Why this matters
The Epstein files document something clinicians who work with adolescent patients already know: that the medical system had multiple chances to see what was happening to those girls, and did not. Some of them were pregnant. Some had abortions coordinated through the network. Some had clinical notes written about them — "emotionally immature," "easily manipulated" — by clinicians who were sitting in front of children and documenting that fact and not reporting it. The files are not the first time this pattern has appeared in a chart. They are simply the most documented.
I want to be precise about what I am saying. The failure mode in adolescent exploitation is rarely dramatic. The patient who is being trafficked does not usually walk into clinic with a sign on her chest. She walks in with a sore throat. With a request for emergency contraception. With abdominal pain that no one has gotten to the bottom of. She walks in with an adult who is doing most of the talking, and the adult is calling himself her boyfriend or her uncle, and the patient is letting him. She walks in and she is fifteen and her answers are slightly too rehearsed and her affect is slightly too flat and her chart says she has been in three times in eight months for sexually transmitted infection screening and once for a missed period, and no one has asked her what is actually going on. This guide is for that visit.
What the evidence says
Child sexual abuse and adolescent sexual exploitation in the United States are not rare. The CDC's most recent national prevalence data continue to indicate that roughly 1 in 4 girls and 1 in 13 boys experience sexual abuse before age 18. Approximately 91 percent of child sexual abuse is committed by someone the child knows. The National Center for Missing & Exploited Children's annual CyberTipline reports document continuing growth in online enticement and child sex trafficking reports; the 2024 figures, the most recent available, exceeded the prior year again.
The American Academy of Pediatrics' 2024 Council on Child Abuse and Neglect clinical report, The Pediatrician's Role in Preventing Child Maltreatment (Stirling et al., Pediatrics 154(2):e2024067608), is the current canonical provider-facing guidance and frames recognition, screening, and reporting as a routine pediatric responsibility — not an exception. The National Child Traumatic Stress Network (NCTSN) hosts the largest current library of provider-facing tools for recognition and trauma-informed response, including screening instruments and free training modules.
The bias-failure mode is now well-documented in the peer-reviewed literature. Black and Native girls are recognized as victims of trafficking and exploitation at significantly lower rates than their white peers presenting with comparable signs, and are more often labeled as runaways, as "fast," as engaged in willing transactional sex, or as participating in criminal activity. The same chart presentation produces different clinical conclusions depending on who is in the bed. That is not an artifact of statistical chance. It is implicit bias operating in the medical workflow, and providers should expect it in themselves.
Red flags — what to notice
- Clinical signs. Recurrent sexually transmitted infections, especially in patients under 16. Unexplained pregnancy or pregnancy loss in early adolescence. Genital, anal, or oral trauma without a clear accidental explanation. Signs of chronic somatic stress — recurrent headaches, abdominal pain, sleep disturbance — without a workup that resolves them. Tattoos or branding (a name, an initial, a barcode, a crown, "Daddy") on the neck, chest, lower back, or thigh, especially if the patient is evasive about who did it or why.
- Behavioral signs in the visit. Scripted or rehearsed answers. Inability to recall details that should be easy — current address, school name, dates. Deference to an accompanying adult for answers to medical questions. Flat affect, dissociation, or sudden hypervigilance when a particular topic is raised. Reluctance or refusal to be examined alone. A patient who tells you the accompanying adult is her "boyfriend" or "uncle" or "manager" and is older than the relationship would suggest.
- Social-context signs. Sudden unexplained income, new phone, new clothing the family cannot account for. Multiple cell phones. School absences accelerating. Running away or staying with adults whose relationship to her she cannot explain. History of foster care, group-home placement, or chronic family instability — all of which significantly raise baseline risk. Hotel key cards. Travel patterns that do not match the patient's stated activities.
- Disclosure signs. The patient describes an "older boyfriend" who is more than three years older and exerts control over her movements, her communication, her clothing, or her money. The patient describes feeling "lucky" to have been "picked" by an older adult. The patient minimizes coercive details — "he just gets a little jealous sometimes" — and is invested in the relationship being framed as love.
- Bias-failure mode (the one most likely to be missed). Black, Native, and Latina adolescent patients presenting with the same constellation above are statistically less likely to have it recognized as exploitation. The clinical question to ask yourself: if this exact patient presented in front of me as a white twelve-year-old from an affluent family, would I be more concerned? If yes, I am underrecognizing this patient. Adjust.
The clinical encounter
The goal of the encounter is not to extract a disclosure in one visit. The goal is to create enough safety in the visit that the patient knows you are a person she can disclose to — now, or next time, or never, but at least the door is on the table.
Three structural moves matter more than any single question.
First: get the accompanying adult out of the room. Routinely. For every adolescent patient. Not as a flag, but as policy. "We see all of our adolescent patients alone for part of the visit — that is our practice for everyone." This neutralizes the situation; it does not single the patient out, and it does not tip off a trafficker that you are suspicious. Document the time the patient was seen alone in the chart.
Second: ask in the way that opens, not the way that closes. Closed yes-or-no questions about abuse get yes-or-no answers and end conversations. Open-ended questions about safety and relationships get information. The HEADSS framework (Home, Education, Activities, Drugs, Sex, Safety) — covered in detail in PG4 — is the structural backbone; the language is what makes it work.
Third: name the door. Tell the patient explicitly, before you ask anything sensitive, what your confidentiality framework is and what your reporting obligation is. Patients who have been groomed by adults are extremely good at reading whether a person can be trusted. They will read you. Be honest with them about what you can and cannot keep private, and they will know whether you are a person they can come back to.
The patient who taught me how to do this was fifteen. She had been seen in our clinic four times in nine months — twice for STI testing, once for an emergency contraception request, once for what was charted as "abdominal pain, likely functional." Different residents had seen her each time. Nothing was wrong. She had been triaged through the standard adolescent visit each time and the resident had documented the relevant negatives and she had walked back out.
The fifth time, I was the attending. I read her chart backwards. Four visits, no continuity, no one had ever seen her without her cousin in the room, and the cousin was twenty-six. I asked the cousin to step out because we always do that part alone. She told me he was not her cousin.
That is the part the textbooks do not tell you. The reason she had been missed four times was not that the signs were not there. The signs had been documented in the chart, by four different providers, none of whom had read the prior notes long enough to see the pattern. The infrastructure of recognition is continuity of care and time to read a chart. When those are missing, exploitation walks in and walks back out, and we are the ones who let it.
I tell residents this now. When you read a sixteen-year-old's chart and she has been in three times in eight months, you are not looking at a frequent flyer. You are looking at a patient who is trying to be seen.
What to document
- The patient's own words, in quotation marks. Not your paraphrase. "Patient states 'my boyfriend is twenty-six and he gets mad when I take too long at appointments.'" Direct quotation matters legally and clinically — it is contemporaneous, it carries evidentiary weight, and it does not require you to have assessed whether her language was accurate.
- What you observed. Affect, demeanor, body language, evasions, who was in the room, who answered which questions, how long the patient was seen alone, what the patient's reported relationship to accompanying adults was. Behaviorally specific. Not "appeared anxious" alone — "patient flinched when accompanying adult re-entered the room, made minimal eye contact for the remainder of the visit, declined to make eye contact with that adult."
- Your clinical concerns, named and justified. "Concern for adolescent sexual exploitation given [specific findings]. Plan: [report made / report deferred pending [specific next step] / consultation with social work / safety planning]." Document the reasoning, not just the conclusion. If you decided not to report, document why; if you decided to report, document the trigger.
When and how to refer
Three referral pathways exist for adolescent exploitation, and they are not interchangeable.
If the patient is in immediate danger — actively being controlled by an accompanying adult, presenting with acute injury, or describing imminent harm — your obligation is to ensure she does not leave with the trafficker. This may require alerting hospital security, calling child protective services from the room, or arranging a clinical reason for an admission. Do not confront the suspected trafficker. Do not signal to him that you have concerns. Get the patient to a place where she can be seen alone and supported, and make the report from there.
If the patient is in ongoing danger but not in immediate crisis — exploitation appears active, but the patient is medically stable and the trafficker is not in the room — your obligation is to make a report to child protective services (if the patient is a minor) and to call the National Human Trafficking Hotline (1-888-373-7888) for case consultation. The hotline does not replace your mandatory report; it supplements it with trafficking-specific safety planning that CPS frequently cannot provide. See PG3 for the legal architecture of the report itself.
If the patient is medically stable, in no immediate danger, but discloses historical or ongoing exploitation — your obligation is the same mandatory report, plus warm-handoff referral to a trauma-informed clinician, plus connection to survivor-services resources. See PG2 for the language of that referral and R1 for the verified directory of resources.
For your own care
You are going to miss patients. I have missed patients. Every provider who has been doing this work long enough has charted a visit and walked out of the room and learned six months or six years later that the patient was being exploited and you did not see it. That is not negligence. That is the operational reality of trying to do recognition work inside a fifteen-minute visit with no continuity of care and a chart you barely had time to skim. Holding yourself accountable for the structural conditions you did not create will burn you out and stop being useful to anyone, including the next patient.
What is yours to hold is what you do with the recognition once it is on the table. The first time you make a mandatory report on an adolescent patient — especially when the disclosure is partial, the evidence is suggestive but not definitive, and you are second-guessing whether you saw what you think you saw — you will feel sick. That is not a sign you reported wrong. That is a sign you reported on a child whose life is going to be hard for a while, and you are aware of it. The feeling does not mean you should not have made the call. It means you are a clinician who has registered the weight of what you just did.
The practice I use after a hard adolescent visit, and the practice I tell residents to use, is the same one I use after a bad outcome on labor and delivery. I do not debrief in the parking lot alone. I do not take it home and not name it. I find a colleague — a peer, not a junior — and I say out loud what I saw, what I did, and what I am worried I missed. Not for advice. For acknowledgment. Sometimes I cry. Sometimes I do not. Both of those are fine.
The trap with adolescent exploitation cases is that the patient never quite leaves you. You will think about her at three in the morning two months later. You will wonder whether the report you made got her somewhere safe or got her killed. You will not know. You have to be able to function in the next clinic anyway. Which means you have to have done the acknowledgment work after the last one. The work does not happen later. It happens that day. With another human being who knows what the work is.
That is what I tell residents. Find your person. Do not do this alone. The patient cannot be held by a clinician who is not being held.