Why empathetic clinicians face fewer lawsuits

AA surgeon with a flawless operative record gets sued twice in five years. A surgeon down the hall, working comparable cases with comparable outcomes, has never been sued. The difference is rarely visible in the operating room. It tends to show up in the pre-op visit, the post-op rounds, the way a complication is disclosed. The relationship between empathy and malpractice risk is one of the most consistent findings in clinical communication research, and the Levinson 1997 study is the paper that put numbers on it. For anyone searching empathy malpractice or communication malpractice claims, this is the body of evidence to start with.

Wendy Levinson and colleagues audio-recorded routine office visits of primary care physicians and surgeons, half of whom had two or more prior malpractice claims and half of whom had none. Coders blind to the claims history scored the recordings on communication behaviors. Among the primary care physicians, those with no claims used more orienting statements, more humor, more facilitation of patient input, and longer visits, by roughly three minutes on average. The no-claims group was about three times less likely to have malpractice claims than the high-claims group. The effect did not replicate as cleanly in surgeons, but the primary care finding has been cited thousands of times because it was the first hard data linking communication behaviors to litigation risk.

“Patients almost never sue clinicians they trust. They sue clinicians who made them feel dismissed, abandoned, or deceived.”

The intuitive reading of the empathy malpractice link is that nice doctors get sued less. The mechanism is more specific than that. Studies of plaintiff motivation, including the foundational work by Beckman, Markakis, Suchman, and Frankel, found that bad outcomes alone almost never trigger lawsuits. What triggers a suit is a bad outcome combined with a breakdown in the clinician-patient relationship: the patient felt the clinician was rushed, dismissive, deserted them after the complication, or was less than honest. Empathetic communication is protective not because it prevents complications but because it preserves the relationship through complications. A patient who trusts the clinician interprets a bad outcome as bad luck. A patient who does not trust the clinician interprets the same outcome as negligence.

The three-minute visit-length difference in the Levinson data is often misread as a finding that long visits prevent lawsuits. That is not quite right. The extra minutes were used for specific behaviors: setting the agenda at the start, asking what else the patient was worried about, signaling transitions, and inviting questions at the end. These are behaviors that the Marvel 1999 study would later document as protective against premature interruption. The structural lesson is that the protective effect comes from how the time is spent, not just from spending more of it. A clinician who runs a tight, well-structured visit can capture most of the relational benefit without adding much wall-clock time.

For medical students and residents reading the Levinson 1997 study for the first time, the practical implications are immediate. The behaviors associated with lower claims risk are learnable, brief, and unrelated to personality. They include opening the visit with an agenda-setting question, summarizing what the patient said before moving on, signaling transitions out loud, and explicitly inviting questions before closing. None of these require a different temperament. They require a different habit structure. The Riess 2012 trial later showed that a brief, neuroscience-informed empathy curriculum can shift these behaviors measurably in residents who were skeptical of the training going in.

The Levinson work pairs naturally with the disclosure-and-apology literature. Programs at the University of Michigan and elsewhere have shown that proactive, empathetic disclosure of adverse events, including an apology where appropriate, is associated with fewer claims and lower payouts, not more. The mechanism is the same one. Patients who feel the clinician took ownership of a bad outcome experience it as a relationship that held. Patients who feel the clinician hid, deflected, or minimized experience it as a relationship that broke, and the legal system is one of the few remaining channels through which they can demand acknowledgement.

The Empathy in Medicine Initiative covers the malpractice evidence base explicitly in its training because trainees are often told empathy matters without being shown the numbers. Knowing that the Levinson 1997 study found a roughly threefold difference in claims rates linked to communication behaviors changes how trainees defend the practice to skeptical attendings. The EIMI ambassador program trains students to teach this material to their peers and to model the agenda-setting and disclosure behaviors that the evidence supports. The free EIMI training and application materials live at /Training and /Apply.

The blunt summary is that *patients almost never sue clinicians they trust*. They sue clinicians who, after a complication, made them feel dismissed, abandoned, or deceived. The communication behaviors that build trust before a complication are the same behaviors that protect the relationship after one. They are inexpensive, learnable, and on the available evidence, among the most underappreciated risk-management tools in clinical medicine.

A common objection to applying the Levinson data is that the cleanest findings were in primary care, not in procedural specialties. The replication in surgeons was messier, and surgical claims have a different distribution of root causes. Even so, the qualitative literature on surgical malpractice suit motivation maps closely onto the primary care findings. Plaintiffs in surgical cases describe feeling abandoned after the complication, unable to reach the surgeon, given conflicting explanations by the team, and treated as a problem rather than a patient. The protective behaviors are slightly different in surgery, the post-complication phone call, the bedside visit that did not have to happen, the explicit acknowledgement of difficulty, but the underlying mechanism is identical. A relationship that holds through a bad outcome rarely produces a suit.

For program directors and risk-management offices reading this evidence base, the operational implication is that empathy training is not soft programming bundled in alongside wellness. It is a risk-management intervention with a quantified effect size, comparable in magnitude to many of the structural risk-mitigation programs hospitals already invest in. Treating it as such, funding it, measuring it, and tracking communication behaviors as a quality metric, would represent a meaningful realignment of how most institutions currently allocate attention. The Beckman, Markakis, Suchman, and Frankel deposition study from 1994 remains the qualitative counterpart to the Levinson numbers and is worth reading alongside it. Plaintiffs in that study repeatedly described being deserted at a time of need, devalued, delivered information poorly, or failing to understand the perspective of the patient and family. None of those four themes are technical. All four are addressable through training.

A final note on professional defensiveness. Clinicians sometimes hear the empathy and malpractice literature as an accusation that bad communication causes lawsuits. The more accurate reading is that good communication is protective against the suits that bad outcomes would otherwise trigger. The outcome is the precipitating event; the relationship determines whether the patient interprets that event as a misfortune or a wrong. Reframing the work this way removes the implicit blame and centers the protective behavior as a learnable craft rather than a personality verdict.