Reproductive health is an important issue in the care of adolescents and young adults (AYA). Unfortunately, many AYAs, particularly those with chronic medical conditions, may not regularly access primary care to address these issues. This study evaluates the impact of an inpatient reproductive health consult service on reproductive health care.
A reproductive health-focused screening consult service was initiated in June 2017 at an academic teaching hospital. Patients aged 15 to 26 admitted to pediatric ward teams were eligible for screening. A retrospective chart review was conducted from December 2016 to June 2019 to determine the effect of the consult service on the primary outcome, documentation of a psychosocial assessment, and reproductive health concerns.
Nine hundred twenty-nine encounters were analyzed (345 preintervention and 584 during intervention), representing 675 patients. The proportion of encounters with a documented psychosocial assessment increased from 14.8% to 41.3% during the intervention (P < .001); a reproductive health screening consult was responsible for 37.3% (109 of 292) of the documented assessments. There were high self-reported rates of sexual activity (38%), substance use (47%), and mood concerns (48%) among hospitalized AYA; all behaviors were documented at statistically significant increased frequencies (P < .001) during the intervention compared with preintervention.
Initiation of an inpatient reproductive health screening consult service led to increased documentation of psychosocial assessments, including increased documentation of sexual health history and other risk factors. With improved screening of reproductive and psychosocial needs, targeted interventions can meet underrecognized needs among hospitalized AYA.
Addressing reproductive health for adolescents and young adults (AYAs) is lacking in many health care settings. Given known negative consequences of inadequate reproductive health care, including unintended pregnancy and sexually transmitted infections (STIs), providing adequate reproductive health care, especially for AYAs with chronic medical conditions, is important. Previous studies have found that sexual health screening is often missed in the primary care setting,1,2 which has led to efforts to expand access to reproductive health care in alternative sites, including urgent care and the emergency department.3–5 Studies have also explored increasing access in the inpatient setting and shown that both adolescents, as well as providers, were open to reproductive health evaluation.6–9 Additionally, research has shown that AYAs with chronic medical illnesses may be more likely to engage in risky health behaviors compared with healthy adolescents.10 At the same time, they are more likely to identify subspecialists, who may be less likely to provide comprehensive preventative care,11,12 as their main source of health care.13,14
Despite recommendations on universal reproductive health screening for AYAs,15,16 research exploring implementation of these recommendations in the inpatient setting found that screening is far from universal; biases exist with male patients and those presenting without certain clinical risk factors (eg, urinary symptoms, pelvic pain, etc), leading to lower rates of screening among these groups.17,18 In an effort to expand screening and enhance access, the Division of Adolescent Medicine developed a targeted reproductive health screening consult service to be implemented on the inpatient units. Reproductive health screening was completed utilizing the home, education, activities, drugs, sexuality, and suicidality/depression (HEADSS) assessment. Although substance use and mental health issues are critical to address for AYAs in their own right, their impact on reproductive care is well established, and therefore these factors were included as part of the assessment.19–21 The consult service was focused on patients admitted to both general and subspecialty pediatric teams. This study describes the effect of implementation of a reproductive health screening consult service among AYAs admitted to the inpatient service at a children’s hospital in Northern California.
Methods
A reproductive health-focused consult service run by the Division of Adolescent Medicine was initiated in June 2017 at an academic teaching hospital. A consult service model was chosen as the basis of this intervention because it was an existing clinical service to which a new clinical focus could be imposed. All patients aged 15 to 26 admitted to any of the 7 pediatric ward teams (2 general pediatric and 5 subspecialty) were eligible for screening, including non-English speaking patients. Patients <15 were excluded in this pilot phase to ensure adequate staffing of the consult service. Patients admitted to surgical teams were also excluded. Leadership from the reproductive health screening consult service team met with stakeholders from each ward service to review the intervention and gain consensus (see Fig 1 for timeline of intervention by service team). Hospital revenue leadership was also consulted before implementing the screening component of the consult service. As a result of targeting early adopters in our implementation, there was a staggered rollout of the consult service over time. All participating services agreed to their patients being screened via chart review by the consult team to assess the patient’s eligibility for the reproductive consultation. The project was determined to be exempt from the Stanford University institutional review board. The consult service consisted of an adolescent medicine attending with either an adolescent medicine fellow or a pediatric resident. All interns were taught how to conduct a HEADSS assessment during their orientation and a refresher session was given as part of didactic teaching during their adolescent medicine rotation. Logistics of the screening consult service were also reviewed during each monthly rotation. A list that was automatically populated with eligible patients was created in the electronic health record and was reviewed each weekday by the adolescent medicine team. The resident and/or fellow would review the chart to determine if a HEADSS assessment was already documented for the patient during the current admission. If there was no HEADSS assessment documented during the current admission, but 1 was documented by adolescent medicine within the last 6 months, then the patient was ineligible for a screening consult during that admission because of concerns about staffing during the pilot phase. However, those encounters were included in the retrospective chart review. If a HEADSS assessment was not documented after 24 hours of admission, the resident and/or fellow would contact inpatient teams for each eligible patient and inquire about whether there was a specific reason why the patient would not be appropriate for a reproductive screening consult. Adolescent medicine was also available for formal consults during the same time period for patients who did not receive a screening consult. After completing the consult, the adolescent medicine service would document a consult note and bill according to time and/or complexity for the services provided. Of note, there have been no concerns with regard to billing despite it being a theoretical concern on behalf of the primary team during the initial phases of implementation.
Data Collection
Variables of interest were obtained via retrospective chart review for every admission that met inclusion criteria in the 6 months before the intervention and 2 years postintervention for each respective service team. Demographic information collected via chart review included age, sex, race/ethnicity and insurance type to evaluate for potential biases in screening rates. The primary outcome was the documentation of a comprehensive HEADSS assessment. Secondary outcomes included use of contraceptives, use of teratogenic medications, and STI screening. REDCap was used for data collection and management. The Stanford University REDCap platform (http://redcap.stanford.edu) is developed and operated by the Stanford University School of Medicine research information technology team. The REDCap platform services at Stanford University are subsidized by (1) the Stanford University School of Medicine Research Office, and (2) the National Center for Research Resources and the National Center for Advancing Translational Sciences at the National Institutes of Health, through grant UL1 TR001085.
Charts were reviewed retrospectively by a team of 4 investigators. To ensure reliability and consistency between individual reviewers, a subsample of 15 charts was reviewed by all 4 investigators. Any disagreements were resolved by consensus. If a comprehensive HEADSS assessment was documented for that admission, secondary outcomes were collected, including additional details from the HEADSS assessment note, review of medications during that admission, and STI laboratory results within the past year. Each admission for a patient was reviewed separately, given that it is standard of care for AYAs to have a HEADSS assessment during each separate admission to the hospital.
Statistical Analysis
We analyzed demographic information by patient and not by encounter. As some patients had multiple admission encounters during the study period, primary and secondary outcomes were analyzed by admission encounter during the 2 intervention periods (preintervention versus during intervention). Bivariate analysis was performed for our study outcomes of interest. Chi-squared tests were used to compare the distributions of the primary and secondary outcomes (proportions) before and during intervention. Fisher’s exact tests were performed for all categorical outcomes that had at least 1 stratum with <5 encounters. Multivariate logistic regressions were performed to evaluate the impact of the interventions upon the primary and secondary outcomes. Generalized mixed-effect models were implemented to obtain robust standard errors by taking into account the correlations within patients and/or within the subspecialty teams. Patients’ demographics, including age, sex, and primary insurance group, were adjusted in the model. Race/ethnicity was not included in the model because of lack of statistical significance when comparing the pre- and during-intervention groups, as well as limited power to include the variable on the basis of sample size and higher clinical relevance for the other demographic variables included. The odds ratio (OR), corresponding 95% confidence interval (CI), and P value were reported from the model. Subgroup analysis was performed in a similar fashion by further stratifying the study cohort into the 5 major service groups. Complete case analysis was performed as low rates of missing data were observed in our study cohort (<5%). Type I error was controlled at 0.05. All analysis was performed using R statistical programming language, version 3.4.3.
Results
Between December 2016 and June 2019, a total of 929 admission encounters met inclusion criteria. A total of 345 admissions occurred during the preintervention phase (December 2016–May 2017) and 584 admissions during the intervention period (June 2017–June 2019). Intervention start dates differed by ward team (Fig 1). Overall, 71% of the admissions were to a subspecialty service team.
There was a total of 675 individual patients represented in the 929 admission encounters included in the study. The demographic characteristics by patient are shown in Table 1. The preintervention group was noted to have 41.0% female patients as compared with 51.3% in the intervention group. The mean age of both groups was approximately 17 years with SD of 2.2, (P = .67). The median length of stay was 3.7 days (interquartile range of 1.7–7.4 days) preintervention and 3.6 days (interquartile range of 1.8–8.0 days) during intervention (P = .80). There was no statistical difference between the proportion of patients with public versus private insurance status in the preintervention and during-intervention groups (P = .14).
Overall, the total number of encounters with HEADSS assessments completed was 292 in our entire study cohort, increasing from 14.8% preintervention to 41.3% after the initiation of our intervention consult service (P < .001, Table 2). A reproductive health screening consult was responsible for 45.2% (109 of 241) of the documented assessments during the intervention. Of note, 32 (3.4%) encounters had a limited HEADSS assessment documented by psychiatry providers, which included some, but not all, of the variables of interest. A total of 371 encounters were found to document a patient’s sexual activity status, which increased from 24.9% preintervention to 48.8% after the initiation of our consult service (P < .001). Out of the 292 encounters in which a HEADSS assessment was documented, 114 of 292 (39.0%) encounters included documentation of a patient ever being sexually active, and 88 of 292 (30.1%) encounters included documentation of a patient reported being currently sexually active. Out of these 88 encounters, 18 (20.5%) of those were not using a contraceptive method and 20 (22.7%) were taking a teratogenic medication. Of those patients who were sexually active and taking a teratogenic medication (n = 20), 17 (85%) were on a contraceptive method. Out of the 114 encounters in which a patient reported having ever been sexually active, STI screening was recommended in 51 (44.7%) of adolescent medicine consult encounters and was completed in 38 (33.3%) of those. There was no statistically significant difference in STI screening of those who were ever sexually active in the pre- versus during-intervention time periods.
Out of 357 encounters with substance use documented in our cohort, 113 (31.7%) reported substance use within the last year (most commonly marijuana, alcohol, and tobacco). The documentation of substance use also increased from 23% preintervention to 47% after the initiation of our intervention consult service (P < .001). Out of 362 encounters with mood history documented, 163 (45%) reported mood concerns. The documentation of a mood disorder increased from 23.2% preintervention to 48.3% during our intervention (P < .001). Depressive symptoms were the most commonly documented mood concern (Table 3).
Regression modeling showed a significant increase of the odds of having a HEADSS assessment documented (OR: 5.3 [3.4–8.4], P = .003) during intervention as compared with preintervention, adjusting for age, sex, and insurance status, as well as clustering within subspecialty team and within patient. Subgroup analysis performed by the type of service team showed that the odds of having a HEADSS assessment documented for patients on the general pediatrics service during the intervention phase were significantly higher compared with the preintervention phase (OR 2.63, 95% CI 1.25–5.53, P = .011). For patients admitted to subspecialty services, the increase in the odds of completion of HEADSS assessment was even greater (OR 7.36, 95% CI: 4.18–13.0, P < .001).
Discussion
This study reviews the impact of psychosocial assessment before and during the implementation of a targeted inpatient reproductive health screening consult service. In our preintervention group, we found remarkably low rates of psychosocial assessment documentation during the patient’s admission. Although our inpatient hospital manages extremely complex patients with one of the highest case-mix indexes in the country, many academic children’s hospitals likely have similar inpatient populations that rely on similarly specialized care.22,23 Perhaps the low rates of screening reflect the medical acuity of patients at the time of admission; however, conducting a psychosocial assessment once a patient is stabilized remains critical for identifying risk factors that may affect their reproductive health, as well as other medical conditions. The low rates of psychosocial documentation underscore the opportunities for improving the clinical care of AYAs; relevant and necessary recommendations can only be made after comprehensive screening identifies unmet needs.
Our study shows that implementation of a reproductive health screening service increased the documentation of HEADSS assessments and reproductive screening for AYAs admitted to pediatric ward teams. Although this increased screening was noted among all pediatric ward teams, the effect was even more apparent among subspecialty teams. This has clinical implications given that subspecialty teams are often caring for patients with chronic conditions: a population of patients who often have lower utilization of primary care but with increased health risk behaviors.10–14 The potential for exposure to teratogenic medications for AYAs receiving subspecialty care is likely higher, as well, and further underscores the need for focused reproductive health care. Using the inpatient venue may allow intervention with a higher-risk population who are not accessing care in other settings.
Not surprisingly, when AYAs are asked about health behaviors, they endorse sexual activity (including unprotected sex), substance use, and mental health concerns.24 Although the percentages of participants endorsing engagement in these behaviors increased during the intervention, we believe this change is a result of better screening, not necessarily an increase in actual behavior when compared with national trends over the same time period.25–27 Nearly one-third of patients in the intervention period reported being currently sexually active. Reassuringly, about 75% were on a contraceptive method, though this remains a gap still to be addressed. Additionally, about one-quarter of patients who were sexually active were taking a teratogenic medication. Although some drug manufacturers require reproductive health counseling with the prescription of a teratogenic medication, this represents only a small fraction of the teratogenic medications commonly prescribed.28 More robust screening and counseling, such as that provided by a reproductive consult service, should be considered to uniformly understand patient risk factors and manage them appropriately.
Though not necessarily the focus of the reproductive health screening consult service, we did screen for other health risk behaviors, including substance use and mental health concerns. About 1 in 6 patients endorsed substance use within the last year during the intervention period, with types of substances reflective of national trends among AYAs.29 Even more striking was the fact that about half of participants reported mood concerns, most commonly depression and anxiety. The presence of substance use and/or mental health issues is especially salient for patients with chronic medical conditions given potential impact on adherence with treatment recommendations.10
Although we’ve demonstrated that implementation of a reproductive health screening consult service increases psychosocial assessments among AYAs, it is noteworthy that we were only able to achieve screening in less than half of admissions. There are many factors that may be contributing to this less-than-ideal outcome, including patient level factors such as inability to participate in a HEADSS assessment because of severe intellectual disability or medical instability (eg, severe pain or altered mental status). Though multiple attempts were made to complete the HEADSS assessment by the adolescent medicine service once patients were clinically improving and able to participate, this was not always achieved. Family or patient refusal was another factor, though both were relatively rare. Scheduling may have affected completion of a consult because outreach by the consult team was only available during the weekdays; therefore, patients admitted and discharged during the weekend would not have been captured. Finally, given data were obtained via retrospective chart review, we cannot comment on the specific reasons the HEADSS assessment was not done because that was not reliably documented. Because chart review was based on documentation of HEADSS assessments performed, it is possible that screening occurred but was not documented and, thus, current rates in our study could be an underestimation of assessment frequency.
In addition to the limitations mentioned above, another factor that could affect our results is that the consult intervention was rolled out to service teams over several months. Given crossover of medical trainees to different ward services, there may have been a blunting of the intervention effect if a trainee had previously been exposed to the intervention while rotating on a “preintervention” service. We were unable to account for this in our regression modeling, although we do not anticipate that the results would be significantly affected given that trainees have received formal teaching on the HEADSS assessment as part of their medical education.
Conclusion
This study demonstrates that a targeted, inpatient, reproductive health screening consult service resulted in increased rates of documented HEADSS assessments in AYAs. We found that patients endorsed high rates of health risk behaviors, which would benefit from additional counseling and management. Though rates are improved with the consult service, they are still far from ideal. Further work is needed to address barriers preventing true universal screening, as well as identifying patient- and provider-facing tools and workflows that may better facilitate routine screening on a consistent basis.
Acknowledgments
We thank Dr Nivedita Srinivas for being instrumental in developing the study design and implementation strategy for this article.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: Dr Goldstein is a Merck Nexplanon Clinical Trainer. The other authors have no conflicts of interest relevant to this article to disclose.
Drs Carlson and Goldstein conceptualized and designed the study, supervised and participated in data collection, and prepared and revised the manuscript; Drs Keppler and Pineda collected data and reviewed and revised the manuscript; Ms Weng assisted with study design, conducted the statistical analysis. and prepared and revised the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
