ACR Bulletin

Covering topics relevant to the practice of radiology

Reading the Signs

Jump to Article
Radiologists can, and should, add value to the care of intimate partner violence patients — in and beyond the reading room.

Recognizing the signs of abuse and sharing your findings with other clinicians can change a life — maybe save a life.

—Annie Lewis-O’Connor, NP, PhD
January 13, 2020

“Sadly, you may see patients who die as a result of intimate partner violence (IPV),” says Annie Lewis-O’Connor, NP, PhD, founder and director of the Coordinated Approach to Resilience and Empowerment Clinic at Brigham and Women’s Hospital in Boston. “Recognizing the signs of abuse and sharing your findings with other clinicians can change a life — maybe save a life.”

Raising awareness of what the Centers for Disease Control and Prevention (CDC) calls “a serious and preventable public health problem” should be the goal of all clinicians, Lewis-O’Connor says. Relying on the expertise of radiologists to help identify patients who present with injuries commonly associated with IPV can be a powerful tool when reaching out to women and men in abusive relationships.

IPV is defined by the Violence Prevention arm of the CDC as “abuse or aggression that occurs in a close relationship,” past or present. While IPV encompasses abuse beyond physical injury — including sexual assault, stalking, and psychological aggression (using verbal or non-verbal communication to harm or gain control over another) — radiologists are in a unique position to identify the signs (and patterns) of physical injuries that may suggest IPV.1,2

Unbiased Witnesses

While considerable research has focused on abuse cases of children and the elderly — leading to increased training and prevention efforts — relatively little literature focuses on how radiologists can play a larger role in helping IPV patients.

“Radiologists are trained to simply report traumatic findings from the current examination without making any active effort to highlight any possibility of this life-threatening issue,” says Bharti Khurana, MD, radiology fellowship director for emergency MSK radiology at Brigham and Women’s Hospital and assistant professor at Harvard Medical School.

Fostering awareness of IPV and familiarizing radiologists with the most common imaging findings of abuse can aid in proper diagnosis and better patient care. Radiologists are often able to form unbiased conclusions based solely on imaging, without having direct contact with the victim or the abuser.2

“There is much to be done in terms of raising awareness among radiologists and physicians in general about IPV,” says Elizabeth George, MD, neuroradiology fellow at University of California, San Francisco, and one of the authors of a defining IPV study while chief radiology resident at Brigham and Women’s Hospital. There is a need for more multidisciplinary research to integrate clinical and imaging data — and to create robust systems for the identification and ongoing care of IPV victims, she says.3

“It is being increasingly recognized that radiologists have a significant role to play in the identification of IPV,” George says. “We have access to a wealth of information in the form of current and prior imaging,” she notes. “Equipped with this objective data, we can work closely with referring physicians and healthcare clinicians as an unbiased witness to improve patient care.”

Telling Injuries

“Radiologists are only starting to understand the spectrum of imaging findings in IPV,” George notes. “IPV-related findings have not yet been part of radiologists’ training.” Victims of IPV receive more imaging studies and have a higher frequency of potential violence-related imaging findings when compared with control subjects of the same gender and age range.3 “A lot of the injuries are usually distal on the body, often signally defensive injuries,” Lewis-O’Connor says. “If you are being punched, you are going to put your hand up to protect yourself. If you are being kicked in the abdomen, you are going to pull your legs up. These injuries are red flags for me.”

“The face is considered a target area, especially mid-face contusions and periorbital fractures. In the presence of defensive injuries, such as forearm or hand fractures or contusions, the likelihood of these injuries due to violence becomes high,” Khurana says. “By recognizing the high imaging utilization, location, and imaging patterns specific to IPV — as well as old injuries of different body parts on prior studies and injuries inconsistent to the history — the radiologist can generate an objective report,” she says.

“We are already trained to identify these injuries in isolation,” George notes. “Understanding the pattern of associated and prior injuries — and being mindful of them until it becomes routine — will help us put IPV detection into practice.”

Radiologists can, and should, add value to the care of IPV patients — in and beyond the reading room. “What might at first glance seem to be an accidental injury, on careful review of additional and prior findings, could be indicative of ongoing nonaccidental trauma,” George says. By developing expertise in IPV recognition, having discussions with referring providers, and understanding the coordinated care that follows, radiologists can further the goal of patient-centered care and make a life-changing difference for their patients.

Any type of injury can happen because of IPV. But if there are specific findings that we can give the probability for, we can increase radiologists’ role — and give them the confidence to make the invisible visible.

—Bharti Khurana, MD

Untidy Circumstances

Motivation and diligence will not go unchallenged, however. IPV continues to be profoundly underdiagnosed, mainly due to a lack of early detection which can result from the reluctance of victims to report it to healthcare providers. Screening with IPV in mind can lead to the detection of characteristic injuries or patterns that may inform a conversation that prevents future violence.4

The burden of identifying IPV is not the sole responsibility of radiologists, but falls on the healthcare team when a study shows injuries consistent with IPV, says Lewis-O’Connor. While conducting a team huddle, it may become clear that imaging results don’t match up with the patient’s history. Plus, radiologists may find healing injuries the referring provider didn’t know were there.

When IPV is suspected, all members of the healthcare team must be extremely mindful of a patient’s situation — even when they have the patient’s best interests in mind. Only a handful of states in the country allow or require reporting of IPV, Lewis-O’Connor says. You can ask patients questions related to their situation — and ask if they want help. In the majority of states, law enforcement can’t be called unless the patient requests it. Allow the patient to self-determine, provide choices, and respect their decisions.

While many lives are lost to IPV each year nationwide, Lewis-O’Connor says, pursuing a suspected case could ultimately make things worse for the victim once they leave a healthcare setting. Thus, providing a safe space in a non-judgmental manner allows patients to engage in the future.

Many cases of IPV go unreported by victims because of feelings of guilt, shame, or fear of reprisal — especially against their children, who are also at risk. The overwhelming majority of IPV patients are women, and Lewis-O’Connor notes that she has seen many come in soon after having a child.

Concerns of patients are real — “What happens if there’s not enough evidence to arrest an abuser, but the abuser finds out it was reported?” she asks. “What if the victim fears for her child or depends on the abuser for housing, food, or money?” Considering reporting is complicated and some find it more harmful than good, she says, you have to be careful when explaining options to patients and listen without prescribing. “It’s not as tidy as everybody would like it to be,” she says.

To put that into perspective, Lewis-O’Connor says that during her career she has had two patients die of breast cancer and three murdered as a result of IPV. Knowing that the worst can happen may prompt healthcare providers to share findings with other clinicians and hospital social workers, she says, so that potential victims are offered timely assistance.

Promising Inroads

While identifying victims of abuse is arguably the biggest challenge in combating IPV, opportunities exist to connect with patients.

Researchers at Massachusetts General Hospital (MGH) have explored integrating IPV screening when women present for breast imaging or annual mammograms. Women are given a questionnaire posing questions such as, “Do you feel safe at home?” or “Do you feel safe in your relationship?”

If women indicate not feeling safe at home, they are referred to the institution’s Helping Abuse and Violence End Now (HAVEN) program that is located on campus or provided with contact information for offsite HAVEN centers.5

While this type of patient self-reporting can have positive outcomes, providers need more guidance on IPV, according to Khurana. Screening questions can motivate a patient to disclose information, but if a patient decides not to, a provider might not raise their own concerns about IPV. “Right now we are essentially depending on patients’ self-reporting,” she says. “Even if a patient does not disclose IPV, services and safe numbers can be provided as part of universal education,” adds Lewis-O’Connor.

If a past injury shows up on new imaging, a greater awareness of IPV might prompt a radiologist to raise questions about abuse. However, Khurana believes that expecting radiologists to seek out IPV findings and then raise concerns with the appropriate clinician or healthcare support staff is not realistic without some kind of systemic help.

Learning Patterns

Along with a team of multispecialty physicians from Brigham, MGH, Harvard School of Public Health, and other institutions, Khurana is now leading an effort to use machine learning to narrow findings that suggest the probability of IPV injuries and integrate those findings into radiology reporting systems.

“Our goal is to create a fully integrated, multidimensional clinical decision support tool that uses patterns derived from expert analysis of historical radiological and clinical data, classification models, and statistical evidence to classify injuries for their likelihood of being due to IPV,” she says. Clinicians would be automatically alerted if a patient’s injuries have low- or high-risk probability of IPV.

Providers may overlook the signs of IPV because of their unconscious bias toward a victim’s or abuser’s physical appearance, education level, or socioeconomic background. Research acknowledges that some healthcare providers can be hesitant to suggest IPV, often for fear of offending patients or their partners. The automated prediction of IPV based on historical radiological and clinical data could avoid such bias and help validate radiologists’ concerns.6

Khurana hopes her work with data scientists will lead to an alert system for radiologists based on patients’ imaging history. Using machine learning to recognize signs of IPV on current and prior images, the alert would provide a visualization of risk factors, empowering healthcare providers to open a dialogue with potentially at-risk patients. Once validated, Khurana hopes to make the algorithm accessible through
ACR’s Data Science Institute™ and integrate outputs into radiology reports.

“In addition, our multidisciplinary team plans to design conversational guides using medical images for training social workers and clinicians to approach patients identified as high-risk for IPV,” she says. Visually pointing out an injury on imaging studies to a victim may encourage them to talk about their situation.

Further research and training is needed to create awareness of IPV among radiologists who might be the first physician to suspect violence when presented with serial imaging studies. “We as a specialty should lead this work, educate ourselves, and increase awareness among our colleagues,” George says. “To make a meaningful impact in the multidisciplinary care of these patients, radiologists must work together with clinical colleagues in integrated groups.”

“IPV is so common, but these patients often get missed,” Khurana says. “Any type of injury can happen because of IPV. But if there are specific findings that we can give the probability for, we can increase radiologists’ role — and give them the confidence to make the invisible visible.”


1.Centers for Disease Control and Prevention (Violence Prevention). Preventing intimate partner violence. Accessed Dec. 10, 2019.
2. Bhole S, Bhole A, Harmath C. The black and white truth about domestic violence. Emerg Radiol. 2014;21(4):407–412.
3. George E, Phillips CH, Shah N, Lewis-O’Connor, Rosner B, Stoklosa HM, Khurana B. Radiologic findings in intimate partner violence. Radiology. 2019;291(1):62–69.
4. Boas G. The role of radiology in identifying intimate partner violence. Radiology Rounds. April 2019; Vol.17, Issue 4.
5. Narayan AK, Lopez DB, Miles RC, Dontchos B, Flores EJ, et al. Implementation of an intimate partner violence screening assessment and referral system in an academic women’s imaging department. J Am Coll Radiol. 2019;16:631–634.
6. Khurana B, Seltzer SE, Kohane IS, Boland GW. Making the ‘invisible’ visible: transforming the detection of intimate partner violence. BMJ Quality & Safety. 2019.

Author Chad Hudnall, senior writer, ACR Press