Workplace violence in the healthcare industry remains a principal concern of executives and administrators throughout the world. J. P. Phillips, in a review article in the New England Journal of Medicine wrote that, “Health care workplace violence is an underreported, ubiquitous, and persistent problem that has been tolerated and largely ignored.”(1) He mentioned that a majority of these attacks are verbal, but can be much more serious. Attacks on healthcare workers contribute to professional burnout and also risk lower-quality provider–patient interactions.(2-4) Such attacks occur in countries all around the globe.(5,6)
Acts of violence can include both verbal and physical assaults; intentional homicides are, however, quite rare. Robiner and coworkers reported on 944 killings of healthcare workers in 10 categories from 2003 to 2020, based on statistics compiled by the National Violent Death Reporting System.(7) However, these were all-cause killings. The authors noted that “few of these homicides were related to professionals’ work.”(7)
Healthcare executives, to help protect the lives of their workforce, need to understand what motivates patients to murder providers, especially those with whom the killer has a professional relationship. To accomplish this objective, we need to analyze a substantial number of cases where a patient (or family member) murdered his or her healthcare provider, a challenging task when quantifying a low-frequency event. Moreover, certain classes of at-risk providers — for example, nurse practitioners and physician assistants — have existed only since the mid-1960s, making statistical analysis difficult for these and similar categories of workers.
Professional relationships between physicians and patients, on the other hand, have existed since the beginning of recorded history. Likewise, patient-instigated doctor killings generate far more news coverage than the typical homicides that occur daily in our cities and towns. In the past, as now, newspapers devoted considerable space to physician-involved killings, whether as perpetrator or victim.
Optical character recognition (OCR), a century old invention originally created to help the blind read,(8) has been used to scan and digitize billions of pages of newsprint, simplifying searches for information about past newsworthy events. We used such technology to develop a database of cases where patients killed their doctors.
Methods
To create our database, we queried ProQuest, Factiva, and other newspaper archives, extending our search back into the 1800s. With the help of our university’s research librarian, we set up a search protocol that included the terms patient* AND physician* OR doctor* OR surgeon* OR dentist*, or other kinds of healthcare practitioners including psychiatrists and psychotherapists as well as chiropractors and osteopaths. We also added the search terms kill* OR murder*, as well as shoot*, shot, stab*, strangl*, and so forth. This yielded many thousands of newspaper articles. (The asterisk signifies any possible ending to the root word: murder* = murdered, murdering, murders, and so forth.)
We read and eliminated most of the articles, because they describe cases that were not appropriate to the survey, such as when a physician murdered the patient, or a when a doctor was killed during a random street robbery. The search also located cases about legal euthanasia, which we deleted. Likewise, there were many newspaper articles describing a physician forensically examining the body of a person who had been killed or murdered, or testifying at a murder trial; all of these were eliminated as well. Altogether, more than 1000 newspaper articles were thus discarded.
We eliminated all cases where a doctor was murdered because he or she was performing abortions, this being a distinct class of killings that would not shed light on our objective. Such killings are not accomplished by patients of doctors offering abortion services, but, instead, are done by individual opposed to abortions.
We limited our search to doctor killings in the United States and Canada, because societal factors in other parts of the world, not typical of our culture, may have motivated the murderers. As we extended our search to more regional newspaper archives, we began to see names already in our database, indicated that we had identified most of the killings that were reported in newspapers over the past 160 years.
We eliminated almost all cases where the killer was not identified, because these could have been random in-office robberies or drug-seeking behavior where the practitioner’s death was a byproduct of that intent. In cases where the killer was a known patient of the doctor, but the motive was not listed in the newspaper article, we included the case, listing the murder cause as “unknown.”
Using Excel’s “forms” feature, we generated an entry document for each doctor murdered by a patient. The form included the doctor’s and killer’s first and last names; the location of the killing; the relationship of the killer to the patient; the gender of the doctor, patient, and killer; the murder weapon used; the patient’s diagnosis (if recorded); whether or not the killer committed suicide after the killing; and the motivation for the killing. There was space in the form for any additional comments.
Results
After eliminating all cases that had insufficient information or that were not useful for other reasons, 123 newspaper reports remained of practitioners who were killed by an individual (or the relative of an individual) with whom that healthcare provider had a professional relationship. (Because of missing information in newspaper articles, not all statistical information adds up to 123.)
The earliest murder in our series occurred in 1869, and the latest happened in July of 2023. The peak decade for murders was 1901 to 1910 (22), for reasons that are unknown, but is likely a statistical aberration. For three decades thereafter, slightly more than 10 murders per decade were recorded, with the numbers falling off to about five murders per decade over the next eight decades (Figure 1).
Figure 1. Murders of doctors by decades, 1861 to present.
The murder victim usually was described in the newspapers as physician (55), by far the most common identification, with dentist (12) and psychiatrist (7) being the second and third most common identifier of the practitioner’s profession. Next came surgeon (6), orthopaedic surgeon (5), and then two each for urologist, ENT, psychotherapist, chiropractor and pediatrician. Only one murder each was listed for osteopath, plastic surgeon, dermatologist, physical medicine and rehabilitation physician, cardiac surgeon, gastroenterologist, and emergency physician.
Most of the reported murders were committed in the doctor’s office; with about half the cases occurring there. Another 26 murders occurred in the patients’ homes, typically in an era when doctors made house calls. Hospital settings, the doctor’s home, and a public street or building rounded out the rest of the murder locations (Figure 2).
Figure 2. Locations where doctors are most often murdered.
In most cases, the murder weapon was a handgun (Figure 3). Almost all the murdered practitioners were men, with only three women among the 123 that we identified in the articles. The killers usually were men (106), although some were women (17). Almost one-fourth of the killers committed suicide, either immediately following the murder or shortly thereafter.
Figure 3. Murder weapon or mode.
The most common recorded reason (42) for murder of a doctor was that the killer was unhappy with the medical or surgical treatment of either themselves or a close family member. For 21 of the murders, the cause that led to the killing was listed as “unknown.” Twenty doctors were killed for what we called “alienation of affection.” In all such cases, a husband murdered the practitioner for having an affair or some other inappropriate relationship with that killer’s wife (Figure 4).
Figure 4. Reasons why doctors are murdered.
Twenty murders were attributed to the patients’ having psychosis or schizophrenia. Five murders involved a political or business deal or other involvement of the killer with the practitioner that we recorded as “nefarious.” For instance, one killing was related to the troubles in Northern Ireland, and another was traced to Chinese group relationships. Monetary loan issues led to the killing of two physicians.
Only a small number of providers (4) were killed during the course of a robbery by one of their patients. Three patients killed the doctor for not extending disability, and we could find only one case where a patient murdered the doctor for not providing more medication.
Discussion
Violence toward physicians is a subset of violence toward healthcare workers in general, which, in turn, is a manifestation of violence in all workplaces. Governments around the world have created taskforces to deal with the problem of increasing workplace violence. They generate pamphlets containing suggestions designed to reduce attacks and enhance worker safety.(9) Yet in spite of these efforts, workplace attacks continue to increase, with ever greater numbers of casualties, especially at mass shooting events.(1,10-14) These incidents are taking on epidemic features, yet we believe that efforts directed at reducing workplace violence are not likely to lead to fewer murders of professional practitioners by their patients. Here’s why.
When a physician is murdered by a patient, the incident usually is a targeted killing. The killer probably will circumvent measures generally directed at lowering workplace violence. For example, hospitals now recommend keeping scissors and other potential weapons out of sight.(9) Although this may reduce the number of spur-of-the-moment attacks by a patient on a healthcare worker, it will not deflect a planned killing. The same can be said for using locked doors with electronic latches between the waiting room and clinical areas in a practitioner’s office. Without the use of stadium-type metal detectors, such measures will do nothing to prevent a patient from killing a doctor with a concealed weapon.
Even keeping patients known to be disruptive out of the office entirely would not stop a premeditated murder. Such a perpetrator could simply wait in the parking area and kill the doctor there. It would take presidential-type protection to shield a healthcare worker from being killed by someone who has a mind to do so. Therefore, certain preemptive measures are necessary to protect practitioners.
First and foremost, providers must avoid sexual dalliances with patients, even if they are consensual and occur outside of business hours and locations. Aside from generating civil lawsuits when lust fades — and problems with regulatory agencies — such behavior is an invitation to attacks by a spouse if the patient is married or otherwise involved with someone else.
Attacks on healthcare workers by mentally ill persons is a unique problem faced especially by aides and nurses employed where individuals with significant psychopathology are confined in a locked-down facility.(15-17) For instance, disputes about smoking in non-designated locations are among the most common stimuli for nonlethal attacks.(16) Typically, the patient will aggressively twist the arm of the healthcare worker, or engage in some similar activity. Psychiatrists sometimes, but not always, can defuse potentially murderous behavior, as our survey revealed.(15)
We recommend that healthcare executives inform their provider staff that the best way to reduce the likelihood that a dissatisfied patient will kill a practitioner is to establish a positive provider–patient relationship throughout the entire sequence of encounters. Because patient dissatisfaction with medical or surgical care is, according to our findings, the most common cause of murders of doctors by patients, healthcare providers should be alert to the concerns of unhappy patients.
Not every medical or surgical intervention results in clinical success. One potentially effective way to reduce the likelihood of unhappy patients is pretreatment selection, taking into account those variables associated with suboptimal outcomes. One general principle is that one should avoid pain-relieving operative interventions when the pretreatment pain level is not very great, or when significant pain is not accompanied by corresponding organic findings. A useful model of pain recognizes a biopsychosocial origin of such symptoms,(18) proposing that an underlying issue can be amplified (or repressed) by the patient’s personality and intrapsychic issues, as well as external social factors, such as secondary gain, illness role behavior, and so forth. Such things as pain catastrophizing(19) or an unconnected but simultaneous “injustice experience” (such as being fired at work)(20) can amplify pain beyond any factor associated with injured bodily structures.
Many well-known risk factors can contribute to actual or perceived poor surgical or medical treatment outcomes, including smoking, diabetes, obesity, malnutrition, socioeconomic status, Workers’ Compensation, and patient’s education level.(21,22)Although a practitioner cannot cherry-pick for treatment only those individuals without any risk factors, an honest discussion about the potential for a suboptimal outcome should prepare the patient for less-than-hoped-for results. This may reduce the likelihood of a future violent attack.
Currently there is a tendency for providers who are confronted by an unhappy patient to send that individual to a pain management or other specialist for further care. A wise provider, rather than dismissing such patients from their care, will maintain contact with those individuals, with periodic telephone calls and office visits to check on progress. Five minutes on the phone may be all that’s needed to prevent a practitioner death.
In spite of all precautions, even thoughtful, caring patient interaction accompanied by a high-quality outcome, there is still the possibility that a provider may be killed by a patient. Tennessee’s senate minority leader, speaking after the killing of a doctor in his state, said, “Tragedies like this underscore the urgent need for common sense — like reinstating background checks and gun licenses and establishing new reforms like an order of protection so police can remove firearms from a person who is threatening others. Whether you’re at school, a grocery store or on the job, you deserve to live and work free from gun violence.”(23)
In light of our finding that 89% of patient-instigated physician murders involved firearms, stronger gun laws seem more likely to reduce the incidence of such occurrences than any other feasible measure. Considering the prevalence of firearms in America, however, we can’t depend on such reforms to save many providers’ lives. For this reason, establishing caring provider–patient interactions is our best hope.
References
Phillips JP. Workplace violence against health care workers in the United States. N Engl J Med. 2016;374:1661-1669. https://doi.org/10.1056/NEJMra1501998
Arnetz JE, Arnetz BB. Violence towards health care staff and possible effects on the quality of patient care. Soc Sci Med. 2001;52:417-427. https://doi.org/10.1016/S0277-9536(00)00146-5
Hacer TY, Ali A. Burnout in physicians who are exposed to workplace violence. J Forensic Leg Med. 2020;69:2-5. https://doi.org/10.1016/j.jflm.2019.101874
Weisberg J, Sagie A. Teachers’ physical, mental, and emotional burnout: impact on intention to quit. J Psychol. 1999;133:333-339. https://doi.org/10.1080/00223989909599746
Lorettu L, Nivoli AMA, Daga I,( )et al. Six things to know about the homicides of doctors: a review of 30 years from Italy. BMC Public Health. 2021;21(1):1318. https://doi.org/10.1186/s12889-021-11404-5
Pich J, Roche M. Violence on the job: the experiences of nurses and midwives with violence from patients and their friends and relatives. Healthcare. 2020;8:522. https://doi.org/10.3390/healthcare8040522
Robiner WN, Freese RL, Barnes RD, Palmer B, Kim MH. Homicides of psychologists, physicians, nurses, pharmacists, social workers, and other health professionals: National Violent Death Reporting System data 2003-2020. J Clin Psychol. 2023;79:2932-2946. https://doi.org/10.1002/jclp.23589
Morgan RS. Optical readers. Computers and the Humanities. 1970;5(2): 75-78. https://doi.org/10.1007/BF02402284
Cal/OSHA Guidelines for Workplace Security, 1995: Sacramento, CA.
Kowalenko T, Waters BL, Khare RK, Compton S. Workplace violence: a survery of emergency physicians in the state of Michigan. Ann Emerg Med. 2005;46:142-147. https://doi.org/10.1016/j.annemergmed.2004.10.010
Kumar M, Verma M, Das T, Pardeshi G, Kishore J, Padmanandan A. A study of workplace violence experienced by doctors and associated risk factors in a tertiary care hospital of South Delhi, India. J Clin Diagn Res. 2016;10(11):6-10. https://doi.org/10.7860/JCDR/2016/22306.8895
Lanza ML. Factors relevant to patient assault. Issues Ment Health Nurs. 1988;9:239-257. https://doi.org/10.3109/01612848809140927
Lipscomb JA, Love CC. Violence towards health care workers: an emerging occupational hazard. Am Assoc Occup Health Nurs. 1992;40:219-228. https://doi.org/10.1177/216507999204000503
Paola F, Malik T, Qureshi A. Violence against physicians. J Gen Intern Med. 1994;9:503-506. https://doi.org/10.1007/BF02599220
Madden DJ, Lion JR, Penna MW. Assaults on psychiatrists by patients. Am J Psychiatry. 1976;133:422-425. https://doi.org/10.1176/ajp.133.4.422
Nelson B. Acts of violence against therapists pose lurking threat. The New York Times. June 14, 1983:C1.
Squires S. Attacks on psychotherapists: assaults by violent patients are an increasing concern. The Washington Post. July 2, 1990.
Wade DT, Halligan PW. The biopsychosocial model of illness: a model whose time has come. Clin Rehabil. 2017;31:995-1004. https://doi.org/10.1177/0269215517709890
Hampton SN, Nakonezny PA, Richard HM, Wells JE. Pain catastrophizing, anxiety, and depression in hip pathology. Bone Joint J. 2019;101-B:800-807. https://doi.org/10.1302/0301-620X.101B7.BJJ-2018-1309.R1
Sturgeon J, Seward J, Rumble D, Trost Z. Development and validation of a daily injustice experience questionnaire. Eur J Pain. 2021;25:668-679. https://doi.org/10.1002/ejp.1702
Dy CJ, Bozic KJ, Pan TJ, Wright TM, Padgett DE, Lyman S. Risk factors for early revision after total hip arthroplasty. Arthritis Care Res (Hoboken). 2014;66:907-15. https://doi.org/10.1002/acr.22240
Wylde V, Trela-Larsen L, Whitehouse MR, Blom AW. Preoperative psychosocial risk factors for poor outcomes at 1 and 5 years after total knee replacement. Acta Orthop. 2017;88:530-536. https://doi.org/10.1080/17453674.2017.1334180
Bella T. Patient shoots and kills surgeon in exam room, police say. The Washington Post. July 12, 2023. https://www.washingtonpost.com/nation/2023/07/12/tennessee-doctor-shot-patient-benjamin-mauck/