(Neiko Ng for The Washington Post)

Because I am a doctor, my friend Sophia told me the following story. “I go to a walk-in clinic with neck pain and a low-grade fever. I never go to the doctor. I know something is seriously off. ‘Would you test me for strep?’ I ask. ‘You’re overreacting. You just have a cold,’ this young doctor says. Would he have liked to hear me make a bigger deal about how badly I felt? I almost had to beg for a strep test. When it comes back positive, I’m so angry, I can barely speak to him. He was incompetent. Or trying to save money. Or maybe he was just lazy. He was certainly unkind.”

Her doctor, apparently, was a lousy diagnostician. But more than that, she was put off by his attitude. What stuck in my mind weeks later was her choice of the word “unkind.”

It’s reasonable to expect a doctor to be kind at every visit. Kindness may be less important to us when the visit is urgent, when we are in terrible pain and barely listening as we wait for relief, when the problem is diagnosed and fixed quickly. But generally, most of us assume that it matters. In ancient Greece, medical texts advised physicians to “be solicitous in your approach to the patient, not with head thrown back [in arrogance] or hesitantly with lowered glance, but with head inclined slightly as the art demands.” Today, medical schools teach and evaluate kindness at patients’ bedsides and through role-playing. As Leslie Jamison, who acted as a patient, writes in “The Empathy Exams,” “Checklist item 31 is generally acknowledged as the most important category: ‘Voiced empathy for my situation/problem.’ ”

Yet doctors and patients alike have lamented that fully booked appointment schedules, the laptop’s intrusion during history-taking, billing pressures and edicts from insurance companies are squeezing kindness out of the exam room. So what exactly do we lose when we lose kindness? It may improve doctor-patient relations and patient satisfaction, but does kindness matter for patient outcomes? Can it lower the risk of hospitalization or death? Can kindness save lives?

It seems obvious: When doctors are kind, patients do better. But when the hard-nosed and unsentimental scientist demands proof, persuasive evidence is hard to find.

One approach to studying this subject has been to try to correlate doctors’ scores on empathy scales with patient outcomes. Empathy (historically first addressed in terms of an observer’s feelings before a work of art) has been defined in the medical context as “an uncritical understanding of the patient’s experiences, emotions and feelings” that is communicated to the patient. Empathy is as close as medical researchers have come to suggesting the notion of kindness, which is less well defined (although certainly understood by patients). Empathy scales ask a doctor’s level of agreement with a series of statements such as: “My patients feel better when I understand their feelings.” An empathy scale can’t capture the texture of an actual patient interaction, with its shifting tones and positive talk, its nonverbal cues of eye contact and gesture and body language, but self-reports by doctors can give a broad sense of their attitudes toward patients.

Using such scales, researchers have categorized doctors’ empathy as high, moderate or low, and studied patient outcomes that may be affected by empathy. In the few published studies, patients of physicians with high empathy were more likely to have greater control of their diabetes than were patients of physicians with low empathy, the suggestion being that physician empathy is associated with improved patient outcome.

But these findings don’t necessarily demonstrate cause and effect. It may be that doctors, through their empathic treatment, cause their patients to better manage their diabetes, or it may be that patients who do better managing their diabetes cause their doctors to think of themselves as more understanding, more involved, more empathetic.

I recently saw a man who came to see me again and again with dangerously high blood sugars. He seemed unable to bring his diabetes under control, on some days overeating, on others skipping an insulin dose. I was frustrated with him, impatient and terse, blaming him for intransigence and not interested in investigating its many possible causes. Each time he returned unable to meet my eye, knowing he had disappointed me; I dreaded seeing his name on the day’s schedule; his persistently lousy outcomes affected my mood and made me dour with other patients. When his self-care mysteriously improved (his sister had come to live with him), he was happier, and I, too, was pleased. I would have rated my empathy (which I might have imagined was a stable quotient) quite differently depending on whether it was the week before or after his sister moved in.

Skeptics would say this reverse causality is the same misinterpretation economic researchers make when they are out to demonstrate the value and worth of kindness in the business world. Seeking the “kindest leaders,” researchers may convince themselves that when employees are treated better, the company is more successful. The opposite conclusion — successful companies permit everyone to be a little softer, a little more generous — is equally likely but does not support the “kind leader” hypothesis. Everyone wants to find that kindness helps, but it’s difficult to prove.

An alternative approach would be to ask patients to rate their clinician’s empathy. In the single published example of this kind of study, patients with symptoms of a common cold were seen by a provider they’d never met before. Immediately after the visit, they answered 10 questions that assessed the aspects of the encounter related to what the researchers termed empathy. Did the doctor make them feel at ease? Allow them to “tell their story”? Take an interest in them as a whole person? Fully understand their concerns? Show care and compassion? For the 14 days following, patients recorded whether they still believed they had cold symptoms.

Patients who gave their doctors perfect empathy scores had shorter colds — by nearly a full day — and a greater activation of one key protein of the immune system than patients who rated their doctors as less than perfect. Each of the six clinicians in the study saw a comparable proportion of patients reporting perfect empathy scores, precluding the possibility that these results were driven by one clinician being naturally more empathic than all the others. But whereas one might have predicted a relationship between empathy and recovery across the full range of scores, among the three-quarters of patients who gave their clinicians less-than-perfect scores, a better rating did not predict quicker cold improvement. It’s unclear what the lack of a linear response to empathy tells us, but perhaps the perception of empathy is an on-off phenomenon, such that a patient either feels completely connected to her doctor (a “perfect score”) or she doesn’t.

The contrarian could raise further objections. The researchers didn’t take into account an assortment of other variables that contribute to medical outcomes, such as physician competence (were symptom-relieving medications prescribed?), patient adherence to recommendations, social support or concurrent medical conditions. They assigned patients to new doctors, when prior relationships with providers who knew them well may have influenced the positive effects of empathy. Maybe the empathy scale was faulty; it included questions about whether the doctor “explained things clearly” and “helped create a plan of action,” which may not be components of empathy.

It would be best to do a randomized trial, our gold standard for scientific evidence, to demonstrate that the kinder the doctor, the more successful the patient. But few patients would agree to participate in a study where they might be assigned to unkind doctors.

A slightly less ethically fraught study might select a group of doctors, measure the outcomes of some of their patients (researchers could choose a relatively homogeneous group, perhaps those with medication-treated diabetes) over a period of time, then train the doctors to be more empathic and measure the outcomes of a second cohort of their diabetic patients going forward (while also measuring the maintenance of the doctors’ empathy training effect) to see if these patients do better than the group from before the training. We would predict that those doctors who raised and maintained their empathic behavior scores would have better results than doctors who never improved or let their empathy skills lag.

This would be a tough, long and expensive study to perform. And even if the results indicated that empathy improved the outcomes of people with diabetes or asthma, we’d still be at a loss to explain the mechanisms whereby kindness works. Does better, more empathic listening lead to better physician communication, which in turn leads to better patient understanding, which translates to greater adherence to medications or other medical recommendations? Does empathy inspire optimism or positivity in the patient? Does having someone on your side serve as an antidepressant, and does defusing anxiety simmer down the sympathetic nervous system and lower cortisol? Is trust an immune activator?

Given the limits of the evidence, the kindness skeptic could argue that a common cold is a self-limited condition and that this single study hasn’t demonstrated the effect of empathy (or kindness) on more serious, chronic concerns. We wouldn’t know if a doctor’s empathy leads to greater likelihood of cancer remission or fewer flares of lupus or faster healing of wounds.

There are older studies of people with mental-health problems suggesting that the variability in patient depression outcomes — whether the treatment is pharmacology or psychotherapy — depends on the psychiatrist. Psychiatrist Paul Crits-Christoph and statistician Robert Gallop propose that “poor rapport” (another way of saying unkindness?) leads to ineffective care. Put another way, there is the implication that the clinician is not only a provider of treatment but also a means of treatment. These studies, however, don’t explicitly measure patient perceptions and therefore move us only a short way toward demonstrating that kindness matters.

At the moment, the best answer to the kindness contrarian is: Even if the evidence in favor of the therapeutic benefits of empathy is weak, there is no evidence that refutes the idea that empathy improves care. And too many patients have stories of how unkindness or the sheer obliviousness of doctors can be devastating and indelible.

Kindness carries with it a commitment to a certain way of thinking and being rather than to a particular pre-defined endpoint. By showing that they are open to patients’ experiences, doctors are helping them feel better, or at least feel at ease during office visits. Many long-standing medical recommendations (an annual physical examination, a total-body skin cancer check) are being reevaluated, and the makers of guidelines often determine that “there is not enough evidence to recommend.” Such old-fashioned medical interventions, absolutists suggest, could lead to over-diagnosis or over-treatment. But kindness at every visit is never too much to ask. Sophia was right: There is no burden added to the work of doctors if we expect them to be kind. Sometimes doctors don’t need to wait for evidence to do what is right.


Read more from Outlook:

When medical care is delivered in 15-minute doses, there’s not much time for caring

Nurses make fun of their dying patients. That’s okay.

Our unrealistic view of death, through a doctor’s eyes

How voters personal suffering overtook reason — and brought us Donald Trump

and follow our updates on Facebook and Twitter.