Abstract
Background:
Transgender and nonbinary youth (TNBY) experience discrimination which has been linked to suicidal ideation and self-harm. Few studies have examined this relationship systematically. We studied the relationship between gender dysphoria coding and hospitalization for suicidality and/or self-harm in a large representative database from the United States.
Methods:
Using the 2016 and 2019 Kids’ Inpatient Database, we identified TNBY youth with gender dysphoria between 6 and 21 years of age using ICD-10 codes. We identified suicidal ideation using explicit “suicidality” codes, and self-harm using 355 self-harm codes. Prevalence of suicidality and/or self-harm was compared between youth with and without gender dysphoria. Multivariable regression tested for an association between gender dysphoria and suicidality or self-harm.
Findings:
Gender dysphoria had a prevalence of 161 per 100,000 admissions between 6 and 21 years of age in 2016 and 475 per 100,000 admissions in 2019. Subjects with gender dysphoria were disproportionately White, privately insured, and from higher median income than those without dysphoria. Prevalence of suicidality in subjects with gender dysphoria was 36% in 2016 and 55% in 2019, which increased to 41% and 66% for either suicidality or self-harm, compared with 4 to 6% for subjects without gender dysphoria. In multivariable modelling, subjects with a gender dysphoria diagnosis had a 3 to 5-fold greater prevalence of suicidality and/or self-harm relative to those without gender dysphoria codes.
Interpretation:
In a large representative national sample, TNBY with gender dysphoria were frequently hospitalized for suicidality and/or self-harm. The fewer non-White, publicly insured, and low-income youths with gender dysphoria suggest that underlying inequities may shape the identification and management of gender dysphoria. Structural and provider-level interventions are needed to reduce discrimination and expand gender-affirming competencies to prevent adverse outcomes for hospitalized TNBY with gender dysphoria.
Funding:
NIH K23-HL136688 (NY)
INTRODUCTION
Transgender and nonbinary youth (TNBY), those whose gender identity does not align with their assigned sex at birth, experience stigma and discrimination which has been linked to adverse mental health outcomes,1,2 including an increased prevalence of suicidality and self-harm.1–6 Survey and interview studies from North America have reported rates of suicidal ideation as high as 60% among TNBY adolescents,2,6 and suicide attempts > 50% in transgender men,3 both estimates being several-fold higher than for cisgender youth.
While it is important to investigate and document inequities in rates of suicide or self-harm between cisgender and TNBY, it is particularly important to appreciate this in the inpatient setting, as youth who have attempted suicide or self-harm frequently find themselves acutely hospitalized.7 Many inpatient pediatric providers have not been trained to provide gender-affirming care, and this has only been introduced as a topic in medical school and residency in recent years.8 Furthermore, many TNBY may avoid disclosing their gender identity when hospitalized due to fears of discrimination.9
Despite the high rates of discrimination in healthcare experiences, little is known about the patterns of and reasons for hospitalizations among TNBY.2–4 As prior studies primarily involved surveys of students, the demographics and prevalence of TNBY among hospitalized pediatric patients is unknown. However, a key limitation in the field is the lack of comprehensive measures to identify gender minority populations, precluding accurate epidemiologic studies. A related unanswered question is the prevalence of hospitalized TNBY with gender dysphoria, a condition related to the stress associated with incongruence between assigned sex at birth and gender identity. Gender dysphoria is more reliably coded, and may serve as a lower-limit estimate for TNBY youth. Therefore, we aimed to 1) describe the prevalence of TNBY with gender dysphoria among hospitalized patients in a large, nationally representative database of pediatric inpatients in the United States, 2) describe the sociodemographic characteristics of inpatient TNBY with gender dysphoria, and 3) explore the association between a gender dysphoria diagnosis and admission for suicidality or self-harm.
METHODS
Study design:
This was a population-based, serial cross-sectional study using data from the 2016 and 2019 releases of the Kids’ Inpatient Database (KID). The Children’s Hospital of Philadelphia Institutional Review Board determined that ethical approval and IRB oversight were not required as this study analyzed de-identified, publicly available data.
Setting and Participants:
The 2016 submission of KID was comprised of 3,117,413 patients admitted to 4,200 hospitals in the United States of America, accounting for 80% of pediatric discharges and included 47 states. The 2019 release includes 3,087,884 admissions in 48 states and the District of Columbia. Both general and specialty hospitals are included. Patients aged 6 to 21 years (not yet had 21st birthday) were included. Patients transferred to a different hospital were excluded so as not to be counted twice. A subset of TNBY with gender dysphoria were identified using ICD-10 codes consistent with gender dysphoria (F64·0 Transsexualism; F64·1 Gender identity disorder in adolescence and adulthood; F64·2 Gender identity disorder of childhood; F64·8 Other gender identity disorders; F64·9 Gender identity disorder, unspecified). These labels are pathologizing and no longer appropriate; however, this was the best way to identify TNBY in KID prior to the updated “gender dysphoria” terminology. We also acknowledge that this approach will not capture all gender diversity, with or without dysphoria, and that some of these categories do not necessarily imply dysphoria. Furthermore, we note that ICD-10 coding is an imperfect method to determine gender dysphoria and is highly dependent on provider documentation. Subjects with one of these codes constituted the primary exposure treated as a binary variable.
Variables:
Suicidality was identified using ICD-10 codes (R45·851 suicidal ideation and T14·91 Suicide attempt). Self-harm was identified using one of 355 self-harm codes (https://www.hcup-us.ahrq.gov/reports/ataglance/HCUPanalysisPedEDVisitsSuicide.pdf). Suicidality (primary outcome), self-harm, and suicidality and/or self-harm were treated as binary outcomes.
Demographic variables included age, recorded sex, payer group, median household income for patient’s ZIP Code (subdivided by KID into quartiles), and a combined race/ethnicity category reported by KID as White, Black, Hispanic, Asian and Pacific Islander, Native American, and Other. Given reduced numbers, we grouped Asian and Pacific Islander, Native American, and Other as “Other” for reporting and analyses. Variables relating to the patient admission included weekday versus weekend admission, whether a patient was admitted through the emergency department (ED), elective vs non-elective admission, and quarter of the year in which the admission occurred. Variables relating to hospital-level characteristics included hospital location (rural or urban), hospital size (small, medium, or large), teaching hospital status and hospital control (government or private). Hospital location was subdivided into US census regions (Northeast, Midwest, South, and West).
The dataset was examined for missing data. Race/ethnicity had the highest missingness (7%), with remaining variables < 2%. Because of the conceptual problems with imputing race based on the representativeness of the remaining variables,10,11 analyses were conducted after combining missing with “Other” Race and analyzing complete cases.
Statistical Analysis
Analyses were done using Stata version 17 (StataCorp, College Station, Texas, USA). The characteristics of the 2016 and 2019 cohorts were stratified according to presence or absence of gender dysphoria codes and compared using Χ2 tests. Prevalence of a gender dysphoria-related diagnosis was calculated. Unadjusted associations with suicidality, self-harm, or both was performed using modified Poisson regression to estimate prevalence ratios (PR).12,13 Multivariable Poisson regression models were constructed, with potential confounders selected using a causal framework via directed acyclic graphs (Supplementary Figure 1). The final models adjusted for individual-level (recorded sex, age, race/ethnicity, insurance type, median household income for patient’s ZIP Code), admission-level (elective admission, admitted from emergency department, weekend admission, admission quarter), and hospital-level (bed size, region, hospital ownership, rurality/teaching status) confounders. We used the robust (sandwich) estimator of variance and, as suggested by KID, regressions were performed with sampling weights applied in order to generate nationally representative estimates. We purposefully included race/ethnicity as a confounder as our conceptual model suggested that exposure to systemic racism could plausibly impact the likelihood of receiving a gender dysphoria diagnosis and be associated with suicidality.14,15 As both gender dysphoria and suicidality and self-harm are more commonly reported in teenagers, this analysis was repeated stratifying the cohorts according to age (6 to 12 years and 13 to 21 years). As a sensitivity analysis to assess the dependence of our conclusions on the inclusion of race/ethnicity in our models, we repeated the multivariable analyses after excluding race/ethnicity as a confounder. In a final post-hoc analysis, in order to assess whether geographic region modified the association between gender dysphoria and suicidality, we repeated the main analysis stratifying according to geographic region. Given a prevalence of gender dysphoria diagnoses of ~0.2% and of suicide of ~5%, we would have 90% power at α = 0.05 to detect a PR of at least 1.30.
Role of the Funding Source
Study sponsors had no role in study design, data collection, analysis, interpretation of data, writing the manuscript, or submission for publication.
RESULTS
Description of the Cohorts
There were 3,117,413 discharges in KID 2016, and 3,087,884 in 2019 (Supplementary Figure 2). Of those excluded due to age < 6 years, 6 subjects from 2016 KID had gender dysphoria codes and none < 6 years of age in either dataset had suicidality or self-harm codes. Of the remaining 1,090,544 eligible subjects between 6 and 21 years of age in 2016, 1755 had gender dysphoria codes (Table 1), for a prevalence of 0·161% (161 per 100,000 admissions of ages 6 to 21 years). When restricting the cohort to those between 13 and 21 years, the prevalence was 0·204% (204 per 100,000 adolescent admissions) in 2016. Of the 1,026,752 eligible subjects in 2019, 4872 had gender dysphoria codes, for a prevalence of 0·475% (475 per 100,000 admissions) for ages 6 to 21 years and 0·585% (585 per 100,000 admissions) for ages 13 to 21 years. In both 2016 and 2019, subjects with gender dysphoria were more likely to have a primary psychiatric diagnosis (Supplementary Table 1), were more likely to be White race, privately insured, and from higher income ZIP codes (Table 1), and less likely to be from the South or from rural hospitals (Supplementary Table 2) than subjects without gender dysphoria (all p < 0·001).
Table 1:
Demographic characteristics of the cohort stratified by presence or absence of gender dysphoria codes. All data are presented as N (%).
| 2016 | 2019 | |||||
|---|---|---|---|---|---|---|
|
| ||||||
| Variable | All | No gender dysphoria codes | Gender dysphoria codes | All | No gender dysphoria codes | Gender dysphoria codes |
|
| ||||||
| N | 1,090,544 | 1,088,789 | 1755 | 1,026,752 | 1,021,880 | 4872 |
|
| ||||||
| Sex | ||||||
| Male | 407,424 (37) | 406,894 (37) | 530 (30) | 498,467 (49) | 497,264 (49) | 1203 (25) |
| Female | 682,846 (63) | 681,641 (63) | 1205 (69) | 527,185 (51) | 523,594 (51) | 3591 (73) |
| Missing | 274 (0) | 254 (0) | 20 (1) | 1100 (0) | 1022 (0) | 78 (2) |
|
| ||||||
| Age | ||||||
| 6-12 | 281,550 (26) | 281,445 (26) | 105 (6) | 267,674 (26) | 267,240 (26) | 434 (9) |
| 13-17 | 344,900 (32) | 343,776 (32) | 1124 (64) | 335,337 (33) | 332,545 (33) | 2792 (57) |
| 18-20 | 464,094 (43) | 463,568 (43) | 526 (30) | 423,741 (41) | 422,095 (41) | 1646 (34) |
|
| ||||||
| ICD10 codes | ||||||
| Suicide | 55,986 (5) | 55,351 (5) | 635 (36) | 39,017 (4) | 38,831 (4) | 2680 (55) |
| Self-harm | 14,217 (1) | 13,996 (1) | 221 (13) | 12,321 (1) | 12,263 (1) | 731 (15) |
| Either | 67,227 (6) | 66,510 (6) | 717 (41) | 48,257 (5) | 47,006 (5) | 3201 (66) |
|
| ||||||
| Race | ||||||
| White | 505,228 (46) | 504,067 (46) | 1161 (66) | 468,199 (46) | 464,730 (45) | 3469 (71) |
| Black | 197,821 (18) | 197,704 (18) | 117 (7) | 170,441 (17) | 170,090 (17) | 351 (7) |
| Hispanic | 229,823 (21) | 229,658 (21) | 165 (9) | 202,270 (20) | 201,778 (20) | 492 (10) |
| Other | 81,235 (7) | 81,131 (8) | 104 (6) | 111,916 (11) | 111,580 (11) | 336 (7) |
| Missing | 76,437 (7) | 76,229 (7) | 208 (12) | 73,926 (7) | 73,702 (7) | 224 (5) |
|
| ||||||
| Payer group | ||||||
| Public | 581,188 (53) | 580,537 (53) | 651 (37) | 524,670 (51) | 522,794 (51) | 1876 (39) |
| Private | 424,732 (39) | 423,741 (39) | 991 (57) | 421,995 (41) | 419,350 (41) | 2645 (54) |
| Other | 83,131 (8) | 83,021 (8) | 110 (6) | 78,399 (8) | 78,058 (8) | 341 (7) |
| Missing | 1493 (0·1) | 1490 (0·1) | 3 (0·2) | 1688 (0·2) | 1678 (0·2) | 10 (0·2) |
|
| ||||||
| Median income by ZIP ($10,000) | ||||||
| < 43 | 363,371 (33) | 363,066 (33) | 305 (17) | 313,159 (30) | 312,141 (31) | 1018 (21) |
| 43 to 54 | 268,573 (25) | 268,185 (25) | 388 (22) | 250,527 (24) | 249,382 (24) | 1145 (24) |
| 54 to 71 | 245,449 (23) | 244,962 (23) | 487 (28) | 247,447 (24) | 246,078 (24) | 1369 (28) |
| > 71 | 196,966 (18) | 196,420 (18) | 546 (31) | 204,324 (20) | 203,033 (20) | 1291 (26) |
| Missing | 16,185 (2) | 16,156 (2) | 29 (2) | 11,294 (1) | 11,245 (1) | 49 (1) |
All comparisons between those with and with gender dysphoria codes were significant at p < 0·001 by Χ2 testing. The 2016 and 2019 datasets were analyzed separately.
Association with Suicidality and/or Self-Harm
Subjects coded with a diagnosis of gender dysphoria had greater prevalence of coexisting suicidality both in 2016 (36% versus 5%; unadjusted PR 7·19, 95% CI 6·75 to 7·66) and 2019 (55% versus 4%; unadjusted PR 5·45, 95% CI 5·30 to 5·60) relative to youths without gender dysphoria codes (Tables 1 and 2). This association persisted after adjusting for confounders in both 2016 (adjusted PR 5·02, 95% CI 4·67 to 5·41) and 2019 (adjusted PR 4·14, 95% CI 4·02 to 4·28) datasets (Table 2; Supplementary Table 3). Increased PR for subjects with gender dysphoria, relative to those without, were also seen for self-harm and when combining suicidality and self-harm in both 2016 and in 2019 (Table 2). Similar results were seen in a sensitivity analysis restricted to subjects between 6 and 12 years of age (Table 3), as well as subjects between 13 and 21 years (Table 4). Conclusions were unchanged and effect sizes were similar when removing race/ethnicity from the multivariable models (Supplementary Table 4). In an unplanned post-hoc analysis, geographic region modified the association between gender dysphoria and suicide in both 2016 and 2019 datasets (both interaction p < 0.001), with the strongest association between gender dysphoria and suicide in the South and West (Supplementary Table 5).
Table 2:
Crude and adjusted associations between gender dysphoria codes and suicidality and/or self-harm in subjects aged 6 to 21 years.
| Model | 2016 | 2019 | ||
|---|---|---|---|---|
|
| ||||
| Prevalence ratio | 95% confidence intervals | Prevalence ratio | 95% confidence intervals | |
|
| ||||
| Association with suicidality | ||||
| Unadjusted | 7·19 | 6·75 to 7·66 | 5·45 | 5·30 to 5·60 |
| Adjusteda | 5·02 | 4·67 to 5·41 | 4·14 | 4·02 to 4·28 |
|
| ||||
| Association with self-harm | ||||
| Unadjusted | 4·80 | 3·97 to 5·80 | 5·16 | 4·73 to 5·62 |
| Adjusteda | 3·64 | 2·99 to 4·44 | 3·75 | 3·40 to 4·14 |
|
| ||||
| Association with suicidality or self-harm | ||||
| Unadjusted | 6·75 | 6·37 to 7·15 | 5·34 | 5·22 to 5·47 |
| Adjusteda | 4·77 | 4·46 to 5·09 | 4·04 | 3·94 to 4·17 |
Adjusted for recorded sex, age, race/ethnicity, insurance type, median household income for patient’s ZIP Code, elective admission, admitted from emergency department, admitted at weekend, admission quarter, hospital bed size, hospital region, hospital control and hospital teaching status. Regressions were performed with sampling weights applied.
Table 3:
Crude and adjusted associations between gender dysphoria codes and suicidality and/or self-harm in subjects aged 6 to 12 years
| Model | 2016 | 2019 | ||
|---|---|---|---|---|
|
| ||||
| Prevalence ratio | 95% confidence intervals | Prevalence ratio | 95% confidence intervals | |
|
| ||||
| Association with suicidality | ||||
| Unadjusted | 12·17 | 9·16 to 16·18 | 10·17 | 9·45 to 10·94 |
| Adjusteda | 4·88 | 3·58 to 6·65 | 3·90 | 3·49 to 4·35 |
|
| ||||
| Association with self-harm | ||||
| Unadjusted | 18·03 | 6·86 to 47·43 | 13·29 | 9·49 to 18·60 |
| Adjusteda | 3·82 | 1·23 to 11·87 | 3·40 | 2·40 to 5·11 |
|
| ||||
| Association with suicidality or self-harm | ||||
| Unadjusted | 12·74 | 9·81 to 16·56 | 10·37 | 9·72 to 11·07 |
| Adjusteda | 4·90 | 3·63 to 6·61 | 3·86 | 3·48 to 4·28 |
Adjusted for recorded sex, age, race/ethnicity, insurance type, median household income for patient’s ZIP Code, elective admission, admitted from emergency department, admitted at weekend, admission quarter, hospital bed size, hospital region, hospital control and hospital teaching status. Regressions were performed with sampling weights applied.
Table 4:
Crude and adjusted associations between gender dysphoria codes and suicidality and/or self-harm in subjects aged 13 to 21 years.
| Model | 2016 | 2019 | ||
|---|---|---|---|---|
|
| ||||
| Prevalence ratio | 95% confidence intervals | Prevalence ratio | 95% confidence intervals | |
|
| ||||
| Association with suicidality | ||||
| Unadjusted | 6·19 | 5·80 to 6·60 | 4·71 | 4·58 to 4·85 |
| Adjusteda | 3·80 | 3·52 to 4·10 | 3·23 | 3·12 to 3·34 |
|
| ||||
| Association with self-harm | ||||
| Unadjusted | 3·80 | 3·13 to 4·61 | 4·21 | 3·85 to 4·61 |
| Adjusteda | 2·68 | 2·17 to 3·30 | 2·90 | 2·62 to 3·21 |
|
| ||||
| Association with suicidality or self-harm | ||||
| Unadjusted | 5·71 | 5·39 to 6·06 | 4·58 | 4·46 to 4·69 |
| Adjusteda | 3·59 | 3·35 to 3·85 | 3·15 | 3·06 to 3·25 |
Adjusted for recorded sex, age, race/ethnicity, insurance type, median household income for patient’s ZIP Code, elective admission, admitted from emergency department, admitted at weekend, admission quarter, hospital bed size, hospital region, hospital control and hospital teaching status. Regressions were performed with sampling weights applied.
DISCUSSION
Using the two most recent releases of a large representative national database, we identified that approximately 0·16% of hospitalized children between 6 and 21 years, and 0·20% of hospitalized adolescents between 13 and 21 years, had a diagnosis of gender dysphoria in 2016. Prevalence in 2019 was nearly triple that of 2016, likely reflecting changes in attitudes and approaches to gender diversity in and out of hospital settings. Having gender dysphoria-related diagnostic codes were highly associated with both suicidality and self-harm. Actual numbers of hospitalized TNBY (with or without dysphoria) are likely larger, given that some TNBY do not share their gender identity, may not have gender dysphoria, and among those that do may not be assigned a corresponding administrative code. Furthermore, some of our chosen codes (e.g., F64·0 Transsexualism) do not imply gender dysphoria, upwardly biasing the estimated prevalence of dysphoria. Population-based reports such as the Youth Risk Behavior Survey suggest that the prevalence of adolescents identifying as TNBY is approximately 2%,16 and potentially higher in more recent reports.17 The fewer non-White, publicly insured, and low-income youths with gender dysphoria codes in our cohort suggest potential differences in the context and consequences of sharing gender identity with healthcare providers or inequitable access to gender-affirming care among racial minority and economically disadvantaged youth.18 Under-representation of subjects with gender dysphoria from the South of the United States, relative to other regions, suggests geographic inequities, which may correspond to inequities in likeliness of disclosing transgender or nonbinary identity, uneven access to gender-affirming providers, or inequities in systemic assessment for gender identity.
This is the first nationally representative assessment of the prevalence of gender dysphoria diagnoses among pediatric hospitalizations, demonstrating a high prevalence of suicidality and self-harm. In 2018, the American Academy of Pediatrics recognized that TNBY youths are an increasing and underserved population, at risk for significant health disparities and discrimination. This prompted a policy statement aimed toward outpatient providers and primary caregivers to improve health and well-being of these children.19 Less attention has been paid to the care of TNBY in the inpatient hospital setting. However, since a substantial portion of inpatient pediatric encounters with TNBY, particularly TNBY with gender dysphoria, involve suicidality and/or self-harm, inpatient practitioners are in a unique position to provide appropriate care in an affirming environment, and increase access to critical gender-affirming services and treatment. Our results also support providing gender-affirming care to all patients, particularly if hospitalized for suicidality and/or self-harm, as discrimination due to gender identity may be a contributor to the hospitalization, irrespective of recorded gender dysphoria.
TNBY have increased risk of attempted suicide or self-harm relative to cis-gender youth in prior studies,1–6,20,21 most of which were survey-based from the United States and Canada. In one of the largest such studies (over 120,000 adolescents), female-to-male transgender adolescents reported the highest rate of attempted suicide (51%), followed by nonbinary (42%), male-to-female transgender (30%), questioning (28%), cis-gender female (18%), and cis-gender male (10%).3 Our study confirms an association between gender dysphoria diagnoses and increased prevalence of suicidality. However, as administrative coding may not reliably differentiate between assigned sex and gender, we could not verify whether specific categories of TNBY (e.g., transgender men) with gender dysphoria drove this association. Disaggregating different types of gender diversity will be an important focus for future studies, and improved recording of gender identity in administrative databases would substantially facilitate understanding how different forms of stigma may have distinct impacts on the physical and mental health of youth.
It is important to note that just as when using race as a proxy for exposure to systemic racism,14 we are using gender dysphoria coding as a proxy for exposure to systemic transphobia, which is the more plausible causal factor for increased suicidality or self-harm. Data suggest that lack of access to gender-affirming care, discrimination, and bullying are the main drivers of high rates of anxiety, depression and suicidality in this population.1,2,20,22,23 Specific to the medical setting, a recent systematic review encompassing 844 TNBY from 17 countries documented pervasive discrimination in encounters with the health care system, with feelings of vulnerability and uncertainty surrounding decision-making.9 These risks can be mitigated by both increasing access to health care generally that is gender affirming and by increasing access to specific medical treatments such as gender-affirming hormone therapy and psychosocial support.24,25 Thus, inpatient pediatric providers are in a position to take affirmative steps towards reducing discrimination by bringing awareness, respect, and appropriate affirming practices to these encounters. For example, progression of diagnostic language from previously-used pathologizing terminology (such as “gender identity disorder”) to likely use of “gender incongruence” in ICD-11 is a welcome change that will more accurately reflect a person’s identity. Some specific steps in the inpatient setting include staff name badges with their pronouns (modeling acceptance of TNBY), having the electronic health record consistently and automatically use patients’ stated names and pronouns in charts and documents, and avoiding use of incorrect names (i.e., dead-naming) when performing two-person identification checks.
We found evidence of socioeconomic, racial, and geographic inequities in the prevalence of gender dysphoria coding in KID. Race/ethnicity had the highest missingness in our dataset, but as this was not our primary exposure, we categorized those with missing race/ethnicity coding as Other. This approach would not affect the estimates for gender dysphoria (our primary exposure) reported in multivariable analysis, and our findings were robust to removal of race/ethnicity from the models. A prior analysis of over 5000 transgender adults noted both individual (racial/ethnic minority, lower income) and geographic (South and West United States) factors associated with medical care refusal by providers.26 When adjusting for state-level factors, the percentage of the state population voting Republican was positively associated with care refusal. In our study, we found effect modification by region, with higher prevalence of suicidality for subjects with gender dysphoria than those without in the South and West, confirming the significance of regionality in future studies. At the time of this manuscript, it is notable that multiple state legislatures have attempted to ban gender-affirming care in pediatrics.27 Our study reinforces that social determinants of health is a broad domain affected by region-specific culture and politics which could impact rates of adolescent suicide and self-harm.28
Our study has limitations. KID cannot identify repeat admissions, and so it is possible some subjects are represented more than once. Additionally, administrative coding is an imprecise method to identify primary cause for admission. As KID is limited to hospitalized subjects, our results cannot mitigate the bias related to difficulties in accessing healthcare. However, a larger limitation is the reliance on a formal diagnosis of gender dysphoria. The prevalence of gender dysphoria among TNBY is difficult to estimate as there are not reliable ways to determine gender identity in administrative health databases. Indeed, the demographic and regional imbalances reported here suggest potential misclassification, with reporting or coding of gender dysphoria differing according to race, income, and geography. Relatedly, it is difficult to operationally define TNBY with gender dysphoria in a manner that is inclusive of all identities, and some of the ICD-10 codes we used capture gender diversity without implying dysphoria and thus could inflate the prevalence of dysphoria. Therefore, because only a subset of hospitalized TNBY with gender dysphoria would be captured via coding, the precise prevalence of suicidality among TNBY with gender dysphoria, and the proportion of total hospitalizations for suicidality and/or self-harm attributable to gender dysphoria, is likely biased. Importantly, if practitioners are more likely to code gender dysphoria for admissions related to self-harm than for other medical or surgical admissions, the association between gender dysphoria coding and suicidality would be upwardly biased. Similarly, higher risk subjects are potentially more likely to be hospitalized, which would also upwardly bias our reported PRs. Despite this likely misclassification, the strong association between gender dysphoria codes and suicidality and/or self-harm in our study is consistent with previously reported high rates of increased suicidal ideation among TNBY.3,4,20,23
Despite these limitations, our study does provide a likely lower-bound for the prevalence of hospitalized TNBY with gender dysphoria, as well as evidence of demographic and geographic disparities in the United States with coding gender dysphoria. Thus, our results are most directly relevant to inpatient pediatric providers to provide a framework for assessing their own personal and institutional practices surrounding gender identity. The association with suicidality and/or self-harm, while potentially inflated in magnitude, likely represents a real association consistent with community survey data.
CONCLUSION
Using the 2016 and 2019 releases of the nationally representative KID database, we identified a cohort of TNBY using gender dysphoria-related diagnostic codes. Having a clinical diagnosis of gender dysphoria was associated with hospitalization for suicidality and self-harm. The fewer non-White, publicly insured, and low-income youths with gender dysphoria suggest inequities in access to gender-affirming care among racial minority and economically-disadvantaged youth. Regional disparities raise concerns about discrimination within and outside of healthcare settings, point to potential inequities in clinician preparedness to provide gender-affirming care, and raise concerns about discrimination preventing youth from self-identifying. Structural and provider-level interventions are needed to reduce discrimination and improve access to gender-affirming care to prevent adverse outcomes. These interventions should focus on youth at-risk for suicide and self-harm and can include linkage to gender-affirming care and support services, targeted follow-up for those previously hospitalized for self-harm, ensuring gender-affirming competencies among staff, expanding gender-affirming outpatient clinics, and addressing stressors related to gender identity from multidisciplinary perspectives. In the inpatient setting, these interventions can include use of affirmed names and pronouns and improving recognition of gender dysphoria as a possible contributor to a hospitalization for self-harm. Finally, inpatient pediatric providers are in a unique position to provide TNBY with gender dysphoria an affirmative environment generally, and to link them directly to supportive care and services when hospitalized.
Supplementary Material
RESEARCH IN CONTEXT.
Evidence Before This Study:
We explored Pubmed without date limitations using historical and contemporary terms related to gender identity (including transgender, non-binary, gender identity disorder, and gender dysphoria) and suicidality. Evidence was comprised primarily of survey studies and self-identified samples. Overall, the data suggested that transgender and nonbinary youth (TNBY) were placed at higher risk for suicidality and self-harm than cisgender due to discrimination and stigma, but little was known about demographics and patterns of TNBY with gender dysphoria admitted into a hospital for suicide attempts or self-harm.
Added Value of This Study:
Using the nationally representative 2016 and 2019 Kids’ Inpatient Database from the United States, we report prevalence of gender dysphoria in pediatric admissions (161 per 100,000 admissions between 6 and 21 years in 2016; 475 per 100,000 in 2019) using administrative codes, with evidence of differential coding according to demographics and geography. Gender dysphoria coding was strongly associated with suicide or self-harm diagnoses.
Implications of All Available Evidence:
TNBY subjects with gender dysphoria identified using administrative codes are frequently hospitalized for reasons related to suicidality or self-harm. The inpatient setting offers a unique opportunity to provide gender-affirming care to this at-risk population. Evidence of disparate coding may reflect a stigma of transphobia manifesting as differential comfort with sharing experiences of gender dysphoria or with administrative coding, rather than actual differences in TNBY or gender dysphoria prevalence according to demographics or region. Our study will not capture all hospitalized TNBY youth, and administrative coding has limitations as a method to determine gender dysphoria, but our results provide an approximation for prevalence of hospitalized TNBY with gender dysphoria.
ACKNOWLEDGEMENTS
We would like to thank Ross Perfetti for their expertise and assistance in writing this manuscript.
FUNDING
Dr. Yehya is supported by National Institutes of Health (NIH) grant number K23-HL136688.
DECLARATION OF INTERESTS
Dr. Yehya reports funding from Pfizer outside the scope of this manuscript. The remaining authors declare no conflicts of interest.
DATA SHARING
This paper uses publicly accessible data from Healthcare Cost and Utilization Project (HCUP) Kids’ Inpatient Database (KID).
REFERENCES
- 1.Jackman KB, Dolezal C, Levin B, Honig JC, Bockting WO. Stigma, gender dysphoria, and nonsuicidal self-injury in a community sample of transgender individuals. Psychiatry Res 2018; 269: 602–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Eisenberg ME, Gower AL, McMorris BJ, Rider GN, Shea G, Coleman E. Risk and Protective Factors in the Lives of Transgender/Gender Nonconforming Adolescents. J Adolesc Health 2017; 61(4): 521–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Toomey RB, Syvertsen AK, Shramko M. Transgender Adolescent Suicide Behavior. Pediatrics 2018; 142(4). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Thoma BC, Salk RH, Choukas-Bradley S, Goldstein TR, Levine MD, Marshal MP. Suicidality Disparities Between Transgender and Cisgender Adolescents. Pediatrics 2019; 144(5). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Sorbara JC, Chiniara LN, Thompson S, Palmert MR. Mental Health and Timing of Gender-Affirming Care. Pediatrics 2020; 146(4). [DOI] [PubMed] [Google Scholar]
- 6.Bauer GR, Pacaud D, Couch R, et al. Transgender Youth Referred to Clinics for Gender-Affirming Medical Care in Canada. Pediatrics 2021; 148(5). [DOI] [PubMed] [Google Scholar]
- 7.Plemmons G, Hall M, Doupnik S, et al. Hospitalization for Suicide Ideation or Attempt: 2008–2015. Pediatrics 2018; 141(6). [DOI] [PubMed] [Google Scholar]
- 8.Dowshen N, Nguyen GT, Gilbert K, Feiler A, Margo KL. Improving transgender health education for future doctors. Am J Public Health 2014; 104(7): e5–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Chong LSH, Kerklaan J, Clarke S, et al. Experiences and Perspectives of Transgender Youths in Accessing Health Care: A Systematic Review. JAMA Pediatr 2021; 175(11): 1159–73. [DOI] [PubMed] [Google Scholar]
- 10.Elliott MN, Fremont A, Morrison PA, Pantoja P, Lurie N. A new method for estimating race/ethnicity and associated disparities where administrative records lack self-reported race/ethnicity. Health Serv Res 2008; 43(5 Pt 1): 1722–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Labgold K, Hamid S, Shah S, et al. Estimating the Unknown: Greater Racial and Ethnic Disparities in COVID-19 Burden After Accounting for Missing Race and Ethnicity Data. Epidemiology 2021; 32(2): 157–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Zou G A modified poisson regression approach to prospective studies with binary data. American journal of epidemiology 2004; 159(7): 702–6. [DOI] [PubMed] [Google Scholar]
- 13.Petersen MR, Deddens JA. A comparison of two methods for estimating prevalence ratios. BMC Med Res Methodol 2008; 8: 9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Lett E, Asabor E, Beltran S, Cannon AM, Arah OA. Conceptualizing, Contextualizing, and Operationalizing Race in Quantitative Health Sciences Research. Ann Fam Med 2022; 20(2): 157–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Flanagin A, Frey T, Christiansen SL, Committee AMAMoS. Updated Guidance on the Reporting of Race and Ethnicity in Medical and Science Journals. JAMA 2021; 326(7): 621–7. [DOI] [PubMed] [Google Scholar]
- 16.Johns MM, Lowry R, Andrzejewski J, et al. Transgender Identity and Experiences of Violence Victimization, Substance Use, Suicide Risk, and Sexual Risk Behaviors Among High School Students - 19 States and Large Urban School Districts, 2017. MMWR Morb Mortal Wkly Rep 2019; 68(3): 67–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kidd KM, Sequeira GM, Douglas C, et al. Prevalence of Gender-Diverse Youth in an Urban School District. Pediatrics 2021; 147(6). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lett E, Abrams MP, Gold A, Fullerton FA, Everhart A. Ethnoracial inequities in access to gender-affirming mental health care and psychological distress among transgender adults. Soc Psychiatry Psychiatr Epidemiol 2022; 57(5): 963–71. [DOI] [PubMed] [Google Scholar]
- 19.Rafferty J, Committee On Psychosocial Aspects Of C, Family H, et al. Ensuring Comprehensive Care and Support for Transgender and Gender-Diverse Children and Adolescents. Pediatrics 2018; 142(4). [DOI] [PubMed] [Google Scholar]
- 20.Connolly MD, Zervos MJ, Barone CJ 2nd, Johnson CC, Joseph CL. The Mental Health of Transgender Youth: Advances in Understanding. J Adolesc Health 2016; 59(5): 489–95. [DOI] [PubMed] [Google Scholar]
- 21.Heino E, Frojd S, Marttunen M, Kaltiala R. Transgender identity is associated with severe suicidal ideation among Finnish adolescents. Int J Adolesc Med Health 2021. [DOI] [PubMed] [Google Scholar]
- 22.Clements-Nolle K, Marx R, Katz M. Attempted suicide among transgender persons: The influence of gender-based discrimination and victimization. J Homosex 2006; 51(3): 53–69. [DOI] [PubMed] [Google Scholar]
- 23.Almeida J, Johnson RM, Corliss HL, Molnar BE, Azrael D. Emotional distress among LGBT youth: the influence of perceived discrimination based on sexual orientation. J Youth Adolesc 2009; 38(7): 1001–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Turban JL, King D, Carswell JM, Keuroghlian AS. Pubertal Suppression for Transgender Youth and Risk of Suicidal Ideation. Pediatrics 2020; 145(2). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Branstrom R, Pachankis JE. Reduction in Mental Health Treatment Utilization Among Transgender Individuals After Gender-Affirming Surgeries: A Total Population Study. Am J Psychiatry 2020; 177(8): 727–34. [DOI] [PubMed] [Google Scholar]
- 26.White Hughto JM, Murchison GR, Clark K, Pachankis JE, Reisner SL. Geographic and Individual Differences in Healthcare Access for U.S. Transgender Adults: A Multilevel Analysis. LGBT Health 2016; 3(6): 424–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Gordon CM. Caught in the Middle: The Care of Transgender Youth in Texas. Pediatrics 2022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Montoya-Williams D, Fuentes-Afflick E. Political Determinants of Population Health. JAMA Netw Open 2019; 2(7): e197063. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
This paper uses publicly accessible data from Healthcare Cost and Utilization Project (HCUP) Kids’ Inpatient Database (KID).