Boston doctor builds tool to identify patients at risk for intimate partner violence
0 9 mins 2 mths


BOSTON – Dr. Bharti Khurana, an Emergency Radiologist at Brigham and Women’s Hospital was leaving a trauma “hackathon” in 2016 when a nurse practitioner began talking about Intimate Partner Violence. For all the advances in medicine, the NP told the group, there was little attention paid to the root cause of the injuries that sends many patients to the Emergency Department. Dr. Khurana stopped to listen. What she heard struck a chord. “I take it very personally. I’m an Emergency Radiologist. I’m a gatekeeper. I look at the imaging studies, decide if the patient should stay here and get admitted or discharged… For non-accidental trauma in children, we do such an excellent job. And then, when it comes to adults, we were not doing anything. So, I got motivated.” 

Dr. Khurana was already extolling the virtues of artificial intelligence, specifically in detecting fractures. While many radiologists viewed AI as a threat to their job security, she was eager to build a tool that would create an annotation for fractures even before radiologists looked at the images. But her focus began to shift as she wondered whether she could use AI to build a tool that would identify patients at risk for intimate partner violence. Six years later, she and her team (Dr. Khurana is the founder and director of the Trauma Imaging Research and Innovation Center) have done it. The tool, an automated clinical decision support tool for Intimate Partner Violence Risk and Severity Prevention (AIRS), uses radiological data and a patient’s clinical history. “On average,” she explains looking at an X-ray, “we can detect four years before the patient self-reports intimate partner violence.” She explains that because domestic violence tends to escalate over time, knowing earlier can protect patients against more severe injuries. Its accuracy is now roughly 80%. 

Obvious red flags

Obvious red flags include “target” and “defensive” injuries. An abuser’s target is often the victim’s head and face: orbital bones near the eye and cheekbones. Mid-facial fractures are, sadly, common in patients experiencing Intimate Partner Violence. Defensive injuries occur when patients try to protect themselves. Dr. Khurana explains that a broken forearm bone (ulna) near the pinky finger can be a sign that a patient has raised an arm to protect his or her face. By contrast, a broken arm bone (radius) near the thumb is common among patients who try to brace themselves for a fall. Radiological studies provided key data for the AIRS tool’s creation. But it also includes so much more. Information in a patient’s history (past ER visits, old fractures, medication, canceled screenings, etc.) that would take a long time for a busy radiologist to find and compile human being to compile is quickly gathered using the tool and provided-as an assessment-in real time while a healthcare provider reads the image. 

An urgent public health issue

Dr. Khurana built the tool on Brigham and Women’s patients and validated it on Mass General patients. With information from those who were reporting IPV and a group that was not, she discussed the findings with trauma surgeons, patient advocates, and an entire multi-disciplinary team to further develop the tool. Knowing that barriers to care exist for both patients and clinicians, she also developed conversation guides. While all patients are asked-verbally or on a questionnaire-whether they feel unsafe in their relationship, even patients experiencing violence will often answer “no.” 

Dr. Bharti Khurana
Dr. Bharti Khurana, an Emergency Radiologist at Brigham and Women’s Hospital.

CBS Boston


The AIRS tool provides objective information that can make it easier for a clinician to revisit that answer in a sensitive way. It is not an easy conversation. Dr. Khurana sympathizes with everyone involved. “We never got trained in identifying Intimate Partner Violence. Not in radiology and also not in medical school. To this day there are many physicians who consider this a social issue, not a medical issue.” There is no debate that IPV is an urgent public health issue. According to the CDC, at least one in four women and one in seven men will experience domestic violence in their lifetime. It is also involved in roughly half of all homicides against women in the U.S.

Stopping the cycle of abuse  

Survivors often deal with chronic, debilitating injuries including concussions. “We talk so much about them in football players. But we don’t talk about the micro concussions happening almost every day to these women,” Dr. Khurana says looking at an X-ray. “Because they cannot think right, they make more errors. They are losing their confidence. They are no longer working anymore. They are not getting help outside. They are completely dependent on their partners.” It is a vicious cycle that exacerbates a patient’s sense of shame and vulnerability. Dr. Khurana is confident that the AIRS tool’s ability to identify who’s at risk will help educate patients, stop the cycle of abuse, and save lives. “We do mammograms. That’s a screening study to diagnose breast cancer before cancer spreads to the entire body. When I talk to other radiologists, I say, ‘Think of Intimate Partner Violence as the same thing.’ It is going to destroy you completely. In fact, we are not just talking about the patient. The children who are witnessing IPV have adverse health effects too.”

“Empower guides”   

Dr. Khurana has learned many lessons developing the AIRS tool beyond the use of technology. Many patients don’t want to leave their relationship despite the abuse. With that in mind, she and her team have started a new project to predict future health risks associated with IPV including neurological problems, gastrointestinal issues, mental health disorders and substance use disorder. In addition to offering resources and a safety plan, healthcare providers will be able to treat these related conditions. It is important to note that patients are under no pressure. They may choose not to change anything about their lives or relationship. “All we are doing is providing a resource,” Dr. Khurana explains. “Even if they don’t disclose, we offer ’empower guides’ with AIRS. We talk to these patients about universal education. If they don’t want to talk, we offer QR codes.” The QR code makes it possible to explore resources without taking home any physical information that might anger an abuser. “Patient safety and privacy is critically important.”

IPV annotation does not appear in the MGB Patient Portal. Any alerts are saved in a patient “safe zone” available only to the patient and healthcare providers-some who know, personally, how it feels to cope with abuse. Since developing the AIRS tool, Dr.Khurana says people have thanked her and revealed their own experiences with Intimate Partner Violence. “There were many physicians at the Brigham and Mass General Hospital who came to me and shared their personal stories,” Dr.Khurana said.

Tool implemented at Brigham and Women’s Hospital   

With a new National Institutes of Health (NIH) grant, 7,000 new images and a team of eager trainees, Dr. Khurana is now refining the tool. Its current iteration is being implemented in the Emergency Department at Brigham and Women’s Hospital and at some primary care sites. She says developing the AIRS tool has made her work even more meaningful. She is thankful to her team, her mentors, and the patients themselves who tell her that there is empowerment in knowing that they are not alone. “That empowerment feeling is very strong… community help is extremely important,” Dr. Khurana said. “We can provide that.” Six years after asking whether AI could be used to help patients who experience Intimate Partner Violence, Dr. Khurana’s answer could one day change outcomes around the world. “I am very happy and satisfied,” she says smiling, “that finally we are doing something for these patients.”

Dr. Khurana is looking for survivors and volunteers to participate in this project.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *