There are areas in healthcare where tracking patient survival is routine, allowing clinicians in those fields to improve upon their practice and, as a result, patient outcomes. Among clinical care systems that track survival are oncology and cardiology. Michael Schoenbaum, Ph.D., Senior Advisor for Mental Health Services, Epidemiology, and Economics at the National Institute of Mental Health, says missing from this arena is mental health. He says tracking mortality, or turned around, survival, would allow the field to learn and evolve. “We can do better for people by tracking from diagnosis or other index events like discharge from the emergency department or psychiatric hospitalization.” Survival, says Dr. Schoenbaum, is a patient-centered outcome. “It seems obvious, but, that’s why people seek out healthcare: to prolong their lives. They expect not to die.”
Oncology, in part, because cancer is a reportable condition, is the medical field with the most robust tracking system. If a person is diagnosed with cancer, it must be reported to the Centers for Disease Control and Prevention (CDC); the person is tracked from the day of diagnosis until the day she or he dies. Dr. Schoenbaum says this has resulted in rich information on survival patterns and correlatives. If, for example, a person is diagnosed with Stage 4 colon cancer, the system can tell the clinician the fraction of patients with that cancer and stage who survived to varying lengths of time, such as 6 months, 12 months, 5 years, and so on. It can even include correlatives like age, gender, race, ethnicity, and geographic location, allowing researchers and physicians to narrow down a person’s most likely outcome. For instance, if the patient is a woman 35-40 when diagnosed with Stage 3 ovarian cancer, the clinician can quickly find out the number of similarly situated people who survive and to what length of time. The tumor registries also collect information on the care people receive, giving clinicians and researchers information on what happens when people have a particular type of surgery or chemotherapy. “These represent statistical associations. You can look at the 12-month survival of a person who had a particular surgery and compare it to someone who had another surgery or chemotherapy. Of course, we can’t say that the differences are causally due to the different treatments, but researchers can use these data to develop hypotheses.” It creates, at a minimum, a basis for researchers to say, “Well, that’s something we should look into.” The data collection also helps determine a survival time trend: whether the five-year survival for a patient at a certain age, cancer, and stage, is better today than it was a decade ago.
Determining whether survival rates are improving over time requires tracking survival in the first place. In oncology, they can be broken into subsets, including geography, census region, and gender. This helps researchers to look at practice patterns, which often vary geographically, helping to pinpoint areas where outcomes are superior. “This allows researchers to note the places that are doing better and dig into why.”
Heart surgery outcomes are also carefully tracked in the National Cardiovascular Data Registry (NCDR), giving data on 30-day post-discharge mortality. The Society for Thoracic Surgeons (STS) started the clinical database in 1989 and began a voluntary reporting program with the magazine Consumer Reports in 2011, allowing consumers to pick up the magazine and see the survival rate of heart surgery patients by hospital. This empowers the consumer with information on hospitals and surgery types.
Dr. Schoenbaum says the contrast between mental health and oncology and cardiology is stark. Determining mental health outcomes, and whether they are improving over time, is impossible because, with few exceptions, there is no representative data point for a single moment in time regarding survival. “We don’t know the survival of patients discharged from psychiatric inpatient care by hospital, state, or from year to year; we don’t know it for any year. We don’t have a benchmark for it.” There are some exceptions. For example, as a result of the Interdepartmental Serious Mental Illness Coordinating Committee (ISMICC) recommendations, researchers from the Department of Veterans Affairs (VA), earlier this spring, published information on VA patient survival after discharge from psychiatric hospitalizations for a multi-year period. “We don’t have similar information on non-VA patients and hospitals.” Even so, just this release of information, which isn’t as comprehensive as that collected by the NCDR—it doesn’t give information per VA hospital, for example—sheds essential light on the fact that people discharged from psychiatric hospitalization are at high risk for premature mortality.
Another mental health event that puts people at high risk of mortality, suicide or otherwise, is a suicide attempt. The CDC reports there are between 1 and 1.5 million ED visits by people who’ve harmed themselves on purpose. Dr. Schoenbaum says that’s as close as those in the field can get, in the current diagnostic coding, to tracking suicide attempts. “The problem is that deliberate self-harm and suicide attempt are not one and the same: a suicide attempt is when the person who did the deliberate self-harm intended to die.” He says what’s clear is that someone who self-harms, and needs medical care as a result, is at serious risk for premature mortality after discharge. That’s why suicide prevention efforts are aimed to help people who survive intentional self-harm, intending to help them stay alive. “But we don’t know what fraction of people is still alive 3, 6, or 12 months after they go to the ED with intentional self-harm. We don’t have one plausibly representative statistic on that for the United States overall or even for a single state.”
Dr. Schoenbaum points to opioid overdose as yet another area of tremendous vulnerability that is not tracked. He says when people who overdose need medical care in the ED and survive, it’s the overdose equivalent to a suicide attempt. “We have every reason to think that when they leave the ED, they will still have the opiate addiction that led to the overdose in the first place, and so they have a high risk for another overdose and that one, or the one after that, could kill them.” He says the goals of health care, after people have been identified as having survived an overdose, is to keep them alive, “but we don’t even know the rate at which they are dying because we don’t measure it.”
Two years ago, Dr. Schoenbaum co-authored a study looking at mortality in young people with a first or new episode of psychosis, using a national database of healthcare claims of mostly Optum beneficiaries. The researchers looked at all-cause mortality because that’s what they were able to get. They were able to find out whether a person died, not how he or she died. The results of the study revealed that during the first year after a first or new episode of psychosis in young people (ages 16-30), at least 2 percent of them had died. This percentage is striking because it’s about 25 times higher than similarly aged people in the U.S. without psychosis. Mortality for young people who have experienced a first or new episode of psychosis is more comparable with people who die within 30 days of coronary artery bypass grafting (commonly known as bypass surgery), which is 1.5-2 percent. “It turns out that young people in that first year after first or new episode psychosis die in that first year at similar rates as people who’ve had open-heart surgery. We think of heart surgery as a high mortality concern such that we should be tracking it, and yet, it never occurred to us to track mortality in young people with psychosis.” A subsequent study was conducted by the Mental Health Research Network (MHRN), a consortium of research centers affiliated with 13 large health systems across the United States, which included cause of death information. That study also found very high mortality in young people with a first or new episode of psychosis and that most excess deaths—the number of deaths above what’s expected and caused by a specific disease, condition, or exposure—were due to unintentional injury and, especially, to suicide.
As rates of suicide rise in the United States, Dr. Schoenbaum says tracking survival is essential to creating a benchmark even to begin to understand why, and determining what improves outcomes also will require monitoring. Tracking opens doors to new discoveries, revealing additional areas of vulnerability and successes that should be further examined, but Dr. Schoenbaum says without tracking, the field of mental health is plagued with a litany of questions it cannot answer.
We’d like to hear what you think about this or any of our articles. Here’s how to reach us.