Sexual Health Behaviors and Pregnancy Risk Among Hospitalized Female Adolescents

We conducted a cross-sectional survey with a convenience sample of female participants (sex assigned at birth) aged 14 to 21 years and admitted to the medical and surgical units at 2 urban, academic, tertiary children’s hospitals in the Midwest from August 2019 to March 2020. Participants completed an anonymous self-administered Web-based questionnaire. Participants were eligible if they were English speaking and not developmentally delayed or too ill to participate. We excluded participants receiving active treatment of cancer because of the potential emotional burden of discussing reproductive health because their diagnosis and/or treatment could impact their long-term reproductive capability. The institutional review board at each participating institution approved the study.

We reviewed the electronic medical record (EMR) to identify eligible participants. After approval by the medical team, we approached eligible participants for enrollment a single time. Adult participants gave verbal consent, and minors gave verbal assent. The institutional review board waived parental consent, although if a parent was present, we provided a brief overview of the study and requested verbal permission to approach the eligible adolescent. Family members and visitors were asked to leave the room during full study procedures to ensure confidentiality. If family members declined to leave the room, the participant was ineligible. For those who declined to participate, we collected age, race and/or ethnicity, and insurance type from the EMR, although because of a procedural error, the enrollment log was incomplete at one study site (ie, completed on 48 of 63 enrollment days).

The survey contained no personal identifiers, all questions were voluntary, and responses were not linked to participants. Study data were collected and managed by using Research Electronic Data Capture.^29,30  The research team conducted an EMR review to extract age, insurance type, teratogen exposure, and comorbid medical conditions. On survey completion, the research team gave each participant information about local adolescent SRH resources and a $10 gift card.

A multidisciplinary team that included pediatricians from hospital medicine, emergency medicine, and adolescent medicine developed the survey instrument largely on the basis of the previously validated Youth Risk Behavioral Surveillance System (YRBSS) or National Survey of Family Growth as well as relevant literature.^{16,25,26,31,32}  The instrument was used to assess demographics (7 questions), health care use (5 questions), general and sexual health history (24 questions), and acceptability of hospital interventions (4 questions). The survey was pilot tested for comprehension with 5 adolescents who were representative of the study population; we made minor revisions accordingly.

We defined a participant as sexually active if they reported ever having had vaginal intercourse or if they selected any answer other than “I have never had vaginal intercourse” to the question “How old were you when you had vaginal intercourse for the first time?” We defined currently sexually active as vaginal sex in the previous 3 months. Consistent with the YRBSS, we defined early sexual debut as sex before age 13 years.³³  We did not categorize participants with a history of only oral or anal sex as sexually active, following guidelines for calculating pregnancy risk. Condom and contraceptive use was determined through report of use at last sex and frequency of condom use in the past. We defined reversible contraception as use of a birth control pill, patch, ring, injectable, or implant or any type of intrauterine device. We defined dual contraception as use of a reversible method of contraception in addition to a condom. Participants reported acceptability of SRH interventions using a 5-point Likert scale (eg, “I think it is helpful for healthcare providers to talk about birth control and safe sex practices to young people like me” and “I think a hospital stay is a good opportunity for healthcare providers to talk about birth control and safe sex practices to young people like me.”)

We calculated PRI using previously validated methods on the basis of self-report of current sexual activity (vaginal sex in the past 3 months), methods of contraception used at last sex, and method-specific contraceptive failure rates.²⁴  Participants were assigned a score of 0 if they did not report current sexual activity.^28,34  We calculated an overall PRI for the study sample and among subgroups (eg, age, race and/or ethnicity).²⁶ 

Other pregnancy-related outcomes that were evaluated included pregnancy history, pregnancy intentions, receipt of contraceptive counseling, medical comorbidities, and teratogenic medication exposure. Pregnancy desire was defined as a positive response (ie, “sort of true” or “very true”) to any one of the following: “I would be happier if I got pregnant,” “I feel it would be good for me to get pregnant at this time,” or “I would really like to get pregnant now.” We determined receipt of contraceptive counseling by asking what type of provider had discussed birth control with them in the previous 12 months. To assess for medical comorbidities, we developed a list of pediatric specific conditions documented in the US Medical Eligibility Criteria for Contraception Use under the category of “conditions associated with increased risk for adverse health events in pregnancy.”³⁵  A participant was determined to have one of the specified medical comorbidities through either survey self-report or EMR review. Similarly, we used a list of teratogenic medications from class D and X medications more commonly prescribed in pediatrics, as had been used in previous literature evaluating teenagers with teratogenic exposure.²³  A participant was determined to have exposure to a teratogenic medication either through self-report by replying “yes” to the question “Do you take any medications regularly that you believe might cause birth defects?” or through EMR review documenting exposure to ≥1 of the medications on the specified list.

We used means with SDs or medians with interquartile ranges to summarize normally and nonnormally distributed continuous data, respectively. We presented categorical data as proportions. We included participants even if they had missing data from skipped questions and noted data missing in frequency calculations. We used independent t tests and analysis of variance to compare grouped data and χ² tests to compare groups on categorical variables, such as reporting contraception discussion in the past year and contraception at last sex. Microsoft Excel 2016 for Windows and IBM SPSS Statistics for Windows (IBM SPSS Statistics, IBM Corporation, Armonk, NY) were used to conduct statistical tests.

We enrolled 177 participants across the 2 sites. Because of the coronavirus disease 2019 pandemic, we stopped recruitment in March 2020, earlier than anticipated. Among those approached for recruitment, we collected participation data for 166 adolescents. Of these, 141 participated (85%), with no differences in age, race and/or ethnicity, or insurance type between participants and those who declined. We removed 1 record because of concern for “jokester effect,” with inconsistent age between self-report and chart review. The mean age was 16 years (SD 1.5). Most participants were cisgender (97%), straight or heterosexual (74%), and non-Hispanic white (55%) (Table 1). Half (47%) had private health insurance, and most (92%) had a medical visit in the past year. Some (19%) reported there was a time in the past year when they thought they should get medical care but did not.

Nearly half (43%; 75 of 175) reported ever being sexually active, with 30% (53 of 175) reporting current sexual activity. Among those with a history of vaginal sex, 5% (4 of 74) reported early sexual debut and 28% (21 of 74) reported ≥4 lifetime partners. Oral sex and anal sex were reported by 38% (67 of 175) and 6% (10 of 175), respectively. Of those without a history of vaginal sex, 5 reported oral sex and none reported anal sex.

Among the 75 sexually active participants, 81% reported use of a condom and/or reversible contraception at last sex. Condom use was reported by 65%, and of these, 32% reported condoms alone and 33% reported dual contraception. Of those reporting reversible contraception (49%), 21% used the pill, patch, or ring; 16% used injectable contraception; and 12% used LARC. A minority (15%) reported “no method/not sure,” and 4% reported only using withdrawal. More than half (53%) reported ever having used reversible contraception, and 42% (31 of 74) reported always using condoms.

The majority (73%; 53 of 73) of sexually active female participants reported discussion about contraception with a health care provider in the past year; these participants reported increased use of reversible contraception (62% [33 of 53] vs 20% [4 of 20]; P = .001) and dual contraception (42% [22 of 53] vs 15% [3 of 20]; P = .03) at last sex. Contraception discussion was not associated with use of LARC (15% [8 of 53] vs 5% [1 of 20]; P = .24) at last sex. Of those with a history of only oral sex, 50% (2 of 4) received contraceptive counseling. Among those without past-year discussion about contraception, 85% (17 of 20) reported at least 1 clinic visit during this period.

Most participants (89%; 156 of 176) did not desire pregnancy. Four participants had a history of previous pregnancy; none reported multiple pregnancies. The PRI among all participants was 4.75 (Table 2). Participants with teratogenic medication exposure had a lower PRI compared with those without teratogenic medication exposure (0.32 vs 5.98; P = .049).

Of all 177 participants, 72% were at risk for pregnancy complications due to an underlying medical comorbidity and/or teratogenic medication exposure. Most (68%) had ≥1 medical comorbidity, most commonly a mental health disorder (51%), diabetes (10%), and inflammatory bowel disease (8%) (Table 3). Nearly one-quarter (22%; 39 of 176) had a teratogenic medication exposure. Of these, 59% were not aware they were regularly taking a medication known to cause birth defects, and 28% reported they had not discussed birth control with a provider in the previous year. Among the 16 sexually active participants with a teratogenic exposure, 88% reported use of reversible contraception and 69% reported dual contraception use at last sex. LARC use was more frequently reported in this group when compared with sexually active female participants without teratogenic medication exposure (31% [5 of 16] vs 7% [4 of 59]; P = .008).

Two-thirds (66%; 114 of 173) of participants agreed it is helpful for health care providers to talk about birth control and safe sex with adolescents. When asked if a hospital stay is a good opportunity to discuss birth control and safe sex practices, 53% (93 of 176) reported they agree or strongly agree, 34% (60/176) reported they were neutral, and 13% (23/176) reported they disagree or strongly disagree. Participants with a history of vaginal sex were more likely to agree or strongly agree to this question (P = .03). There were no differences in responses to this question among participants reporting reversible contraception use at last sex (P = .30), those reporting clinic visits in the past year (P = .20) or those reporting not seeking care when they felt they should (P = .10).

Sexual health behaviors and pregnancy risk among hospitalized female adolescents in our study were similar to those reported nationally. When compared with recent YRBSS data, we found similar rates of sexual activity (43% vs 40%), current sexual activity (30% vs 28%), sexual debut before age 13 (5% vs 7%), and having ≥4 lifetime sexual partners (28% vs 27%).^31,36  The PRI in our population was 4.75, compared with 5 in the general US female adolescent population.³⁷  However, nearly three-quarters of our population had an underlying medical condition associated with increased risk for pregnancy complications or teratogenic medication exposure.³⁷  Thus, a hospitalization represents an important opportunity to identify and address SRH needs in this population.

To our knowledge, this is the first study to describe PRI among hospitalized female adolescents. Our study population did not exhibit the twofold to fourfold higher PRI seen among adolescents receiving pediatric ED care.^25,26,37  In comparison with the ED-based studies, our study population reported higher rates of reversible contraceptive use at last sex (49% vs 32%–34%) and lower rates of no method or unsure of method (15% vs 21%–31%).^25,26  Additionally, our population reported higher rates of condom use at last sex compared to national adolescent data (65% vs 54%). Although these rates of contraceptive use are encouraging, our findings still reveal opportunities for improvement. Only half of sexually active participants reported ever having used reversible contraception, and less than half used these methods at last sex. A small minority reported LARC use.

We found that participants who reported receipt of contraceptive counseling had a threefold increase in report of reversible and dual contraception use at last sex compared with those who did not. However, we did not find an association with contraception counseling and LARC use. Previous literature revealed improved use of any contraception, including LARC, among adolescents who received comprehensive contraceptive counseling.^38,39  Further study is needed to better understand how comprehensive contraceptive counseling impacts contraception and LARC use in hospitalized adolescents.

More than two-thirds of our participants had an underlying medical comorbidity, and one-quarter had teratogenic medication exposure. Encouragingly, we found a lower PRI in those with teratogenic medication exposure due to higher rates of reversible contraception, dual contraception, and LARC use in this group. However, despite higher LARC use in this group, it was only reported by one-third of these sexually active participants. In addition, a majority of participants with teratogenic medication exposure were not aware they were taking a medication known to cause birth defects, and 1 in 4 reported no birth control discussion in the past year. Hence, there are opportunities to improve education regarding teratogenic medication exposure and increase contraception counseling, focused on LARC, in adolescents with medical comorbidities and teratogenic medication exposure.

Just more than half of participants (53%) were in agreement that a hospital stay is a good opportunity to discuss contraception, similar to a previous study (57%).¹⁶  One-third of participants reported feeling neutral on the topic. Those reporting past sexual activity were more likely to agree. Further exploration, including use of qualitative methods, into these findings could facilitate understanding of acceptability of SRH interventions among adolescents in the hospital, including in those with and without past sexual activity and contraception use. This knowledge can inform development of feasible and acceptable SRH interventions in this unique population.

Study limitations include the potential for response or social desirability bias, despite efforts to protect confidentiality and facilitate disclosure (ie, use of an anonymous online survey administered privately). Because we collected a convenience sample, our data are also subject to sampling bias. We were unable to calculate an overall response rate because of a procedural error at one study site. However, for the majority of enrollment days, the participation rate was 85%. The coronavirus disease 2019 pandemic necessitated closing enrollment earlier than expected, limiting our ability to fully investigate PRI differences among subgroups. However, our sample size is similar or larger than those of past studies conducted in hospitalized adolescents.¹⁹  Additionally, we did not collect data on medical versus surgical hospital service, but a majority of our participants had an underlying medical condition. Thus, our findings may not be generalizable to adolescents admitted for surgical diagnoses.

Hospitalized female adolescents have similar sexual health behaviors and a similar PRI to the general adolescent population; however, nearly three-quarters had an underlying risk factor for pregnancy complications due to medical comorbidities and/or teratogenic medication exposure. Participants reporting past-year contraceptive counseling had increased contraceptive use, yet many hospitalized adolescents had not received contraceptive counseling, and few were using the most effective contraceptive methods. Most participants agreed that a hospitalization is a good opportunity to discuss sexual health and birth control. Future study is needed to understand the impact of comprehensive contraceptive counseling in this population, to provide increased education of teratogenic medications, and to further explore acceptability of SRH interventions in hospitalized adolescents. This information can then inform effective delivery of SRH to improve health outcomes among hospitalized female adolescents.

We thank Avleen Bhandal, Angela Nwankwo, Raga Kilaru, Sarah Wells, and Varsha Kandadi for their help with data collection. Ms Bhandal, Ms Nwankwo, Ms Kilaru, Ms Wells, and Ms Kandadi have no financial conflicts of interest to disclose, and no compensation was provided for their assistance with this study.