementation of the Epic SO/GI section. To assess feasibility, we measured the proportion of nursing staff who had collected SO/GI and pronouns at least once (Fig 2).

• First, nurses triaged adolescents ≥11 years separately from caregivers, allowing the patient to answer sensitive questions confidentially and early. Parents of younger adolescents felt uncomfortable with early separation, and most nurses felt uncomfortable asking about SO/GI and pronouns aloud.

• Second, during check-in, the clinic’s front desk staff gave adolescents a small, commercially available paper card designed to collect SO/GI and pronouns.²⁶  Patients were instructed to give completed cards to nurses or clinicians. No patients returned cards.

• The next cycle occurred after a staff training (week 3), and used the initial separate triage method, this time for teens 13 years and older. Only 1 additional nurse felt comfortable asking verbally.

• Finally, the QI team developed a novel card that asked about SO/GI and pronouns, as well as a variety of other questions important to address for all adolescents (“teen health cards,” Fig 3). These additional topics included screenings for suicide risk, substance use, desire for contraception, sexual activity, and abuse. Where available, validated questions were used. In addition, these cards asked whether others were aware of SO/GI, how best to distribute confidential materials, and whether the patient desired to speak with a physician alone. Front desk staff provided the cards to all patients 13 years and older at check-in for acute visits regardless of whether previously completed to simplify the process and ensure SO/GI data were updated over time. They were not used during well visits, when other, longer screening tools were already in use. Adolescents completed the cards while in the waiting room and gave them to the nurse during triage. Nurses entered SO/GI and pronouns into appropriate EHR fields, then gave the card to residents before their entering the room, ensuring clinicians were aware of identifiers and able to budget time to address positive screens. Nurses and clinicians were instructed to note patients’ responses regarding who was aware of SO/GI and pronouns, and to document diverse identifiers only with patient permission. Within several days of beginning this cycle, all nurses had used the method at least once, and it was incorporated into processes easily.

• The card itself underwent multiple PDSA cycles based on patient and staff feedback, including modification from a flat to folded card to promote privacy, and additional changes to format and language. Adolescents completing intake cards were under no obligation to disclose SO/GI honestly. We felt this was acceptable to prioritize patient safety and avoid unintentional “outing.” We also added the question “Do you need to speak with your doctor alone today?” to allow those desiring support but feeling unsafe disclosing SO/GI or concerns on paper an opportunity for such separation, which happens inconsistently during acute visits. Screening tools are routinely used to assess for depression or high-risk behavior during well care, but not for every acute visit. This tool became an opportunity to address a variety of patient needs that may not otherwise be identified during acute visits for LGBTQ+ as well as cisgender, heterosexual patients.

SO/GI documentation in the designated Epic section was assessed by direct chart review for all patients ≥13 years presenting for acute care visits. Initially, only per-encounter SO/GI documentation was recorded. However, some patients presented repeatedly over short periods. Beginning with week 11, we instead noted whether SO/GI had ever been recorded in the Epic SO/GI section (“Collection Method Change” in Fig 1). Chart reviews for SO/GI were considered positive if any elements of sexual orientation, gender identity, or pronouns were documented. This was to respect patient autonomy in selective sharing of identifiers and because any positive element showed that the process of data collection and documentation had occurred. Staff were incentivized to distribute cards and record SO/GI via a competition for a coffee shop gift card (week 7), and then reminded in twice-daily huddles (beginning week 27). Rates of SO/GI documentation were tracked and displayed using run charts (Fig 1). Medians were calculated using modified Institute of Healthcare Improvement methodology, starting at baseline (0) and increasing when observed proportions were higher than the previous median for 8 consecutive weeks.

We initially intended to track the balancing measure of patient satisfaction comparing before and after surveys. However, a brief, unvalidated, anonymous baseline survey of a convenience sample of English-speaking families found that all respondents felt the clinic was already excellent at caring for LGBTQ+ patients. Because we had demonstrated insufficient staff and clinician self-efficacy and lack of SO/GI documentation at acute visits, we sought an alternate method to track satisfaction, asking staff and clinicians to report all relevant patient comments to QI team leads.

This project was deemed exempt by the University of North Carolina’s institutional review board (17-2261). It was funded by the university’s Institute for Healthcare Quality Improvement.

Seventy of 110 (64%) team members completed the survey at baseline and 44 (40%) did so 6 months later. At baseline, fewer of the clinic’s staff and clinicians felt prepared to care for LGBTQ+ patients than has been reported nationally. In our clinic, the proportion of respondents who described feeling prepared to care for lesbian, gay, and bisexual patients was 53% (compared with 80%-96% nationally), and for transgender patients only 30% (compared with 68%-94% nationally).²⁴  At 6-month follow-up, the proportion that felt prepared to care for lesbian, gay, and bisexual patients increased to 68% (P = .07), whereas the proportion feeling prepared to care for transgender patients nearly doubled to 57% (P = .002).

The clinic initially contained no visual cues of LGBTQ+ inclusivity. At the project’s conclusion, 100% of examination rooms, waiting areas, and hallways displayed welcoming signage.

SO/GI documentation during acute care teen visits in the specific SO/GI section in Epic increased from 0% at baseline to a median of 26.5% in the first 9 weeks of the intervention, to 70% at the conclusion of the active intervention, and to 90% during the 20-week sustainment phase (Fig 1). The median number of acute visits for teens was 20 per week, mean was 19.6 per week, and range was 4 to 34 visits per week. The teen health cards were rapidly incorporated into intake protocols by nurses, 100% of whom reported collecting SO/GI data at least once within a week of introducing the cards (Fig 2). SO/GI and pronoun documentation remained stable even after graduation of a project colead (medical student), parental leave of the other colead (attending), and graduation of one-third of the trained residents. This indicates that the intervention was sustainably integrated into processes. Because the same cards that asked about SO/GI also asked about self-harm, substance use, family support, and safety, all LGBTQ+ adolescents identified with this method were also screened for psychosocial comorbidities. Clinicians used existing pathways to address these other needs. During the active phase, 14% of acute visits with SO/GI documentation were for LGBTQ+ patients, and 25% of these had not had routine well care in the preceding 2 years, indicating SO/GI and pronouns would likely not have been known for those patients without this project.

Team members collected anecdotes about patients’ experience with the new intake cards and clinic signage. One family refused the card. Two staff members expressed discomfort about the topic but remained motivated to provide inclusive care and participated in processes. Multiple patients and parents expressed gratitude for signs of inclusivity. At least 1 patient used the intake card to normalize their own gender diversity to parents. At least 1 patient who had been misgendered by staff for years during acute visits used the intake card to remedy this. Overall, staff and clinicians heard many positive sentiments from patients and families about clinic changes.

This project demonstrates that consistent, sensitive, and confidential documentation of teens’ SO/GI and pronouns is attainable in a busy academic resident clinic with brief training and enhancements to the clinic environment. Modernized intake documents and the cultivation of a “disclosure-positive” environment empowered patients to identify themselves and facilitated appropriate psychosocial screening. This project also serves as a template for implementing SO/GI screening in a setting where clinic staff may have variable levels of comfort and knowledge; however, it is clear that ongoing support and training is necessary to ensure all staff can provide inclusive care for LGBTQ+ patients. Supportive leadership and a robust multidisciplinary team that represented all stakeholders ensured that patient-focused change was attainable clinic-wide and sustainable even in the absence of project leads. Sustainability was enhanced by incorporating changes into existing processes and by avoiding reliance upon select team members with special interest in the project.

Methods of collecting and documenting SO/GI and pronouns have largely focused on adult populations.^19–21,27  Far less consideration has been given to the processes necessary to obtain this information privately and safely from children and adolescents, when these patients may be most in need of an affirming adult. This method of SO/GI and pronoun collection can help pediatric practices establish themselves as welcoming spaces and connect adolescents to necessary support, which may improve patients’ health care utilization across a lifetime.

This project had several limitations. Patient satisfaction data were a balancing measure rather than primary aim and were not collected in detail beyond anecdotes. Informal feedback demonstrated minimal evidence of discontent and multiple indications of benefit; therefore, this would not have appreciably altered the project’s course. In addition, fewer respondents completed the follow-up survey about preparedness to care for LGBTQ+ patients, introducing the possibility of selection bias. Data analysis and manuscript submission were delayed by the global coronavirus pandemic; however, the processes described remain pertinent. Finally, although rates of SO/GI documentation improved, we do not know patients’ future willingness to seek care or the effects of SO/GI reporting on health. The ability to measure such longer term outcomes is hindered by the ongoing coronavirus pandemic, which has affected many patients’ mental health and ability to seek care, regardless of clinic inclusivity.

This project showcases a method that other organizations can use to sustainably collect and communicate SO/GI and pronouns early in acute care visits. Doing so empowers staff and clinicians to address patients correctly, permits an opportunity to explore the need for further psychosocial support, and promotes an inclusive environment likely to improve patient access and care quality in the future. This process was seamlessly incorporated and sustained beyond the active intervention phase. Similar efforts in other settings should continue to create inclusive spaces while also prioritizing patient safety and confidentiality. Future studies should investigate if SO/GI disclosures are accurate, if proper SO/GI usage improves likelihood to seek care, and if improved clinician awareness of gender and sexuality optimizes health outcomes.

The authors thank University of North Carolina’s Institute for Healthcare Quality Improvement for funding and coaching; the National LGBTQIA+ Health Education Center for initial staff trainings and surveys; Elizabeth Sandberg, MD, for resident education about gender-affirming care; QI team members not listed as authors; and University of North Carolina Children’s Primary Care Clinic’s residents, clinicians, and staff for their active involvement.

EHR

electronic health record

LGBTQ+

lesbian, gay, bisexual, transgender, queer, inclusive of other identities

PDSA

Plan, Do, Study, Act

QI

quality improvement

SO/GI

sexual orientation and gender identity