Monday, August 15, 2011

A Blood Test for Depression? Really?

This month's Psychiatric Times has an interesting article (requires a subscription) about putative blood tests for diagnosing schizophrenia and depression. Because most of my patients have either depression or anxiety, I skipped to the section on a new depression blood test marketed by Ridge Diagnostics. Called the "MDDScore," the test measures 10 biomarkers, including cortisol, brain-derived neurotrophic factor, prolactin, and several other chemicals I never heard of. According to the Psych Times article (which was based on a presentation at the APA's annual meeting in May), the test has been validated by comparing 80 depressed patients with 50 healthy individuals. A test score of 6 to 9 was "highly predictive" of depression, and statistically differentiated the depressed from the healthy.

But I'm skeptical about this test. Why? First, none of the data have been published in a peer-reviewed journal, though I presume some have been submitted. Second, how useful is a depression blood test if all it can do is to differentiate depressed people from healthy people? I'll wager that my teenaged daughter would be just as accurate as the MDDScore in differentiating a depressed person from a healthy person based on a 5 minute conversation. Imagine buying an electromagnetic field detector to determine whether a shirt is red or blue. Sure, it will work, but the eyes can do it more quickly and for free. 

Here are a couple of crucial questions for Ridge Diagnostics, and for any company hawking a depression blood test:

1. Can the MDDScore differentiate depression from any other psychiatric disorder, such as anxiety disorders, adjustment disorders, bipolar disorder, substance use disorder, ADHD, and psychotic disorders? I'm pretty certain that the necessary studies have not been done. If the test cannot distinguish different psychiatric problems, then the MDDScore is simply a non-specific "biomarker" for emotional difficulties of all stripes, and would be essentially useless.

2. Does the MDDScore detect depression as well as the PHQ-2, which consists of two simple questions: "During the past month have you felt depressed or down? During the past month have you been bothered by having little interest or pleasure in doing things?" A high score on these two questions is 83% sensitive and 92% specific for depression (see reference here). Can this $745 blood test even approach these numbers?

Until Ridge Diagnostics can provide compelling answer these two questions, the money paid for the test is, more than likely, money down the drain.


Dr John said...

Such tests will never be of any real value in psychiatry because they change nothing. Its like having a test to confirm if I like the taste of beer. If the tests comes back negative am I to believe the test which tells me I don't like beer or my own taste buds and the mountain of empty cans at my feet?

Dr Carlat does an excellent job pointing out its problems but its limitations go far beyond data being published in peer review journals. Lots of shit gets published in peer reviewed journals.

If the test was tied to a confirmed understood pathology that had an understood TX based on the pathology itself, well that would be something. I would not hold my breath ever expecting to see this kind of thing in psychiatry.

This test would appear to have both high sensitivity and specificity for gullibility however.

pacificpsych said...

Ooh, a REAL test. Shiny.

I wasn't quite sure just how I felt about this. Luckily, I have here my AllMoodsBloodTest™. I tested my blood and here are the results:

10% despondent
10% bemused
20% amused
15% irritated
10% exasperated
10% calm
35% numb

All in all, the result read as "comfortably numb".

Let them have shiny. What can you do.

Unknown said...

Oh that it was as simple as detecting a blood-born pathogen. When we can't even agree how to define depression from decade to decade, how can a blood test?

Anonymous said...

Statistically speaking, it's a joke to have a blood test with 10 components evaluated in such a small number of test subjects.

Anonymous said...

Danny, is it legal for a company to market such a test? I'm not sure what the FDA rules are on diagnostic tests.

Asslete said...

I actually like the construct. I've seen some patients whose depression appears to be more "biological' than "psychosocial". If such a test existed that could adequately measure their tomcruisian factor, then clearly psychiatry would be legit. Or maybe not.

Take a closer look at Ridge Diagnostics. Their CEO is a former medical products marketer. So does that mean the test is bunk? Of course not...but one can certainly expect hype and answering critical questions in the style best described as "the fog". For their FAQ:

Will MDDScore differentiate MDD from bipolar?

MDDScore describes the likelihood that a person has MDD (i.e. a score of 9 indicates a greater than 90% likelihood that the person has MDD; a score of 2 would indicate a less than 20% likelihood of MDD). The “score” as it stands currently helps to “rule-in” or “rule-out” MDD. Creating a biomarker panel to differentiate between unipolar and bipolar depression is under development.

But this of course is compared to "healthy" people. I'd be most interested in what defines "healthy" and what someone with dysthymia, PTSD, Bipoar, etc would score on this test.

To the question regarding FDA is their response:

Is the MDD Score FDA Approved?

MDDScore has met all regulatory requirements to be offered as a laboratory service in a Clinical Laboratory Improvement Amendments (CLIA)-certified laboratory.

At this point in time, it is far too early to condemn this test...but it also appears to be far too early to endorse it.

pacificpsych said...

>>At this point in time, it is far too early to condemn this test<<

It's never too early to condemn junk, bunk and quackery. It just won't help, just like you can't argue the notion of chemical imbalance with anyone, it's far too entrenched in the public's mind as a 'fact'.

Perhaps if our predecessors, who were active 20-30 years ago, hadn't caved into this nonsense, things would have been different. But they embraced Pharma's rubbish and never stood up to them, so here we are, and it's far too late to do anything about the state of psychiatry. So make up tests, stick 'em, prick 'em, who cares.

The patients do not appreciate psychiatrists who try to fight this - they love the medicalization even if it's fake medicine based on false premises. Don't matter to them. Go with the flow, man, no use fighting it.

Asslete said...

I understand the disdained perspective of "chemical imbalance". But to deny the potential existence of a meaningful biologic construct to explain at least some depression (emphasis on some), makes you a psychiatric atheist

I, however, am more of a psychiatric agnostic. I am unknowing...but open to the possibility.

I am skeptic, not a cynic. And as such, I will await actual studies/data before dismissing the test. I will ask questions and not assume I know the answers.

Unknown said...

You can buy a Mood Ring for $2.99 from Amazon and get instant results.
My question for Ridge Diagnostics is
"How much of a kickback am I going to get if I order this test for my clients?"

Unknown said...

Forgive my previous snide remarks. After thinking about it for awhile, now I get it. This test is for the fifteen minute psychiatrist who does not have the time to conduct the psychiatric interview - takes too much time. Remember the psychiatric interview? One of the questions you ask is "Are you depressed?" Usually, most patients have no problems answering this question with a yes or no. It is one of the first questions I ask. Why do I have to wait for a blood test when I can get an instant answer?

pacificpsych said...

Of course there are biological reasons for depression. Hypothyroidism, anemia, malignancy, name but a few.

That's why you rule out medical causes for depression first.

This sentence from their home page killed me. Apart from being a repetition of the bio mantra (yawn), note who they cite! LOLOL.

" ...I think it would have a huge affect on the field... one of the great challenges in psychiatry is the absence of hard definitive tests to diagnose a disorder. They are all based on clinical questions and clinical assessments which lead to many not believing that they are real diseases..." (Psychiatrists)

Martin said...

This is a great debate. From a patient perspective it's hard to see the point of this, unless as has been pointed out it is able to distinguish between various possible diagnoses and thus aid the choice of treatment.

But even so, I would hope that all you psychs out that can do that without a blood test. I guess that raises the question whether the overworked GP/family doctor can do so. Maybe this would be useful for the.

However, I tend to agree with pacificpsych. The last thing people with depression need is more "junk, bunk and quackery".

Asslete said...

I agree with the need for less junk, bunk and quackery. And based on past topics, I tend to agree with most of pacificsych's view. But there is a sense of closed-mindedness and logical fallacy that we psychiatrist who are dubious of our own field fall into.

Per your analogy, biomarkers such as TSH and Hematocrit could be useful in diagnosing depression secondary to hypothyroidism or anemia. So perhaps Serum Acylation Stimulating Protein, S100B and Resistin might be likewise useful if further research supports this? Sure it is probably just correlative noise...but shouldn't data and not belief systems determine this?

My point is...with an n of 80, and a comparison of "healthy" vs 'depressed", we simply have too little information. So I remain agnostic, and as such, refuse to condemn the test at this point in time.

But that doesn't prevent me from condemning those who are already out promoting it for having a utility which has yet to be validated.

Joel Hassman, MD said...

Catering to the plans set by Obamacare, brought to you by people who would like to see psychiatry become extinct, so people will spend even less time in the doctor's office to get care for mental health needs.

Check out Dr Grohol's post at re the new definition for addiction by the American Society of Addiction Medicine. Let's keep dumbing down the etiology of mental health problems to a point that may reignite the evils of eugenics.

Is Biopsychosocial an extinct term among collegues as of 2011?

By the way, a post since the last American Holiday, should we be glad to know you are still invested in this blog?

Dr John said...

Anybody who buys into this nonsense for a second should spend some time reading basic philosophy of science and what the difference between good science and this is.

This junk test clearly falls under the category of "cargo cult science".

This term was coined by the great physicist Richard Feynman. He based the phrase on a concept in anthropology, the cargo cult, which describes how some pre-scientific cultures interpreted technologically advanced visitors as religious or supernatural figures who brought boons of cargo. Later, in an effort to call for a second visit the natives would develop and engage in complex religious rituals, mirroring the previously observed behavior of the visitors manipulating their machines but without understanding the true nature of those tasks. Just as cargo cultists create mock airports that fail to produce airplanes, cargo cult scientists conduct flawed research that superficially resembles the scientific method, but which fails to produce scientifically useful results.

He was clearly making reference to psychiatry when he warned folks to be on guard for this stuff.

We have "tests" for something we have zero understanding of to begin with. That is pure and simple the essence of "Cargo Cult Science"!

SteveM said...

Seems to me the most glaring Ridge oversight is the lack of evidence of specificity that Dr. Carlat mentions.

I mean cortisol levels are elevated under any kind of stress right? Who's to say that the other Ridge physiological measurements aren't also stress related regardless of the psychiatric disturbance. Or even the non-pathological stress of an episodic event like giving a speech or playing in a big game?

Seems like selectively sloppy science to me.

Parenthetically, this is another example of medical diagnostics getting way ahead of actual clinical utility.

On the Technology page of the Ridge site is this quote:

"A consistent change in these patterns is also used to evaluate the efficacy of antidepressant therapy."

That's a hoot. If a patient's Ridge numbers are improved, he won't be depressed whether he feels that way or not...

John M. Nardo MD said...

The point of a biomarker from where I sit would be either to establish a biological etiology or pathophysiological pathway - maybe to allow us to refine some hopelessly heterogeneous diagnostic category like "MDD." It is a sign of the times that this piece of fluff has been introduced as a commercial product rather than for some useful scientific purpose. Be prepared for more of this multi-marker stuff from Brain Resources and from Trivedi's NIMH study - both aiming for "personalized medicine" avenues for picking drugs.

Asslete said...

Biomarkers will become more and more prevalent in the next decade. Whether any of them can separate the wheat from the chaff when it comes to MDD remains to be seen, but I remain open to the possibility that other organic causes exist for depression beyond hypothyroidism, anemia, etc.

So I will await more information on the MDDScore, but it has much of the nonsensical hyperbolic verbiage that one sees in CAM pseudoscience.

Adding to SteveM's earlier observation...the Technology page makes this claim:

"These biological markers are mapped onto a multi-dimensional hyperspace model to create the hyperspace vectors that construct and define the patterns that indicate major depressive disorder. "

WTF! Hyperspace? Vectors? Shouldn't their be some "Energy Field Alignment" as well?

But hey...I am optimistic because "A consistent change in these patterns is also used to evaluate the efficacy of antidepressant therapy"

Super...except that in their own FAQ they note:

"If a patient is currently being successfully treated with an antidepressant, preliminary studies suggest it is likely they will have an MDDScore that is indicative of a higher than normal likelihood of having MDD."

So whether they are well treated or poorly treated, their score is the same??? How does that help me evaluate efficacy. I have a way to "validate" they are depressed, but no way to "validate" their improvement.


Anonymous said...

The test is obviously a bunch a BS simply by the premise. Like Dr. John stated, it's like having a test a blood test for alcholism and not bothering to ask how much a person drinks. All DSM diagnoses are based on clinical scenarios. It reminds of the quacks in a "neuropsychiatric" practice in the northern Chicago suburbs who routine use qEEG to diagnosis everything from substance abuse to eating disorders. The bigger problem though is the merging of consumer driven medicine and the internet. As someone else commented, people don't want to hear that these special tests and "latest breakthroughs" are a bunch of bunk. People want what they want and consumer driven medicine will give it to them. What's amazing is when even many of our own educated colleagues cannot see through this crap.

Scott A. Oakman said...

But if we have a blood test, then we can tell our moms that we're "REAL Doctors"!

R. Thoryke said...

No matter what markers they measure, we still don't have validation that those values of those markers definitively = depression. So no, their test cannot have been FDA-approved for the diagnosis of depression. A quick check of recent approvals should confirm this. Ridge Diagnostics can only say that the tests they use are approved laboratory techniques for assessing the levels of markers in their magic list. It's a new variation on the Quack Miranda Warning, but it's walking and talking like the same duck.

Daniela said...

I'm also pretty skeptical. blood testing may find a few preliminaries signs for depression, but I believe depression is a much bigger issue than just genetics or physical health, and blood tests can't exam mental state.

Anonymous said...

While I agree that for the majority of patients, this would be useless, I could see a few roles for a blood test for depression:

1. Non verbal patients, patients with low IQ, catatonic patients

2. Patients who have 1-2 weeks of depressive symptoms. A test that could predict who would get better (and have an adjustment disorder) or who will develop MDD and need treatment

Sean Paul said...

Wow. I really can't believe the unverified mumbo-jumbo that these companies put out to try and drain peoples pocket books....

This is not evidence based, nor makes any logical sense.

emagine said...

depressed test can help you get an idea if and if so, how depressed you are . There are several prestigious and credible self-tests that you can take online
1.) CES-D - Center for Epidemiologic Studies Depression Scale
This depressed test was created by Lenore Radloff of Utah State University. It is a 20-question test that helps measure your depressive symptoms by the feelings you have had in the past week.

2.) QIDS-SR - Quick Inventory of Depressive Symptomatology - Self-Report
This 16-question depressed test was created by researchers at the University of Texas Southwestern Medical Center. It is based on 9 different criterion domains and measures 16 factors to determine a person's level of major depression.

3.) Geriatric Depression Rating Scale
This depressed test is a 15-question test with all answers being "yes" or "no." This depressed test helps to measure the level of depressive symptoms particularly in older people.

4.) K10 Test for Psychological Distress
Created by researchers in Australia,

This 10-question depressed test helps to measure the level of psychological distress you are experiencing due to anxiety and depression that you have experienced in the last 4 weeks.

Anonymous said...

To me it is absolutely shocking and alarming to see all the negative responses. The bottom line is that a much greater understanding of the biological illness is desperately needed for both better diagnoses and the development of more effective and efficient treatments. Science is supposed to be the cornerstone of medicine. Current symptom based diagnoses is highly problematic and subjective. Would society accept a Cancer diagnoses without looking for underlying tumors and biopsies or an HIV diagnoses without a blood test??? The status quo for treatment and diagnoses is clearly not working for a large number of patients - and the science and treatment has hardly evolved in the past 20+ years - while great strides have been made in most other medical specialties. This may be acceptable for practitioners - it is not acceptable for patients. Why is this so hard to understand?