Let’s imagine that you decide you want to run a marathon. You haven’t run in six months, and you know you’ll have to train.
On the first day of training, as you run your first mile a friend pops out of nowhere and says, “You’re sweating! That means you’re going up to Mile 14 today! Good job, you’re on your way!”
You may appreciate the encouragement, but does the friend’s assessment have anything to do with your physical reality? Before you’ve run one mile, can an observer of your sweat say with certainty that you’ll run 14 miles on your first day of training?
Yeah. That’s how I felt when reading this 2014 UK study that found:
“Adolescent boys who have high levels of stress hormone ‘cortisol’ along with some symptoms of depression are at a 14 times higher risk of the condition than their peers.”
The researchers latched onto teenagers (12-16 years old, mean 13.7) to assess a psychiatric condition. The researchers stated that a physical effect as common as visible sweat was a biomarker that predicted where some of the teenagers were going with their lives.
The study’s only physical measurements were cortisol from saliva samples at 8:00 a.m. on four consecutive days, then repeated a year later. For comparison, the standard lab test is to measure cortisol from saliva taken four times in one day at 9:00 a.m., 1:00 p.m., 5:00 p.m., and 9:00 p.m.
Cortisol is an effect of multiple potential causes, including stress, which itself is often an effect of multiple potential causes. One common cause of stress and its cortisol byproduct is diet, for example, when a person consumes caffeine.
“Mean time between waking and morning-cortisol collection was 50 min.”
I found it hard to believe that teenagers who:
- woke up at 7:10 a.m.,
- gulped down who knows what for breakfast,
- got ready for, and then
- went to school for an 8:00 a.m. cortisol test
wouldn’t have relatively “elevated morning cortisol” from the resultant stress.
The subjects self-reported depressive symptoms via a 33-item questionnaire initially and again every four months. The subjects were interviewed for psychiatric diagnoses.
The largest separator used for stratification within the subjects was an autobiographic memory test. Without this test, the study wouldn’t have made the main finding, so let’s look at the test’s details:
“Anxious and depressed adolescent patients report significantly elevated levels of over-general categoric memories compared with well controls. Six positive and six negative words are presented on flashcards in pseudorandom order, and participants are instructed to recall a particular memory of an event in their life after each word. Sixty seconds were allowed for each response.
Responses were categorized as specific if they referred to an event with a specific time and place, lasting no longer than 1 d[ay]. Responses were considered overgeneral if they formed a general class of repeated events.”
We can see that the autobiographical memory test only considered the subjects’ verbal expressions – within a short time period – of their recalls of emotionally triggered memories. As informed by the principles described in Agenda-driven research on emotional memories, the recall of an emotional memory is a product of the cerebrum responding to input from limbic system and lower brain areas. When someone describes their recall of an emotionally triggered memory, it’s yet another level further removed from the brain areas that store emotional memories.
We can also see that the test scores of the subjects’ verbal expressions aren’t capable of providing any etiologic evidence for an effect of high cortisol. A correlation is the best that could ever be shown by an autobiographic memory test. Again, the study’s main finding hinged on this third-order observational method of trying to figure out what’s going on in the subjects’ brains.
The researchers developed a control group, and made only a token attempt to trace the control group teenagers’ histories:
“The primary caregiver was interviewed about the quality of the family environment in three epochs (0–5, 6–11, and 12–14 y of age).
Four classes were found: optimal class, aberrant parenting, discordant, and hazardous.”
Were we supposed to believe that any primary caregiver would tell the truth about anything in a teenager’s history that indicated they had damaged their child? Good luck with that.
Anyway, the researchers didn’t act as though the teenagers’ histories had any significant relationships with any present or future conditions of the teenagers. The researchers showed their ahistoric biases by subsequently processing the entire history of each of the control group teenagers into a 1 or a 0 for the model.
The researchers then modeled this binary assessment to be relevant to the study’s main subjects!
The researchers’ agenda led to predetermined findings. Was the reviewer onboard with this agenda?
- By disregarding the study’s main subjects’ histories, the study couldn’t provide etiologic evidence for any present or future effects.
- By measuring only early morning cortisol, are we surprised that the model’s numbers could be processed into some correlation?
- Comparing this sole measurement to the 325 measurements taken of the subjects in Assessing a mountain climber’s condition without noticing their empty backpack made me wonder about the study designers’ real intentions.
News coverage of the study jumped on its flimsy finding to demand that something must be done. What did the researchers offer the teenagers who needed help?
- After citing research that:
“Showed null effects for two active treatments [cognitive behavioral therapy (CBT) and attentional training, respectively]”
they recommended some unspecific:
“New models of public mental health education and intervention in the youth population.”
- After citing research that found:
“Current diagnostic classifications [e.g., the Diagnostic and Statistical Manual for Mental Disorders (DSM) and the International Classification of Diseases (ICD)] have proved to have low diagnostic validity for investigations on the etiology, prevention, or treatment of MD [major depression]“
the study relied on these diagnoses anyway, and then disclaimed:
“It may also be the case that current classifications, as used in this study, such as DSM and ICD are simply not optimally specified.”
The researchers didn’t make their case that the “elevated morning cortisol” effect was an adequate biomarker for teenagers who needed help. The researchers did a disservice to their subjects by neither investigating nor providing any etiologic evidence for the observed effects.
Who really benefited from the underlying agenda? I didn’t see that it was the teenagers who may have actually needed assistance.
Did the study’s funders know that these efforts had enormous lacks? And what did:
“New models of public mental health education and intervention in the youth population”
http://www.pnas.org/content/111/9/3638.full “Elevated morning cortisol is a stratified population-level biomarker for major depression in boys only with high depressive symptoms”