Prevalence of gender dysphoria and suicidality and self-harm in a national database of paediatric inpatients in the USA: a population-based, serial cross-sectional study

Background Transgender and non-binary young people experience discrimination that has been linked to suicidal ideation and self-harm, but few studies have examined this association systematically. We aimed to study the association between gender dysphoria-related diagnostic coding and hospital admission for suicidality or self-harm in a large representative pediatric inpatient database from the USA. Methods Using the 2016 and 2019 Kids' Inpatient Database (KID), we identified transgender and non-binary young people (aged 6-20 years) with International Classification of Diseases (ICD)-10 codes related to gender dysphoria. We identified suicidal ideation using explicit suicidality ICD-10 codes, and self-harm using 355 self-harm ICD-10 codes. Prevalence of suicidality (primary outcome), self-harm, and the combination of suicidality and self-harm (secondary outcomes) was compared between young people with and without gender dysphoria-related codes. Univariable and multivariable regression was used to test for an association between gender dysphoria and suicidality, self-harm, or suicidality and self-harm combined. Findings 1 090 544 individuals were included from KID 2016 and 1 026 752 from KID 2019. Gender dysphoria-related diagnoses were prevalent in 161 per 100 000 hospital admissions in KID 2016 and 475 per 100 000 hospital admissions in KID 2019. In KID 2016 and KID 2019, among individuals who were White, privately insured, or from higher median income ZIP code areas, proportionately more had gender dysphoria-related codes. Prevalence of suicidality was greater in individuals with gender dysphoria-related codes than in individuals without gender dysphoria-related codes in KID 2016 (635 [36%] of 1755 individuals with gender dysphoria-related codes vs 55 351 [5%] of 1 088 789 individuals without gender dysphoria-related codes; unadjusted prevalence ratio [PR] 7 center dot 19 [95% CI 6 center dot 75-7 center dot 66]) and KID 2019 (2680 [55%] of 4872 individuals with gender dysphoria-related codes vs 38 831 [4%] of 1 021 880 individuals without gender dysphoria-related codes; unadjusted PR 5 center dot 45 [5 center dot 30-5 center dot 60]). This association persisted in multivariable modelling adjusting for confounders in KID 2016 (adjusted PR 5 center dot 02 [95% CI 4 center dot 67-5 center dot 41]) and KID 2019 (4 center dot 14 [4 center dot 02-4 center dot 28]). Increased unadjusted and adjusted PRs for individuals with gender dysphoriarelated codes, relative to those without, were also evident for self-harm and when combining suicidality and self-harm in both the 2016 and 2019 datasets. Interpretation In a large representative national sample, transgender and non-binary young people with gender dysphoria-related diagnoses were frequently admitted to hospital for suicidality or self-harm. The lower rates of gender-dysphoria-related codes among young people who were non-White, publicly insured, and from low-income households suggest that underlying inequities might shape the identification and management of gender dysphoria. Structural and health-care provider-level interventions are needed to reduce discrimination and expand genderaffirming competencies to prevent adverse outcomes for hospitalised transgender and non-binary young people with gender dysphoria.