In May researchers published the results of a sweeping study on the deaths of U.S. physicians. Appearing in the journal Academic Medicine, the study gathered data on more than 380,000 resident physicians—doctors who graduated from medical school and were training in their chosen specialties—between 2000 and 2014. The authors identified 324 resident deaths during that time period and compared them with data from the U.S. Centers for Disease Control and Prevention’s National Death Index.

Cancer was the most common cause of death among residents, although at lower rates than people of similar age and gender in the general population. Meanwhile, suicide was the second-leading cause of resident death—and the most common cause of death among male residents.

We’ve known for some time that doctors are dying by suicide. Over the last few years, suicides among U.S. doctors have stirred national discussion. In leading academic journals mental health experts have sought to address suicides among physicians and medical trainees. Media outlets have added coverage with headlines like “Why Do Doctors Commit Suicide?” and “The Hidden Epidemic of Doctor Suicides.” Pamela Wible, a family physician, gave a talk at a 2015 TEDMED conference called “Why Doctors Kill Themselves.”

As a resident physician in psychiatry, I’ve written about suicide among physicians and medical students in the past. So this new study in Academic Medicine felt sadly familiar. But the study also included some startling findings.

Although suicide was one of the leading causes of death among residents, the rates of suicide for residents appeared to be much lower compared with the general population. For people aged 25 to 34, when doctors are commonly in residency training, the odds of suicide among residents were about one fourth that of their age-matched peers. The authors found this “surprising, given the data on higher rates of depression in medical students…and the higher rates of suicide deaths in practicing physicians” compared with the broader public.

Does this mean suicides aren’t an issue for residents? Far from it. Each suicide is a devastating loss not only to family and friends, but also to colleagues and patients. The study identified at least 66 resident suicides from those 15 years. That’s 66 people lost to loved ones. Those are 66 doctors who no longer live and breathe.

Still, this study suggests we have much to learn about the causes of suicide among U.S. physicians. For example, it’s long been believed medical trainees are at high risk for suicide. A 2015 review of studies estimated 22 to 32 percent of resident physicians in the U.S. suffer from depressive symptoms, and multiple studies have shown that residency training places physicians at risk for mental illness and suicidal thoughts. Proposals to combat physician suicide have often focused on mitigating risk factors for suicide like depression and sleeplessness.

But if residents have lower suicide rates than the general population, residents may also have characteristics that protect against suicide. Protective factors might include problem-solving skills, social supports, links to mental health resources, reasons for living and financial stability. Further research into these types of protective factors could help prevent future suicides among doctors. The results of this latest study also force us to confront a reality often overlooked in conversations about physician suicide: medical students, residents and attending physicians are not the same.

Coverage of physician suicide frequently lumps together stories about these groups, but medical professionals face unique stressors at different stages of training. Medical students may struggle with school-related matters like anatomy exams or residency applications whereas residents may wrestle with learning how to independently take care of patients. Attending physicians—doctors who have completed medical school and residency—encounter additional pressures like billing, insurance networks and malpractice suits.

Better data can help us understand how suicide risk evolves for doctors—from before they enter medical school all the way beyond retirement. For instance, the study on resident deaths found 74 percent of suicides occurred during the first two years of residency training. The authors also found that most resident suicides took place during the months of July to September and January to March.

Are there other trends that we should be paying attention to? Can we find more nuanced approaches to suicide prevention for medical students and doctors? The trouble is that these kinds of useful data are hard to come by. Many of the numbers commonly cited regarding physician suicide are either outdated or unreliable. Virtually every article on the issue refers to a statistic from the American Foundation for Suicide Prevention, stating 300 to 400 physicians kill themselves every year. But according to Jill Harkavy-Friedman, vice president of research for the foundation, these numbers aren’t exact and represent estimates from an expert panel held 15 years ago.

Others have written that as many as 150 U.S. medical students die by suicide each year. But recent research suggests otherwise. In 2014 a group from the Baylor College of Medicine published a survey of U.S. medical schools tracking any suicides between 2006 and 2011 among medical students. After receiving responses from about 70 percent of allopathic medical schools nationwide, the authors could identify just six suicides from those five years.

More data on these issues may be forthcoming, though. In Missouri lawmakers are considering legislation that would promote research on mental health among the state’s medical students. The American Medical Association recently announced its support for further study into the mental health of medical students and physicians. Researchers like Katherine Gold at the University of Michigan are collecting large data sets from national registries to better analyze suicides among physicians.

Some of this research may take years to publish. But to prevent suicides among physicians and medical students, we have to do a better job at defining the problem. We can’t develop solutions to a problem that we hardly understand.