Social Navigation for Adolescents in ED

Significant advances have been made in pediatrics in the United States in the last 60 years. Mortality from infectious disease and cancer, for instance, has declined significantly. The more stubborn issues in pediatrics include poor asthma control, obesity, suicide and adolescent homicide. To address these, our profession must begin to systematically address social determinants of health. Scourges such as poverty, unemployment, and domestic violence are known contributors to pediatric morbidity, adverse childhood experiences, and poor adult health and achievement. - The emergency department (ED) may be a strategic venue in which to identify patients at high risk for multiple social issues. We suspect that certain ED patient characteristics may be indicative of those families at highest risk and that comprehensive screening for social determinants of health may be the means by which they can be identified and targeted for intervention.

Social issues in EDs are often addressed individually instead of comprehensively, if done at all. There have been no studies of comprehensive screening for social determinants of health in large, urban, full service pediatric EDs. In addition, no study has attempted to identify those groups of ED patients at highest risk for a large number of social risk factors who might benefit most from rapid interventions. The immediate objective of this proposal is to identify groups of patients at high risk for a large number of social risk factors in a patient-centric manner in a busy pediatric ED, using a single computerized instrument to administer a comprehensive screen and to provide social navigation services to adolescent patients. We will sequentially approach caregivers of patients 0-17 years of age and patients 13-21 years of age to complete surveys regarding social risk factors. The outcomes of interest are 1) the total social risk score and 2) ED recidivism, total social risk score and community resource use in adolescent patients who receive social navigation services vs. enhanced usual care.

We hypothesize that patients with the following ED characteristics will have a higher social risk score than other ED patients: patients with any of nine "high risk" chief complaints (alleged physical abuse , alleged sexual abuse , peer assault , mammalian bites, adolescent reproductive and sexual health issues, intoxication/substance abuse , ingestion/poisoning , psychiatric/behavioral complaints , any complaint with the lowest estimated severity index ); a body mass index in the obese range; frequent non-urgent visits to the ED; or poor asthma control. Further, we hypothesize that adolescents who receive social navigation services will have lower ED recidivism, a lower total social risk score and increased community resource use at 12-month follow up than adolescents who receive enhanced usual care.

SPECIFIC AIM #1: To characterize the cumulative social risk score for patients and families in a large, urban pediatric ED Hypothesis #1: Caregivers of pediatric ED patients and adolescent ED patients will report an increased number of social risks than patients in primary care settings have reported in prior studies.

SPECIFIC AIM #2: To compare the cumulative social risk score in pediatric ED patients with a subset of 9 (nine) high-risk chief complaints, frequent non-urgent visits, obesity, or poor asthma control to the cumulative burden of social risk factors in the general pediatric ED population Hypothesis #2: Patients who present to the ED with any of a subset of nine high-risk chief complaints, frequent non-urgent visits, obesity or poor asthma control will have a higher social risk score than the general pediatric ED population.

SPECIFIC AIM #3: To compare ED recidivism, total social risk score and level of community resource use at 12-month follow up in adolescents who receive social navigation services versus those adolescents who receive enhanced usual care.

IMPACT: It is our intent to revolutionize emergency care, cementing its place as not only a medical safety net, but a social one as well. EDs across the country will benefit from demonstration of a method to intervene on behalf of pediatric patients whose social problems are as important as their medical ones, if not more so. These interventions have the potential to decrease unnecessary ED visits and ED recidivism for chronic problems caused or exacerbated by unmet social needs. The ultimate goal is expansion of social navigation teams in the ED and creation of a follow up clinic to address the longer term social risk factors faced by many of our families.

• Background and Significance: The World Health Organization defines health as "the absence of physical, psychological and social problems". Health care, as it has traditionally been practiced in the United States and Western Europe, has focused primarily on disease and diagnosis. This has proved insufficient, however, at improving disparities in health among patients of different incomes, geographic origins and ethnicities. These disparities may be due, in part, to lack of attention to social determinants of health. Addressing social determinants is a goal of Healthy People 2020 and has now been mandated by the Institute of Medicine. - The pediatric age group, as studied by the life course model of human health , seems especially sensitive to social determinants, and these effects may persist into adulthood. , Downstream effects of poverty, substance abuse and child maltreatment can be seen at each stage in the life course. These issues can strongly affect cognitive functioning, , chronic disease in adulthood - , adult mortality - and adult substance use and mental health, , which themselves can negatively affect future educational/professional achievement and earnings.

Children exist not in a vacuum, but in the larger context of their families and communities. It is difficult to improve their health without addressing the well-being of their adult caregivers and families. Given its accessibility and positioning as both an entry point to and last resort for health care, the pediatric ED may be an ideal venue for screening and brief interventions for social problems.

Relationships between conditions addressed in the ED and social determinants are well established. - It is well known, for example, that unaddressed mental health issues and poor literacy in parents of asthmatic children increase their risk of frequent ED visits and hospitalizations. , These relationships exist for conditions and chief complaints as diverse as peer and adolescent violence, depression, diabetes, and sexually transmitted infections, often seen disproportionately in the ED. Prior work by Fleegler et al. has found that over 50% of patients in a primary care clinic reported two or more unaddressed health-related social problems (HRSPs). The adult population that visits the ED is at even higher risk for HRSPs than the general population , and it is likely that this will hold true in children as well. - Building on previous work in our Division in extending the reach of emergency care to include screening and secondary prevention, we propose the implementation and testing of screening for HRSPs in the ED. We have demonstrated success using audio-assisted computer screening with mobile touch pad technology, a screening method preferred by both patients and providers. Our long-term goal is to identify and intervene to interrupt the deleterious effects of HRSPs on pediatric health. The immediate objective of this proposal is to demonstrate the ability to screen for HRSPs in a patient-centric manner in a busy ED and to provide social navigation services to an intervention group of adolescent patients. We will use audio-assisted computerized patient/family surveys and an innovative multidisciplinary social navigation team to accomplish these objectives.

Preliminary Studies:

1. Screening for behavioral risk factors is acceptable to adolescents and their caregivers. We conducted a computerized survey of 276 adolescent patients and 138 caregivers. Respondents reported >60% acceptability of screening for risk factors such as depression and suicidality, substance use, sexual risk behaviors, violence, and housing instability.
2. Computerized sexual health surveys to guide clinical decision making improves testing for sexually transmitted infections (STIs) for adolescents in the emergency department Dr. Goyal's K23 (HD070910) developed a computerized sexual health survey (SHS) to identify adolescents at high risk for STIs and tested whether implementation of this tool, with results-based clinical decision support, increased STI testing rates. Through a multi-step iterative qualitative study using a Delphi panel of key informants for survey development and content validity, and cognitive interviews with end-users, Dr. Goyal developed a content-valid audio computer assisted self-interview (ACASI) SHS for the identification of adolescents at risk for STIs. This tool was understandable, well-accepted, and rated easy to use by adolescents in the ED. Furthermore, Dr. Goyal demonstrated that it was feasible for successful implementation into the ED workflow.

6.0 * Research Design and Methods: Include a detailed description of all procedures that will be conducted, including those performed as part of regular care. When applicable, include the dates for charts that will be accessed and the system used to access them.

This study will utilize two self-administered, computer-based surveys that test thirteen (13) social domains on a survey administered to caregivers of children 0-17 and thirteen (13) social domains on a separate survey administered to adolescents 13-21. The caregiver survey questions parents regarding the following social domains: housing, employment, income and benefits, safety practices, access to health care, parental depression, food security, education, substance abuse, immigration, legal issues, intimate partner violence and literacy. The adolescent survey questions adolescents regarding the following social domains: housing, food security, education, safety practices, access to health care, mental health, substance abuse, sexual practices, dating violence, immigration, legal issues, human trafficking and literacy. The study will be conducted in the emergency departments at the Children's National Medical Center Sheikh Zayed main campus and the Children's National Medical Center satellite campus at United Medical Center.

Survey Instrument

The survey has been developed through adaptation of previously validated questionnaires, as described below. Once these questionnaires have been collated and adapted for our comprehensive screening tool, we will then pilot the survey with 5 adolescents and 5 caregivers for understanding. If revisions to the survey are required, we will re-pilot the survey once revisions are made and repeat this process until no revisions are required.

The basis of the current survey is The Online Advocate, developed and utilized in a previous study by Eric Fleegler et al and used with permission of the author.

The housing domain includes questions derived from the American Housing Survey. It assesses household size, current housing status, concerns about impending eviction or foreclosure, utilities and housing hazards. Housing hazards include leakage, problems with electricity, no heat for > 24 hours, no toilets functional, any rodent or insect infestation, or no running water in the home. A patient or family will automatically score positive for a housing problem if 1) the family is homeless or living with family or friends for financial reasons ("doubled up"), 2) if there are greater than 8 family members at home, 3) if a family is concerned about an eviction, transfer or foreclosure, 4) if utilities have been shut off for nonpayment in the last 12 months, or 5) if there are two or more unaddressed housing hazards in the last 12 months.

The food security domain questions are the two-item screen designed by Hager et al. and endorsed by the American Academy of Pediatrics in its 2015 guideline Promoting Food Security for All Children. This screen has 97% sensitivity and 83% specificity for identifying food insecurity in a family. , A patient or family will score positive for food insecurity if they answer both questions in the affirmative. An adolescent will also score positive for food insecurity if they qualify for WIC and don't receive it.

The employment and income security domains use questions from the Philadelphia Survey of Work and Family and the Behavioral Risk Factor Surveillance System. A parent will score positive if they are 1) out of work or unable to work and not receiving disability benefits, 2) not receiving benefits (WIC, SNAP, CSFP, TANF or Medicaid) for which they qualify according to household income, or 3) earning less than $10,000 per year. A positive score will be reversed if the parent is out of work because they are 1) in school or a job training program, 2) on family or maternity leave, or 3) retired.

The safety domain includes questions about car seat use, helmet use, smoke alarms, and guns in the home. A patient or parent will score positive if 1) they are not using a car seat/booster seat if appropriate for age and height, 2) their child does not use a helmet regularly, 3) there are no working smoke alarms in the home, or 4) there are guns at home. The adolescent survey adds a question about being a witness to shootings/stabbings/murder. The adolescent will score positive if they 1) sometimes or frequently carry weapons, 2) have been a witness to shooting/stabbing/murder, 3) do not use helmets, or 4) if there are guns at home.

The access to health care domain includes questions from the Behavioral Risk Factor Surveillance System . It assesses health insurance status for both parent and child. Additionally, it assesses availability and utilization of primary medical care (if appropriate) for parent and child. A parent or patient will score positive if either she or her child does not have or has not used their primary medical provider within a 5- or 2-year time frame, respectively, or does not have medical insurance. Questions regarding asthma control were adapted from the Asthma Control Questionnaire (ACQ) for adolescents and the Pediatric Asthma Control and Communication Instrument for the Emergency Department (PACCI-ED) for children under 13 and will be used to test an association between poor asthma control and number of HRSPs.

The parental and adolescent depression questions are the PHQ-2 screen, which has been validated against longer depression screens in adolescents and adults. , A patient or parent will score positive if they earn a score of ≥ 3. ). If they earn a score of ≥3, the screen will convert to the PHQ-9 screen (for adults) or the PHQ-A (for adolescents), which includes questions regarding suicidality. If either the patient or parent has a child under 3 months old, the depression screen will instead be a 3-question subscale of the Edinburgh Postpartum Depression Screen (EPDS), shown to have sensitivity for postpartum depression for adults and adolescents and tested in the pediatric emergency department ; if their score on the 3-question subscale is greater than 3, the survey will automatically branch to a question regarding suicidality. Additional questions exist about recent deaths in the household and bullying. They will score positive if they have had a recent death in the household, have ever lost a parent or guardian, or if they admit to bullying or being bullied frequently (or the parent is concerned that the child is being bullied or is bullying others).

Questions regarding parental education include highest grade completed and number of adults in the household with fluency in English. The child's or adolescent's education is assessed with questions regarding school attendance, failing grades, and truancy. A parent survey will automatically score positive for education if 1) the parent did not complete high school and has not earned a GED, 2) there are no adults in the household who can speak English fluently, 3) if their child has had more than 5 unexcused absences or days home due to suspension in the past month, or 4) if the child has received 3 or more Ds or Fs in the last grading period. An adolescent will automatically score positive for education if they self-report either of the last two items in the affirmative.

The CRAFFT questions screen adolescents for problematic drug or alcohol use. The TICS (two-item conjoint screen) questions screen parents for drug or alcohol abuse or dependence. It was shown to have nearly 80% sensitivity for detection of problematic drug and alcohol use. Two positive answers on either survey constitute a positive screen for potential problematic alcohol or substance use. Parents and adolescents are also questioned regarding household members that they think may be addicted to drugs or alcohol and will earn a point if they answer in the affirmative. They will also screen positive if they admit to smoking cigarettes and the patient has asthma.

Immigration issues are assessed by asking about current immigration status, whether any adult at home is a U.S. citizen or legal resident and any recent deportations or arrests for immigration issues. Undocumented status, no adult U.S. citizen or legal resident in the home, no adult English speakers at home or any deportations in the last 12 months will constitute a positive screen for immigration.

Screening for intimate partner violence is accomplished via the STaT screen. It was tested in an urban emergency department population and has a sensitivity of approximately 96% for intimate partner violence. Any positive response will constitute a positive screen for intimate partner violence. Questions regarding child witnesses to domestic violence are included to screen for possible child protection issues.

Legal issues assessed include involvement in the criminal justice system of either parent/guardian or the child. Involvement of any parent/guardian or the child with the criminal justice system will constitute a positive screen.

The adolescent survey will include questions regarding sexual health. An adolescent will score positive if he/she has had 5 or more lifetime partners, has ever been pregnant or gotten someone pregnant, has had 2 or more partners in the last 3 months, is sexually active with opposite sex partners and is not using any reliable birth control method, or is sexually active with any partner and not using any reliable sexually transmitted infection (STI) prevention method. In addition, an adolescent will score positive if they have had an STI in the last 12 months or if they have not been tested for an STI in the last 12 months (unless not sexually active).

Low literacy by patient or parent as assessed on the initial literacy questions will earn an extra point.

Study Participants

Caregivers of patients ages 0-17 and adolescents ages 13-21 who speak English or Spanish will be approached for survey completion following physician evaluation. Patients will be excluded if they are triaged at estimated severity index (ESI) Level 1 or if they are unable to complete the survey due to developmental level, inability to speak English or Spanish, inability to read the survey when a private area/interpreter is not available to have the survey administered verbally, or mental status. Patients will also be excluded if they are ≥22 years of age or if they present in police custody or as wards of the state. Caregivers will be excluded if they are unable to complete the survey due to developmental level, inability to speak English or Spanish, inability to read the survey when a private area/interpreter is not available to have the survey administered verbally, or mental status.

If a child presents with more than one parent or legal guardian, the researcher will ask that only one parent complete the survey. If an adolescent presents with a parent/legal guardian but is either unwilling or unable to complete the survey, the survey will be scored as if the adolescent presented for care alone (i.e., in the "adolescent alone" group, maximum score 12 [twelve]). If a parent/legal guardian presents with an adolescent who is unwilling or unable to complete the survey but the parent does consent to complete it, the survey will be scored in the "parent alone" group (maximum score 13 (thirteen)).

Study Procedures

A convenience sample of consecutive ED patients during time blocks structured to reflect the variability of ED arrival times will be enrolled.

Following consent but prior to survey completion, the research assistant will enter into the survey the time of day, the patient's chief complaint and chief complaint characteristics, age, ESI level, number of non-urgent visits to CNMC in the last 12 months, ED campus, height and weight in a blank electronic survey identified only by the patient's study number. This information will be accessed from the Cerner FirstNet emergency department tracking software. The adolescent and parent surveys will be linked only by a common study number without utilizing any other identifying information.

If not completed at triage, the research assistant will measure the patient's height using the emergency department stadiometer and standardized technique. Patients who are unable or unwilling to stand will have their length measured if able to lie flat.

Literacy and computer facility will be assessed with four validated screening questions administered to the parent and/or adolescent patient prior to initiation of the body of the survey. If the patient is deemed to have high reading and computer literacy by this standardized assessment, the research assistant will then provide the laptop, equipped with a privacy screen and headphones, to the family for completion of the survey. If they are deemed to have low literacy, the patient or parent will be instructed to use the audio feature of REDCap for each question and will be offered a touch screen laptop or iPad option, or the research assistant will read the questions to the patient and enter the answers. If the research assistant is required to read the questions to the parent, the survey will be conducted in a private room. If none is available within a reasonable time frame, the patient will be excluded from the study. If not required to read the questions to the patient, he/she will conduct a brief tutorial, then exit the room while the survey is completed.

The research assistant will review survey answers immediately following completion. If any of the following occur in any patient or caregiver on non-social navigation days, the procedures described below will be carried out.

1. If the study subject indicates that they wish to speak with a social worker, the emergency department social worker will be notified. Study recruitment will only be carried out when an emergency department social worker is available in the ED. The social worker will only be notified of the fact that the patient would like to speak with them (i.e., not about specific survey answers), unless any of the contingencies described below exist.
2. If an adult study subject indicates that he/she is suicidal, the clinician in charge of the medical care for the adult study subject's child will be notified. The DC Mobile Crisis Services unit of the Comprehensive Psychiatric Emergency Program will be activated. This service is available between 9 a.m. and 1 a.m. Study recruitment will not be carried out outside these hours.
3. If a study subject 17 years of age or under indicates that he/she is suicidal, the clinician in charge of the medical care for that study subject will be notified. An emergency psychiatric consult in the CNMC emergency department will be obtained.
4. If an adult study subject indicates that he/she is currently experiencing intimate partner violence, the study subject will be asked face to face by the research assistant if they 1) would like to speak with a social worker, 2) would like to activate police services in the appropriate jurisdiction, 3) if they feel safe going home, and 4) if their child has witnessed or been physically affected by intimate partner violence in the household. The ED social worker will be notified and will initiate appropriate procedures, including the following:
5. Police services will be activated at the patient's request.
6. If the patient does not feel safe going home, they will be referred to the nearest 24 hour shelter and a warm handoff will be done.
7. If the study subject's child has witnessed or been physically affected by intimate partner violence, a Child Protective Services report will be made.
8. If a study subject 17 years or under indicates that they are currently experiencing intimate partner violence, the study subject will be asked face to face (but away from their caregiver) by the research assistant if they 1) would like to speak with a social worker, 2) would like to activate police services in the appropriate jurisdiction, 3) if they feel safe going home, or 4) if applicable, if their child has witnessed or been physically affected by intimate partner violence in the household. The ED social worker will be notified and will initiate appropriate procedures, including the following:
9. Police services will be activated at the patient's request.
10. If the patient does not feel safe going home, they will be referred to the nearest 24 hour shelter and a warm handoff will be done.
11. If the study subject's child has witnessed or been physically affected by intimate partner violence, a Child Protective Services report will be made.
12. If in the course of conversations about any of the above items a concern for physical, sexual or emotional abuse or neglect, or human trafficking of a person under the age of 18 is uncovered, the child's clinician and social worker will be notified and a Child Protective Services report made.
13. If the patient screens positively for human trafficking, the patient will be questioned privately. If there is continued concern, the National Human Trafficking Hotline will be called.

Adolescent patients in this study will be randomized into an intervention group, which will receive social navigation services and an enhanced usual care group, which will receive screening, the procedures described above, and a packet of community resources. The intervention group will have access to the services of a social navigation team. The social navigation team will consist of trained community health liaisons (undergraduates and graduate students in medicine, public health and social work from neighboring institutions including George Washington University, University of Maryland, Howard University, Catholic University, and Georgetown University), a clinician and a social worker.

The social navigation team will function as follows:

1. Once the computerized screening tool has been completed by the enrolled patient, the results will be printed by the research assistant.
2. The research assistant will then instruct the community health intern to review the results of the completed computerized survey.
3. The community health intern will review the results, create a plan of action, i.e. specific referrals to community agencies and next steps needed by the caregiver and or adolescent (e.g., documents to gather, appointments to make, etc.), following pre-developed protocols for each risk area. Each plan will be reviewed with the social worker prior to presentation to the family. Each family will also receive a packet of community resources relevant to each social domain covered in the screening survey, similar to that provided to the enhanced usual care group.
4. If emergency social concerns are uncovered, e.g., homelessness, abuse, or intimate partner violence compromising safety, the social worker will be notified immediately and she will proceed as per her usual practice, as above.
5. If adolescents report suicidality, abuse or high risk sexual activity, the team clinician will, as necessary, consult psychiatry, order STI testing and prophylaxis, order emergency contraception, notify child protective services or complete any required medicolegal forms.

Patients in the enhanced usual care group will receive risk-tailored printed information regarding community resources from the research assistant based on responses to the computerized survey.

Each enrolled adolescent participant will receive phone follow up at 3, 6 and 12 months by the community health intern. The interviewer conducting the 3-month phone follow up will assess whether any community resources were accessed. If not accessed, reasons for non-use will be assessed. If accessed, satisfaction with resources will be assessed. If participants have not accessed services, the navigator will provide additional resources to the participant to enable access by addressing identified barriers (e.g. making appointments or helping with transportation). If the phone number given is out of service, we will attempt follow up using email or the second phone number provided by the participant. Each participant will be allowed three unsuccessful phone calls and 1 email before being considered lost to follow up.

The 6-month phone follow up will be answer any patient or caregiver questions, paying strict attention to confidentiality, and will confirm contact information for subsequent 12-month follow up. At 12-month follow-up, the patient's chart will be reviewed to assess for follow up ED visits. A phone call will obtain contact information to 1) ask about non-Children's National Health System ED visits, 2) quantify continued community resource use, and 3) to obtain email or phone contact information through which the patient will complete the same computerized screening tool. If adolescents report new suicidality, child-witnessed intimate partner violence, physical or sexual abuse or human trafficking on the follow up survey, we will proceed as described above.

All emergency intervention(s) the provider uses if a patient or caregiver screens positive will be documented on an enrollment form. On non-social navigation days, providers will sign a form acknowledging receipt of positive emergency screens and intervention materials. This form will also allow the provider to ask research to contact social work on their behalf.

Randomization

Adolescent participants will be randomized to either receive or not receive social navigation services using a random time and date generator to create social navigation and control shifts that cover morning and evening shifts at both the Sheikh Zayed campus and UMC. This study will employ a consecutive enrollment sampling technique during randomly selected blocks to minimize selection bias. On social navigation shifts, the community health liaison, social worker and clinician will be available. On non-social navigation shifts (control shifts), patients will receive screening, a packet of community resources and follow-up phone calls to confirm contact information and inquire about community resource use, but no social navigation services.