For people with mental illness, diagnosis is only the first step towards finding the right treatment. Sometimes that initial diagnosis is wrong—patients with bipolar disorder, for example, might have to wait seven to 12 years for a correct diagnosis.
Even when the diagnosis is right, most patients have to try several different medications to find one that actually works. On one Reddit thread, the number of drugs patients tried before arriving at their current treatment ranged from “one” to “too many to count.”
The main reason? We don’t treat mental illness like other types of diseases. That’s mostly because scientists still don’t understand the machinery that causes them in the first place—schizophrenia, for example, is caused by a combination of genetics and innumerable environmental conditions. Scientists don’t even understand how some of the most common drugs, like lithium, work when they treat mental illness, but they prescribe the drugs anyway because they can help relieve symptoms.
Today, clinicians have to use observable symptoms and patients’ self-reports to diagnose a disorder. Then they start testing treatments from there, a technique one scientist called the “shotgun approach” in a recent article in Scientific American. Done this way, it takes longer to get it right, and patients with severe mental illnesses that aren’t receiving proper treatment can become a risk to themselves and to others. Somewhere between two and 15 percent of patients diagnosed with depression commit suicide, with similar or even slightly higher numbers for bipolar and schizophrenia. Others die from accidents during episodes of mania or psychosis.
But what if a quick blood test could help a clinician diagnose a disorder, or find the right treatment more quickly? A number of scientists are working to find these biomarkers for psychological conditions, and some are getting close.
The search for biomarkers in mental illness has been around for decades, says Sabine Bahn, a professor of neurotechnology at the University of Cambridge. Some scientists are looking for genetic hallmarks of diseases, while others are looking for protein biomarkers in the blood, urine, breath, or spinal fluid. There was some exciting research into biomarkers in the 1970s, and again in the 90s, she says, but it was derailed by small details that threw off the sensitivity and accuracy of the tests. “It’s not for lack of trying,” Bahn says.
The search for a single protein that can indicate a psychiatric condition, like the one detected in at-home pregnancy tests, has so far been a fruitless one. That’s because so many of the proteins indicative of mental illness overlap with different conditions. “It’s hard to have markers specific to depression or schizophrenia—80 percent of the proteins that would be on the panel [to identify] schizophrenia would serve for depression, too,” says Daniel Martins-de-Souza, a biochemist at the University of Campinas (UNICAMP) in Brazil. Now scientists are looking for “biosignatures” of illnesses, a combination of proteins (or genes) that together indicate the presence or likelihood of a disease.
To date, no one has come across a biosignature test reliable enough to be sold commercially.
Bahn and her team thought they had found one back in 2010, when they developed a blood test called VeriPsych designed to confirm a schizophrenia diagnosis. Clinicians started using it all over the world, including in the U.S., but in later that year it went off the market. “The test was very expensive, and there were too many markers,” Bahn says. Rules-Based Medicine, the company selling VeriPsych, was then bought by Myriad RBM, which put the test on ice.
That hasn’t dissuaded Bahn. She and her collaborators published a study last year about a blood test to detect schizophrenia before symptoms set in. But Bahn has mostly turned her attention to identifying bipolar disorder because this is where she sees the biggest need. Clinicians often misdiagnose patients with bipolar disorder, thinking they have depression, Bahn says. “It’s only once patients have manic symptoms above a certain threshold that they can be properly diagnosed.” And if a person with bipolar disorder is treated with antidepressants, that can trigger a manic episode, which often results in hospitalization or death.
Bahn’s bipolar blood test would only require a little bit of blood collected onto filter paper, a process so simple that a patient could do it himself. A processing facility would use a mass spectrometer to detect the proteins in a sample and feed those results into an algorithm that can confirm or contradict a clinician’s diagnosis. In theory, this could even be done at a patient’s local hospital to confirm a diagnosis even more quickly.
Bahn and her collaborators are in the process of setting up clinical trials for this test. Even so, her expectations are modest. “I believe in it. I haven’t got the evidence to prove it at the moment,” she says. If the results are clear-cut, she hopes to bring the test (or something like it) to the clinic in the next five years.
Martins-de-Souza, the Brazilian researcher who has worked with Bahn in the past, is skipping the diagnosis altogether; to him, the most pressing need is for a test that can rapidly identify which treatment would work best for a patient. He and his team have identified a series of proteins that change when schizophrenia patients start taking antipsychotics, and change differently depending on whether or not the medication is effective. They don’t know why it happens.
But it might mean that they can develop a blood test to help clinicians to give a patient medicine that works on the first or second try. So far the sample size has been too small to say definitely that the blood test works, Martins-de-Souza notes, so he and his collaborators plan to try it on more patients in Brazil, Germany, and the U.S. If those tests go well, the test is still five to 10 years from arriving in clinics.
The varying focuses of Bahn and Martins-de-Souza’s research hint at a debate going on just below the surface within the psychiatric community: The trouble with diagnosis. So little of the diagnostic criteria is based in biology, and many experts use this fact to explain why many patients receive delayed or inadequate treatment for psychological disorders, or why nearly half of psych patients are diagnosed with more than one disorder.
This has prompted experts to debate how we define these conditions in the first place. In 2013, the National Institute of Mental Health (NIMH) withdrew its support for the DSM-5, the manual that clinicians most commonly use to diagnose mental illness. More clinicians are starting to use the ICD-10, the diagnostic criteria put in place by the United Nations and more common in countries other than the U.S. (it’s like the metric system for mental health), though some posit that it doesn’t differ substantially from the DSM.
Organizations that hold the purse strings for research support this push towards biology. As part of its Research Domain Criteria initiative, the NIMH is putting more funding into research that goes beyond the scripted bounds of specific diseases. “We encourage researchers to take an integrative approach and measure things using more than one unit of analysis,” says Sarah Morris, the program chief of the schizophrenia spectrum disorders program at the NIMH.
That would likely include biomarkers, too, especially if they could help clinicians detect mental illness before symptoms begin. Some experts think that mental illness can even be prevented if it’s detected early, monitored by a clinician, and even preemptively treated.
As scientists get closer to understanding what causes mental illness, our definitions of those illnesses could very well change. Biomarkers, both for diagnosis and for treatment, might help fast-track that understanding, or at least bring the benefits to patients more quickly.
“The technology may or may not be right. But we’re going to keep trying,” Bahn says. “It’s too big a problem to not do something about it.”