Non-fatal strangulation: documenting and reviewing the medical record when the neck shows little

Patients say "he choked me." Police reports repeat it. Triage nurses type it into the chief complaint. Clinically, that word usually describes something else, and the mismatch causes real problems when a prosecutor, a defense expert or a civil jury reads the chart months later.

The StatPearls chapter Strangulation Injuries (Dunn, Sukhija and Lopez, updated 2025) describes strangulation as injury from mechanical force applied externally to the neck, a form of asphyxia that can cut cerebral oxygen delivery by compressing the neck's blood vessels or occluding the trachea. It notes that "choking" is often used loosely for strangulation, but in medical contexts the term "more accurately refers to internal or partial occlusion of the proximal airway": a piece of food, not a hand. Suffocation is a third thing: blocking the nose and mouth from outside.

Strangulation

External pressure on the neck that impedes breathing, blood flow or both. Manual (hands or forearm) or ligature (cord, belt, clothing). Hanging is a form of it.

Choking

Obstruction inside the airway. In a chart, "choking" in quotation marks is usually the patient's word for strangulation; unquoted, it may be a clinician's word for a foreign body. Read the context before you code it.

Suffocation

Blocking the nose, mouth or both from outside, for example with a hand or a pillow. Federal law defines it separately from strangling.

Petechiae

Pinpoint hemorrhages in skin or mucosa, often looked for in the eyes, eyelids, face, mouth and behind the ears after neck compression. Common in the literature, inconsistent in practice, and not specific to strangulation (chapter 9).

Non-fatal or near-fatal strangulation

Strangulation the person survived. "Near-fatal" appears in the 2022 imaging recommendations for patients whose presentation suggests they came close to dying.

Lethality assessment

A structured screen for the risk of future severe or fatal violence by an intimate partner. The 2022 imaging recommendations call for 1 at discharge.

The federal definition

Congress added strangulation and suffocation to the federal assault statute in the Violence Against Women Reauthorization Act of 2013. 18 USC 113(a)(8) punishes assault of a spouse, intimate partner or dating partner "by strangling, suffocating, or attempting to strangle or suffocate" with up to 10 years in prison. The statute applies within federal maritime and territorial jurisdiction, so most strangulation prosecutions happen in state court under state law. Its definition is still the clearest statement of what the records have to show:

"the term 'strangling' means intentionally, knowingly, or recklessly impeding the normal breathing or circulation of the blood of a person by applying pressure to the throat or neck, regardless of whether that conduct results in any visible injury or whether there is any intent to kill or protractedly injure the victim"

Read that clause as a records map. "Impeding the normal breathing" points to the patient's account of breathing difficulty and to anything a clinician observed. "Circulation of the blood" points to symptoms of reduced blood flow to the brain: vision changes, dizziness, loss of consciousness. "Pressure to the throat or neck" points to the account of the method and any neck findings. And "regardless of whether that conduct results in any visible injury" tells you why a normal-looking neck does not end the inquiry. Subsection (b)(5) defines "suffocating" the same way for covering the mouth, nose or both.

The state pattern

The Training Institute on Strangulation Prevention, which tracks this legislation, reports on its legislation page that all 50 states now have felony strangulation laws, and that the laws vary from state to state. They vary in the words that matter to a records review: whether the statute says "impeding" or "obstructing," whether it names breathing, blood circulation or both, whether it reaches suffocation, whether it requires a domestic or dating relationship, and whether it says anything about visible injury. Pull the text of your state's statute before you build the file, and build the element table in chapter 9 from its exact words.

The modern literature on surviving strangulation starts with a records review. In 2001, the Journal of Emergency Medicine published a 3-part series from San Diego on 300 strangulation cases submitted for misdemeanor prosecution to the San Diego City Attorney's Office. Part I (Strack, McClane and Hawley) looked for signs and symptoms that could corroborate a victim's report of being "choked." Its abstract states the central finding plainly: "most victims of strangulation had no visible injuries or their injuries were too minor to photograph," and the authors concluded that police and prosecutors, relying too heavily on visible signs, missed chances for higher-level prosecution. Part II (McClane and colleagues) proposed a clinical protocol for the surviving patient.

Later work in other settings points the same way. In a Western Australian study of 1,064 women examined after recent sexual assault (Zilkens and colleagues, 2016), 79 (7.4%) reported non-fatal strangulation during the assault, and external physical signs were absent in 49.4% of them.

Petechiae depend on where and when someone looks

Petechiae are the classic sign in textbooks. In practice they are uncommon and they fade. A 2026 study from Charité in Berlin (Babigian and colleagues) reviewed 541 adult forensic examinations after non-fatal strangulation. Petechiae were found in 54 (10%). Examinations that found them happened sooner after the event, a mean of 26 hours against 61 hours for those that did not. Conjunctival petechiae were the most common early finding, while petechiae behind the ear persisted longer. Dyspnea, hoarseness and dysphagia were associated with finding petechiae.

2 practical points follow. First, the time between the assault and each examination belongs in the chronology, because a later exam that finds nothing is weaker evidence of absence than an early one. Second, check the location list on the form. An exam that looked at the conjunctivae but not behind the ears, or not inside the mouth, looked at part of the map.

The reason prosecutors, advocates and emergency physicians treat a strangulation history as urgent comes largely from 1 study. Glass and colleagues (J Emerg Med, 2008) used a case control design. They compared 506 women who were killed or nearly killed by a partner with 427 women who had been abused but not killed, using data gathered with the Danger Assessment instrument.

The odds ratios were 6.70 for attempted homicide (95% confidence interval 3.91 to 11.49) and 7.48 for completed homicide (95% confidence interval 4.53 to 12.35). The authors concluded that non-fatal strangulation is an important risk factor for homicide of women and that emergency clinicians should screen for it. The 2022 imaging recommendations separately call for a lethality assessment at discharge.

Where the risk data shows up in the file

• Lethality or danger screens. Police officers, advocates and forensic nurses often complete a structured screen at the scene or at discharge. It may sit in the police report, the advocate's file or the forensic nurse packet rather than in the hospital chart.
• Bail and protective order papers. A prior strangulation history is often cited in release conditions and petitions. Those documents quote the records, so check each quote against its source page.
• Prior encounters. Earlier ED visits, clinic notes and police calls can show a history the patient did not repeat at the index visit, or a history that appears for the first time after the index visit.

What the numbers can and cannot carry

Both sides should read the Glass study for what it is: a population comparison of women who had already been killed or nearly killed with women who had not. It shows that a strangulation history was far more common in the first group. It does not show that any given person strangled anyone, and it does not predict what a specific person will do. Prosecutors use the data to explain urgency, charging and release conditions. Defense counsel may object when risk data is offered to suggest guilt for the charged event, under the rules on prior acts and unfair prejudice in the jurisdiction. The record reviewer's job is simpler: find every place a strangulation history appears, note who recorded it, when, and in whose words.

When the neck looks normal, the symptoms carry the documentation. The 2022 recommendations from the Training Institute on Strangulation Prevention (Smock, Green and Sturgeon, revised November 2022) list the history items that should prompt imaging: loss of consciousness, visual changes such as "spots," "flashing lights" or "tunnel vision," altered mental status, breathing changes, bladder or bowel incontinence, neurologic symptoms (seizure-like or stroke-like symptoms, headache, tinnitus, decreased hearing, numbness, amnesia), neck pain, sore throat or pain on swallowing, and voice change, from hoarse or raspy to unable to speak.

How often do these appear? Bergin and colleagues (J Head Trauma Rehabil, 2022) reviewed 345 women seen at a community emergency department after non-fatal strangulation by an intimate partner between 2008 and 2016.

These rates describe 1 hospital's charts, which means they also describe what that hospital's clinicians asked about and wrote down. A symptom missing from a chart may be absent, or it may never have been asked. That distinction drives most of the disputes in chapter 9.

Documentation checklist for a strangulation encounter

• Patient's own words for the methodhands, forearm, ligature, which hand, from front or behind, in quotation marks
• Number of times and approximate durationas the patient reports it, with "unsure" recorded when that is the answer
• Time of the event and time of each exampetechiae and redness change with time
• Loss of consciousness asked directlyincluding memory gaps and waking up in a different position
• Vision, hearing and neurologic symptomsspots, tunnel vision, tinnitus, numbness
• Voice, swallowing and breathinga functional assessment, beyond a review of systems checkbox
• Incontinence asked directlypatients rarely volunteer it
• Petechiae sites listedeyes, eyelids, face, mouth, behind the ears, neck, even when negative
• Photographs with scale and a photo loginitial and follow-up
• Imaging decision and reasonordered, declined by the patient, or not indicated, with the reason written
• Lethality assessment resultand where the form is kept
• Discharge instructions and return precautionsneurologic warning signs, breathing and swallowing changes

0 of 12 checked

The clinical concern after strangulation is injury that is not visible: dissection or thrombosis of the carotid or vertebral arteries, which can lead to stroke hours or days later, and damage to the larynx, hyoid or other neck structures. The 2022 Training Institute recommendations set 3 goals for the evaluation: rule out acute conditions, evaluate the carotid and vertebral arteries for dissection or thrombosis, and evaluate the airway and the bony, cartilaginous and soft tissue structures of the neck.

2 lines in the document come up often in litigation. First, "observation has no role in ruling out a vascular injury"; observation is reserved for airway concerns. Second, the recommendations apply to delayed presentations up to 1 year after the event. They also suggest considering 325 mg of aspirin if imaging will be delayed, and fiberoptic laryngoscopy when the airway is in question. The document was supported by an Office on Violence Against Women grant, which defense counsel sometimes raise when the recommendations are offered as a standard of care. Whether any protocol defines the standard of care is a question for experts and the court.

What the imaging studies found

Emergency physicians have debated whether routine CT angiography is worth its cost and radiation when the yield is low. The published series do not agree on the answer, partly because they studied different populations.

Read together, the studies show that serious vascular injury is uncommon but real. Matusz and colleagues wrote that their data "suggest, but do not prove, that a selective imaging strategy is safe." Zuberi and colleagues found CTA "rarely identifies clinically significant findings." Khan and colleagues, in a mostly hanging population, advised imaging everyone because no clinical indicator safely excluded injury. An earlier review by Vilke and Chan (2011) noted that carotid dissection after strangulation is rare but documented, and that CTA belongs in the workup when it is suspected.

How each side reads a negative CTA

A negative CTA rules out the vascular injuries the scan can see. It says nothing about whether pressure was applied to the neck, because most strangulation leaves no imaging finding at all. A prosecutor should expect the defense to point to a normal scan, and should be ready to explain what it was looking for. Defense counsel should expect the prosecution to point out that a normal scan is the common result. In a civil claim against a hospital, the question shifts: was imaging indicated, offered and documented, and if not, why not?

A strangulation file is built from at least 6 record holders, and no single request reaches all of them. The hospital chart is the one everyone asks for first. It is often the least useful piece on its own, because the forensic nurse exam, the photographs and the imaging studies may sit in separate systems, and the most immediate evidence of voice and breathing sits with the 911 center and the police.

How the records move

For the patient and her civil counsel, HIPAA gives a right of access to her own records: 45 CFR 164.524 requires a covered entity to act on a request within 30 days, with 1 extension of up to 30 more. Separately, 45 CFR 164.512(c) allows a covered entity to disclose information about a person it reasonably believes is a victim of abuse, neglect or domestic violence to a government authority authorized to receive such reports, when the law requires it, when the person agrees, or under narrower conditions. Prosecutors usually obtain hospital records by subpoena or court order. Defense counsel usually receives them through criminal discovery, or seeks them by subpoena, and many courts review a complainant's medical records privately before releasing them. The rules differ by state. The practical guide to requests is in how to get medical records for a lawsuit.

Many forensic nursing programs use a strangulation specific form alongside the general exam record. Layouts differ, but most ask the same questions in roughly the same order. The mock form below is a composite built for this guide. It shows the fields that matter and what to check in each.

P. calls 911 late on a weeknight. By morning the file holds a CAD log, a police report with photographs, an EMS report, an ED chart, a CTA report, a forensic nurse exam and a discharge summary. 9 days later she sees her primary care physician. Here is the file laid out in time order, with the gaps marked.

Where each finding first appears

The addendum

The ED physician signed a note at 01:32, before the forensic nurse exam. At 06:05 the physician added a labeled addendum. Properly labeled and timed, an addendum is ordinary practice. Each side will still ask what the physician personally saw and what was copied from the forensic nurse's findings.

Here the addendum says "per forensic nurse exam," which answers the question on the page. When an addendum does not say where its new facts came from, the audit trail can show who opened which record and when. The EHR audit trail guide and the altered medical records guide cover how to request and read one.

The same records support opposite readings, and a careful reviewer on either side should be able to state the other side's case in its strongest form. Start with the elements. Using the federal definition as a template, and substituting your state statute's words, each element points to specific records.

Petechiae and voice change have other causes

Defense experts often point out that petechiae are not specific to neck compression. In a series of 37 electrocution deaths (Karger and colleagues, 2002), petechiae were found in 74%, and the authors called the finding non-specific. A published case report (Garland and Tse, 2017) described neck injury and conjunctival petechiae in a death explained by resuscitation and sepsis rather than strangulation, and another (Lambe and colleagues, 2009) described extensive petechiae after a self-inflicted ligature. Those are postmortem and single case reports, and prosecutors will note that none involves a living patient describing a partner's hands on her neck. The point for a reviewer is narrower: record the finding, the site, the time and the examiner, and leave its cause to the experts. Hoarseness has many causes too, which is why persistence into follow-up visits, and any ENT findings, matter.

Hearsay and confrontation

Much of a strangulation record is the patient's own statements. Federal Rule of Evidence 803(4) excepts from the hearsay rule a statement made for, and reasonably pertinent to, medical diagnosis or treatment that:

"describes medical history; past or present symptoms or sensations; their inception; or their general cause."

Symptoms such as "I couldn't breathe" and "I blacked out" sit comfortably within the rule. Whether a statement naming the person responsible is pertinent to treatment is a recurring dispute, and courts answer it differently. Forensic nurse exams raise a second question under the Confrontation Clause. After Crawford v. Washington, 541 U.S. 36 (2004), testimonial statements of a witness who does not testify at trial are barred unless the witness is unavailable and the defendant had a prior chance to cross examine. In Davis v. Washington, 547 U.S. 813 (2006), the Court held that statements to police whose primary purpose is to meet an ongoing emergency are nontestimonial, while statements whose primary purpose is to establish past events for later prosecution are testimonial. A 911 call, an EMS narrative and a forensic nurse interview can fall on different sides of that line, and state courts have reached different results for forensic exams. The reviewer's contribution is factual: who asked, why, when, with police present or not, and whether treatment was provided.

Civil and family matters

The same records appear in protective order hearings, custody and parenting disputes, and civil claims for assault. In family court the standard of proof and the rules of evidence are often looser, and records are sometimes filed whole. See custody evaluation record review and domestic violence injury record review for how those files are organized.

A strangulation file is a good fit for AI medical record review in 1 narrow sense: the evidence is scattered across many short documents from different authors, and the main job is finding and lining up small statements. It is a poor fit in another sense: the details that matter most, such as a quoted phrase, a checkbox marked "unsure" or a photo of a pinpoint hemorrhage, are exactly what a careless tool flattens or invents.

Where software earns its place

• Cross-source symptom search. AI document review and clinical natural language processing (NLP) can find every mention of voice, swallowing, breathing, consciousness, vision and incontinence across a CAD log, a police report, an EMS report and a hospital chart, including synonyms such as "raspy," "blacked out" and "wet herself."
• A cited timeline. An AI medical chronology that puts each entry next to its source page lets both sides check the reading in seconds. Page-level citation is the minimum; without it, the output cannot be used.
• Conflicts and gaps. Medical chronology software can line up the triage template's "denies difficulty swallowing" against the forensic nurse's "painful swallowing," and flag a referral with no matching records.
• Scans and handwriting. Forensic forms and body diagrams are often handwritten and scanned. OCR built for handwritten notes, plus human checking of every extracted checkbox, turns them into searchable text. See can AI read handwritten medical records.

Where it fails

• Paraphrase. A large language model (LLM) asked to summarize will turn "I couldn't breathe" into "reported dyspnea." The quote is the evidence. Require verbatim quotes for every statement by the patient.
• Templated negatives. Generative AI treats a templated "denies" the same as a free text answer. It cannot tell whether the question was asked.
• Ambient AI scribe notes. Notes drafted by an ambient AI scribe and signed by a clinician may smooth over the patient's words or omit an uncertain answer. Compare them against other sources, and note in your chronology that the note was scribe drafted if the chart says so.
• Photographs and images. Text tools do not read photographs of the neck or eyes, and no tool should be trusted to call petechiae from a photo. Imaging needs a radiologist. See can AI read DICOM imaging.
• Hallucination. A model can invent a symptom, a time or a page. In Mata v. Avianca, Inc., 678 F. Supp. 3d 443 (S.D.N.Y. 2023), lawyers were sanctioned under Rule 11 for filing fabricated case citations produced by a chatbot. The same discipline applies to fabricated clinical facts. See is AI accurate enough for court.

What to require from any tool

Legal AI tools vary widely. For a file that holds a complainant's sexual, mental health and injury history, the bar is higher than for a routine contract review. Rank your requirements in this order:

Agentic tools that chain steps on their own, such as requesting records or drafting letters, add a further risk in a case with a protected complainant: an automated step that sends the wrong document to the wrong party. Keep a human at every step that leaves the office.

Each template works for the prosecution, the defense or civil counsel. Replace the bracketed fields, and adapt the record list to your jurisdiction's discovery and subpoena rules.

STRANGULATION RECORDS CHECKLIST
Matter: [CASE NAME / NUMBER]
Patient: [INITIALS]   Event date: [DATE]   Reviewer: [NAME]

PRESERVE NOW (short retention)
[ ] 911 audio and CAD event log            Holder: [COMMUNICATIONS CENTER]
[ ] Body camera and dash camera video       Holder: [AGENCY]
[ ] Scene and neck photographs, native files with capture times

REQUEST BY NAME
[ ] Police report, supplements, written and recorded statements
[ ] EMS patient care report, including narrative and vital signs
[ ] Emergency department chart: triage, physician, nursing, orders,
    addenda with signature times, discharge instructions
[ ] Forensic nurse examination: strangulation supplement, body
    diagrams, photo log, all photographs, evidence collection log
[ ] Imaging reports AND images (CTA, MRA, CT head), native DICOM
[ ] Follow-up: primary care, ENT, neurology, speech therapy
[ ] Lethality or danger screen (check privilege rules first)
[ ] Prior encounters: ED visits, clinic notes, police calls
[ ] EHR audit trail for any note with a late addendum

TRACK
Request sent: [DATE]   Due: [DATE]   Received: [DATE]
Photo count listed: [N]   Photo count produced: [N]
Referrals placed with no matching records: [LIST]

SYMPTOM AND SIGN WORKSHEET
Codes: R = reported (quote it)  D = denied  NA = not asked
       NE = not examined  + = finding present  0 = examined, absent
Record the page cite for every entry.

SYMPTOM        | 911/CAD | POLICE | BODYCAM | EMS | TRIAGE | ED MD | FNE | FOLLOW-UP
Voice change   |         |        |         |     |        |       |     |
Swallowing     |         |        |         |     |        |       |     |
Breathing      |         |        |         |     |        |       |     |
Loss of consc. |         |        |         |     |        |       |     |
Vision change  |         |        |         |     |        |       |     |
Incontinence   |         |        |         |     |        |       |     |
Neuro symptoms |         |        |         |     |        |       |     |
Neck pain      |         |        |         |     |        |       |     |

SIGN           | Examiner | Time of exam | Sites listed | Result | Photo #
Petechiae      |          |              |              |        |
Neck redness   |          |              |              |        |
Tenderness     |          |              |              |        |

Hours from event to each exam: [LIST]
Method as described in each source (verbatim): [LIST]

You are assisting with a medical record review. Use only the
documents provided. Do not use outside knowledge.

Task: Build a table of every statement or finding about the
following, across all documents: voice change, swallowing,
breathing difficulty, loss of consciousness or memory gap,
vision change, incontinence, neurologic symptoms, neck pain,
neck findings, petechiae (with the sites examined), imaging
results, and the method of the alleged neck pressure.

For each row give:
1. Date and time of the entry
2. Document name and page number
3. Author role (dispatcher, officer, EMT, nurse, physician,
   forensic nurse, other)
4. The exact words, in quotation marks, with no paraphrase
5. Whether it is a patient statement, a clinician observation,
   or a templated field
6. For templated negatives, write "TEMPLATE" and do not treat
   them as the patient's answer

Then list:
A. Any symptom recorded as denied in one source and reported
   in another, with both page cites
B. Any referral, photo count or cross reference with no
   matching record in the set
C. Any item marked not examined or left blank

If a fact is not in the documents, write "NOT FOUND".
Do not infer, diagnose or assess credibility.

Medrecords AI builds a cited chronology from the records you upload. On a strangulation file that means the CAD log, police report, EMS report, ED chart, forensic nurse exam, radiology report and follow-up notes land in 1 timeline, with every line tied to its source page.

• Chronology with citations. Every entry links to the page it came from. See chronology and citations.
• Scans and handwriting. Handwritten forensic forms and body diagram notes are read with handwritten record extraction and OCR, and every extracted field stays open to your review.
• Missing records flags. Referrals with no matching records and cross references to documents not in the set are flagged. See missing records identification.
• Search and questions. Find every mention of a symptom across sources with medical record search and ask cited questions with record Q&A.
• Redaction. Suggested redactions for identifiers and third party information, approved by you, with redaction.

What it does not do: it does not retrieve records, decide whether strangulation happened, interpret photographs or imaging, assess anyone's credibility or decide a case. A flag is a signal for a reviewer, not a verdict. Your experts and the fact finder make those calls.

Medrecords AI holds SOC 2 and supports HIPAA with a signed BAA; see security and HIPAA. Self-Service bills 10 cents a deduplicated page, down to 5 cents at volume, duplicates free; Enterprise On-Prem is an annual license.

What records are needed for a non-fatal strangulation case?

911 audio and the CAD log, the police report with statements and photographs, body camera video, the EMS report, the full emergency department chart, the forensic nurse exam with its photographs and body diagrams, imaging reports and images, follow-up visits, and prior encounters. Request the forensic exam, photographs and images by name.

Can strangulation be proven with no visible injury?

The federal definition in 18 USC 113(b)(4) applies regardless of visible injury, and research has found that many victims show little or none. Whether the evidence is enough in a given case depends on the state statute, the whole record and the fact finder.

What symptoms should be documented after strangulation?

Voice change, painful or difficult swallowing, breathing difficulty, loss of consciousness or memory gaps, vision changes, incontinence, neurologic symptoms such as headache or numbness, neck pain, and any petechiae with the sites examined and the time of the exam.

Is a CT angiogram required after strangulation?

The 2022 Training Institute recommendations call CTA of the carotid and vertebral arteries the gold standard when listed symptoms are present. Emergency medicine studies disagree on how selectively to image. Whether imaging was indicated in a specific case is an expert question.

Why does the time of the forensic exam matter?

Findings fade. In a Berlin series of 541 exams, those that found petechiae took place a mean of 26 hours after the event, against 61 hours for those that did not.

Are statements to a forensic nurse admissible?

It depends on the jurisdiction and the facts. They may fall within the medical diagnosis or treatment hearsay exception, and in a criminal case they may also face a Confrontation Clause challenge that turns on the primary purpose of the exam.

Can AI summarize strangulation medical records?

AI can find and line up symptoms across sources and build a timeline cited to each page. It should quote the patient's words exactly, and a person must check every line. It should not interpret photographs, imaging or credibility.

Is it HIPAA compliant to upload strangulation records to an AI tool?

It can be, with a vendor that signs a business associate agreement, holds SOC 2, does not train on your data and logs access. A consumer chatbot without a BAA is the wrong place for a complainant's records.

Can ChatGPT build a strangulation case timeline?

A general chatbot can draft one, but without page citations, a BAA and verbatim quotes the draft is not usable, and it may invent details. Use a tool built for medical records and check each entry against its page.

Can software decide whether strangulation occurred?

No. Software can organize and cite the records and flag gaps. Whether strangulation occurred, what caused a finding and whom to believe are questions for experts, the court and the fact finder.

Study figures were checked against PubMed abstracts or full text in September 2026. The 2022 imaging recommendations were read in full from the Training Institute's published PDF. Statutes and rules were read on the Legal Information Institute site. Figures that could not be confirmed at a primary source were left out. The worked example, the mock form and the timeline are hypothetical. Nothing here is legal or medical advice.

• 18 USC 113, subsections (a)(8), (b)(4) and (b)(5), added by Pub. L. 113-4, section 906 (2013).
• Training Institute on Strangulation Prevention, Recommendations for the Medical/Radiographic Evaluation of Acute Adult/Adolescent, Non/Near Fatal Strangulation, Smock, Green and Sturgeon, revised November 2022.
• Training Institute on Strangulation Prevention, strangulation legislation: felony laws in all 50 states.
• Strack, McClane and Hawley, J Emerg Med 2001;21(3):303, and McClane and colleagues, part II.
• Glass and colleagues, J Emerg Med 2008;35(3):329 (full text).
• Zilkens and colleagues, J Forensic Leg Med 2016;43:1.
• Babigian and colleagues, Forensic Sci Int 2026;378:112730.
• Bergin and colleagues, J Head Trauma Rehabil 2022;37(1):5 (full text).
• Zuberi and colleagues, Emerg Radiol 2019;26(5):485; Matusz and colleagues, Ann Emerg Med 2020;75(3):329 (doi); MacDonald and colleagues, CJEM 2021;23(6):762; Khan and colleagues, World J Surg 2025;49(3):752.
• Dunn, Sukhija and Lopez, Strangulation Injuries, StatPearls.
• Federal Rule of Evidence 803(4); 45 CFR 164.512(c); 45 CFR 164.524.
• Crawford v. Washington, 541 U.S. 36 (2004); Davis v. Washington, 547 U.S. 813 (2006); Mata v. Avianca, Inc., 678 F. Supp. 3d 443 (S.D.N.Y. 2023).

Related guides: SANE exam records, blunt force trauma records, child abuse medical records, how to read an autopsy report, record review for expert witnesses and record review for legal nurse consultants.