The patient was calm, affable, and appreciative — and I felt an uncomfortable prickle of shame. I was on an overnight shift admitting patients from the ED. This patient was my newest admission, and I had opened the chart to read more about him. Immediately a large, bright yellow pop-up window filled the screen: the EHR’s flag for patient behavioral concerns. Instinctively, my mood soured. The flag warned me about the patient: “aggressive,” “verbally assaulted,” “formal warning,” etc. The evening was already a mess with codes, rapids, and multiple complex admissions — I was not looking forward to being yelled at, or worse. My colleagues commiserated with me — we had all been there before. When I entered the patient’s room, I braced myself for the worst. So, when he was reasonable, measured, and understanding, I felt somewhat guilty. That one alert had primed me for an antagonistic relationship with the patient before we had even met. I was left wondering, was I (or the patient) better off with the presence of that flag?

Patient behavioral flags are alerts in the EHR that provide details to alert health care staff to possible safety concerns. These include physical or verbal violence toward clinicians or those who have brought in weapons. Behavioral flags are one attempt to fix an alarming statistic: Health care workers face more workplace violence than any other profession. Moreover, these rates are rising, increasing 63% from 2011 to 2018. In the ED alone, over half of physicians and 70% of nurses say they have been physically assaulted by a patient.

For those of us in the hospital, these events are grimly expected. I had one physician as a patient who was so severely assaulted by her own patient that she is unable to work  full-time due to chronic orthopaedic issues. Co-residents have been virtually and physically stalked by patients. Another had feces thrown at him. Many have been slapped, punched, and kicked. We have all been yelled at, spit on, and threatened.

My mom (who is a nurse) tells the story of when she was eight months pregnant with twins, facing a violent patient. She recalls the event almost in slow motion: He pulls his fist back aimed to punch her in the stomach, and she was too swollen to move away quickly. Fortunately (for her), a selfless colleague jumped between them to take the punch so that my mom was spared.

The violence is real, pervasive, and unacceptable.

Behavioral flags in a patient’s chart serve as a warning to clinicians that can be helpful, timely, and important. But — as my own experiences have demonstrated — these flags pose a real risk of bias. A study based at the University of Pennsylvania found that while behavioral flags were used relatively infrequently (3.5 flags per 1,000 patients), there were differences in who was flagged and subsequently what their care looked like. Strikingly, Black patients were flagged almost twice as much as white patients (4.0 flags versus 2.4 flags per 1,000 patients). Patients with behavioral flags were more likely to leave without being seen by a clinician and were less likely to be admitted. Disparities persisted within patients who have flags. Black patients with a flag had longer waiting times and underwent fewer lab and imaging tests compared to White patients with a flag.

Of course, there are many confounding factors, and the number of labs and imaging tests ordered does not directly correspond to quality care. But I know I walked into my patient’s room with a preconceived notion about him — it’s not unreasonable to postulate that such biases multiply across many physicians and have adverse effects on outcomes.

Behavioral flags should be just one of many interventions to reduce violence against health care workers. And there should be guardrails in place to help reduce the bias introduced by these flags. These can include:

• Being judicious about when to put a flag in a patient’s chart. Consider behavioral flags for patients whose past behavior poses a credible threat to physician safety. This can include, for example, patients who have a prior record of physically harming clinicians or those who have brought weapons on the premises with an intent to harm. It can also include those who have expressed verbal threats that are specific and credible. For example, there is a categorical difference between a patient who is yelling generally that they are in pain and need medication versus a patient who told one of the nurses I work with “F*** that f****** fat*** nurse b**** c*** who didn’t get my pain meds, she better watch her back, my uncle’s coming with my knife and I’m going to kill her.” Although being yelled at is never pleasant, in health care we accept that we see patients at their very worst and it's important to have compassion. However, sexist and racist verbal abuse and threats of physical violence should never be acceptable. 
• Implementing a mechanism and process for regular review and removal of a flag from a chart. How this is implemented would be institution-dependent. For example, criteria for removal could include resolution of the circumstances that led to the flag being placed or a set time period without recurrence of the behavior that led to the flag. Reviews could occur every one to three years by a multidisciplinary committee. One consideration could be having a patient advocate as part of the committee. 
• Informing patients that a flag has been placed in their chart. One study interviewed patients to better understand patients’ perceptions of behavioral flags. They found that the majority of participants agreed that patients have “a right to know” whether a behavioral flag has been placed in their chart. Many hospitals send a follow-up letter to patients with a formal warning if their behavior warranted security involvement. An asynchronous letter could be sent to the patient in the setting of a behavioral flag being placed, especially if there is concern that informing that person during the visit would escalate the behaviors or increase the threat to clinicians. It is important that patients are informed of the consequences of their actions, and provided the opportunity to address them. 

Ultimately, the factors associated with violence against health care workers are complex and multifaceted — and will not be fixed with patient flags alone. In the end, the goal is to ensure the safety of health care workers while maintaining the integrity of patient care. In the meantime, I strive to approach each patient flag with vigilance and compassion.  

What has been your experience with patients with behavioral flags? Share in the comments.

Corinne Carland is an internal medical resident at the University of Pennsylvania in Philadelphia. She is interested in health tech, clinical informatics, cardiology, and genomics/proteomics. Outside of work she enjoys walking/running along the river, trying new ice cream spots, and exploring museums. Dr. Carland is a 2023–2024 Doximity Op-Med Fellow.

Illustration by Jennifer Bogartz