Pediatric and Adolescent Obstetric and Gynecologic Encounters in US Emergency Departments: A Cross-Sectional Study

Overall, there were 518,244 visits and most were discharged from the hospital (89%), while a smaller percentage were admitted (8%) or transferred (2%) ( Table 1 ). Across the population, the number of presenting patients increased with increasing age (e.g., 1% in infants, 4% in 5 to 9 years, 78% in 15 to 18 years). Almost half (46%) of these patient visits occurred at private, not for profit hospitals, and most occurred at teaching hospitals (67%). Medicaid was the primary payor (66%). 72% of the patient population lived in a zip code where the median household income was less than $35,000. In Table 2 , clinical categories are organized by prevalence. Abnormal uterine bleeding, pregnancy, and vulvovaginal disorders were the most common presentations. The most common diagnoses varied by age group ( Table 3 ). The top diagnosis category for infants, toddlers, and school-age children was vulvar and vaginal disorders, for pubescent patients was physiologic/unspecified ovarian cysts and for adolescent patients was pregnancy. Female congenital anomalies were in the top diagnoses for infants and toddlers (12.3% and 4.44%). Sexual abuse and exploitation were in top diagnoses among toddlers and school aged children, while injury to the female genital tract was only a top diagnosis (2.13%) among school aged children. Given the range of admission and transfer by varying clinical diagnoses, we analyzed a common diagnosis with a low admission/transfer rate (vulvar/vaginal disorders) and two uncommon diagnoses associated with higher morbidity (ovarian torsion and ectopic pregnancy). Admission and transfer were more likely for patients with diagnoses for ectopic pregnancy and ovarian torsion compared to those without this diagnosis (ectopic: admission OR 1.96, CI 1.34 to 2.89 and transfer OR 3.37, CI 1.62 to 7.03; torsion: admission OR 6.29, CI 4.27 to 9.27 and transfer OR 4.82, CI 3.37 to 6.91). These increased rates were not seen with vulvar/vaginal complaints (admission OR 0.35, CI 0.30 to 0.41 and transfer OR 0.48, CI 0.41 to 0.57).

This is a retrospective cross-sectional survey design utilizing the Nationwide Emergency Department Sample (NEDS) from 2018. vii NEDS is part of the Healthcare Cost and Utilization Project (HCUP) and is a comprehensive database compiling state inpatient (SID) and emergency department databases (SEDD). The survey design is a stratified, single-stage cluster sample that is stratified into five strata. It includes patients who were treated and released from the ED and those admitted to the same presenting hospital. NEDS is a limited dataset not requiring IRB approval for use. Female patients from birth to 18 years old were included. Utilizing billing and coding data, the NEDS has a total of up to 35 diagnoses (International Classification of Diseases, or ICD codes) associated with each visit. All obstetric and gynecologic (OB/GYN) codes were placed into twenty clinically relevant OB/GYN categories ( Appendix A ). These categories encompass multiple different ICD codes related to OB/GYN care and were organized based on clinical applicability. For example, the treatment of bleeding in early pregnancy is very different than the treatment in the third trimester. Some categories have multiple ICD codes (e.g., pregnancy) versus just one (e.g., differences of sexual development derived from the ICD code “Indeterminate sex and pseudohermaphroditism”). See Appendix A for a complete list. Additional variables included ED disposition, hospital type, age at presentation, expected primary payor type, and median household income by patient zip code. For ED disposition, outcomes included treat and release, transfer (either to a short-term hospital or other facility), admitted as an inpatient to presenting hospital, home healthcare, left against medical advice, died in the ED, or unknown. Hospital type was defined into three major categories: (1) Metropolitan teaching, (2) Metropolitan non-teaching, and (3) Non-metropolitan, non-teaching. Non-metropolitan is defined by the Urban Influence Codes as counties with <49,000 residents not adjacent to a metropolitan county. 7 Thus, there are no defined non-metropolitan, teaching hospitals. Teaching hospital status is defined as a hospital with an approved residency program by the American Medical Association or membership in the Council of Teaching Hospitals. Missing data points for the above variables were categorized as unknown and analyzed independently. Primarily this is a descriptive analysis. Logistic regression was also utilized with probit models for multivariable analysis that did incorporate the hospital weights to account for sampling design. All analysis was performed utilizing Stata’s SVY function with multiple strata, cluster design (College Station, TX: Stata Press. StataCorp. 2019).

Little research has examined OB/GYN related ED presentations among pediatric and adolescent patients. Diagnoses with higher overall prevalence but lower admission rates include physiologic/unspecified ovarian cysts, vulvovaginal disorders, abnormal uterine bleeding, dysmenorrhea, and pregnancy. Diagnoses with low transfer rates were also notable for conditions that are rarely life threatening or emergent such as dysmenorrhea, breast disease, vulvovaginal disorders, pelvic inflammatory disease, and abnormal uterine bleeding (all <2% transfer rate). Many of these conditions can and should be treated in the outpatient setting rather than in the emergency department. Clear follow up guidelines and referrals to local, outpatient resources should be developed for these common, yet non-emergent conditions. There were several diagnosis categories with low overall prevalence but high admission rates, including differences of sexual development, gender identity disorder, injury to the female genital tract, ovarian torsion, and ovarian neoplasm. Given that these presentations are rare in this population, higher admission and transfer rates may indicate that these diagnoses require additional resources and interventions not always available in the emergency department. Diagnoses with high morbidity within this population are fortunately rare and include ectopic pregnancy (n= 218) and torsion (n= 656). These two diagnoses are associated with higher admission and transfer rates suggestive of both need for intervention as well as potential requirement for different level of care. There are a variety of providers that evaluate pediatric and adolescent patients in the emergency department after the initial evaluation by the emergency medicine provider. Depending on the emergency department, additional providers can include pediatrics, adolescent medicine, internal medicine, general surgery, pediatric surgery, and obstetrics and gynecology providers. At some children’s hospitals and academic centers, there are also specialized and potentially fellowship-trained pediatric and adolescent gynecologic providers. While educational efforts should always aim to focus on emergent conditions such as ovarian torsion and ectopic pregnancy, it is also important to address common presentations and diagnoses such as vulvovaginal disorders and abnormal uterine bleeding. There is also a potential role for telemedicine to engage specialist providers in hospitals without access to these subspecialties. Given the common prevalence of vulvovaginal disorders, it is critical that all providers caring for this population feel comfortable performing an initial vulvar or vaginal exam that is age appropriate. Vulvovaginal disorders are not typically life-threatening or emergent, ix and the most common ICD codes include vulvovaginal inflammation and diseases of the Bartholin gland. Abnormal uterine bleeding is another diagnosis where an exam is an important factor in diagnosis. In terms of emergency medicine guidance, there are few texts discussing female congenital abnormalities which is more common in younger patients as well as limited discussion regarding higher morbidity presentations within this population (e.g., torsion and ectopic). x There is some guidance regarding differential diagnoses by age group and vulvar/vaginal disorders. ix The strengths of this study include that it is the first known to date that utilizes national data to quantify obstetric and gynecologic emergency department visits within this population. By creating clinical categories specific to OB/GYN, we were able to evaluate more granular presentations than national data typically allows. Furthermore, we were able to give national estimates for emergency department visits throughout the United States and quantify rarer diagnoses such as ovarian neoplasms and injury to the female genital tract. Another perceived strength is the breadth of diagnoses included. Given the paucity of data on this subject, we decided to be overly inclusive rather than exclusive when creating our dataset. We utilized all 35 possible ICD-10 codes associated with each ED encounter rather than the primary code alone, to overcount rather than undercount OB/GYN related complaints; we wanted to include patients who may have had a secondary or tertiary OB/GYN complaint. Additional research should further identify emergency department outcomes related to specific gynecologic emergencies associated with significant morbidity such as ovarian torsion or ectopic pregnancy in this population. Quality improvement studies could focus on hospital specific protocols and referral patterns for varying gynecologic and obstetric presentations within this population.

Pediatric and adolescent patients are seen in the emergency department with presentations varying from pregnancy to vulvovaginal complaints to genital injury. Given that most pediatric and adolescent patients present to hospitals without access to subspecialty care, there is a potential role for the expansion of pediatric-specific gynecological emergency care education, development of clinical protocols specific to the most common and emergent presentations, creation of outpatient pathways for non-emergent presentations, use of telemedicine for subspecialty consults, and creation of interfacility transfer agreements that would likely benefit this population.

Limitations are mainly related to limitations in the data set. There are inevitable errors, differences, and biases in billing and coding practices that impact the reliability of this data, and this likely increases with the number of non-primary codes utilized. viii Additionally, we were limited in the variables that the NEDS database collects. NEDS neither includes state level data nor detailed information on patients other than the variables listed above. Furthermore, this study is limited in its ability to determine clinical outcomes and rather compared ED dispositions as a surrogate. There also is an inability to determine the severity of illness or presentation that likely greatly impacted the ultimate patient disposition.

Little research has evaluated obstetric and gynecologic presentations among pediatric and adolescent patients in U.S. emergency departments (ED). Despite increases in insurance coverage, emergency care for pediatric populations remains high, and EDs continue to be a key entry point to medical care. i , ii Furthermore, a significant portion of ED visits for pediatric and adolescent patients are at community hospitals. iii , ii Within pediatric and adolescent gynecology, common presentations include abnormal uterine bleeding, adnexal masses, pelvic pain, sexually transmitted infections, and vulvar or vaginal complaints. iv , v Little is known about the prevalence of presentations specific to the ED. Some national data suggests that genitourinary complaints are in the top ten of pediatric emergency department visits vi , but little data explores the granular presentations, patient demographics, and ultimate dispositions of these patients. Thus, the goal of this study is to describe the rates and associations of diagnoses, admissions, and transfers associated with obstetric and gynecologic emergency department presentations for pediatric and adolescent patients.