Mental Health Diagnoses May be Based on Context, Not Symptoms

New study examines the biasing effect of negative contextual information on mental health diagnoses


A new study published in the journal Clinical Psychological Science examines the biasing effect of irrelevant contextual information on judgments about a person’s mental health. The researchers found that if the context provided was negative—even if that context appeared unrelated to the mental health symptom—then participants were more likely to rate a child as having a “disorder.”

Photo Credit: Flickr

The researchers go on to note that this use of context is actually the opposite of how the DSM recommends context be used. In the DSM, context is to be considered as providing alternative explanations for distress—that is, diagnoses should be ruled out if a negative context could explain the symptoms.

However, the participants in this study did the opposite. They tended to rule out the diagnosis more often if there was a positive context for the child—that is, if context did not seem to “fit” the participant’s idea of what a mental health concern looks like.

The research was conducted by Jessecae K. Marsh at Lehigh University and Andres De Los Reyes at the University of Maryland at College Park. The same authors have previously conducted several studies on this effect.

According to Marsh and De Los Reyes, “When people encounter individual symptoms of a disorder, they are not interpreted independently of the surrounding context but rather will be seen in the light of that context.”

That is, “symptoms” of ‘mental illness’ are interpreted based on negative, neutral, or positive information about a person—information that may have nothing to do with the mental health concern being diagnosed.

The authors write that “In a real-world setting, these easily observable elements of a child’s life are profoundly influencing the way symptoms of disorder are interpreted.”

The previous studies examined conduct disorder and ADHD, which are known as “externalizing” disorders because they depend on acting-out behaviors. A study they conducted in 2011 found that mental health clinicians were more likely to diagnose conduct disorder in children if given negative contextual information about that child.

The current study consisted of two experiments. The researchers were interested in examining whether their previous conclusions about the biasing effects of context were limited to “externalizing” disorders like ADHD and conduct disorder, or whether contextual information would bias the diagnosis for “internalizing” disorders like panic disorder as well.

In the first experiment, 120 participants read a list that included three pieces of contextual information as well as a single symptom of panic disorder. In all cases, the contextual information was not relevant to the diagnosis. For half the participants, the contextual information about the child was negative, while the contextual information provided for the other half was neutral.

The authors provided the following example of the contextual information:

  • “His parents have had a difficult time finding him a regular babysitter because his parents need a sitter at inconvenient times during the day.” (neutral)
  • ““His parents have had a difficult time finding him a regular babysitter because people often refuse to babysit him more than once.” (negative)

The researchers found that the participants were more likely to overlook the symptom when given neutral information. When given negative contextual information, the participants were more likely to diagnose the child with a disorder.

It is important to note that in both cases, the actual symptom was the same. Only contextual information, designed to be unrelated to the diagnosis, was changed.

In the second experiment, Marsh and De Los Reyes attempted to determine if the contextual influence on ADHD diagnosis was greater than the contextual influence on panic disorder diagnosis. 149 participants were randomly assigned to one of three groups. Each group read lists with contextual information. Group 1 also had a symptom of panic disorder; group 2 had a symptom of ADHD, and group 3 had a symptom of both panic disorder and ADHD.

The researchers found that participants were more likely to use negative context to diagnose ADHD than panic disorder. The authors write that this was as expected—as an externalizing disorder, ADHD involves interaction with the environment/context, so it was to be expected that participants would use contextual information to inform this diagnosis.

However, it is still important to note that this contextual information was not pertinent to the actual symptoms of ADHD—meaning that participants were much more likely to conclude that the child had ADHD if the child was described as difficult or unlikeable than if the child was described neutrally.

The researchers again found that context influenced the diagnosis of panic disorder, as well—but not to the extent found with ADHD. A limitation of the current study is that the participants were not mental health clinicians. However, previous studies indicate that this effect is present for clinicians as well.

Additionally, Marsh and De Los Reyes suggest that much of the evidence used by clinicians to evaluate a child for ADHD and conduct disorder is based on the reports of parents and teachers (in the current study, about half the participants were parents, and about a third were teachers). If parents and teachers tend to base their reports on contextual information that is not relevant to the criteria for the disorder, clinicians will likewise be biased toward diagnosis based on this information.

The authors call for further research, because “Understanding more about the influence of contextual factors brings us closer to understanding the real world demands of seeing patients and decoding what mental health issues they are experiencing.” They add that the issue of diagnosis is complex, and it is yet unclear how context should be used in the process.



Marsh, J. K., & De Los Reyes, A. (2017). Explaining away disorder: The influence of context on impressions of mental health symptoms. Clinical Psychological Science, 1–14.


  1. This is exactly what happens when you create “diagnoses” that are based on subjective criteria. People “diagnose” what they don’t like or don’t understand. The DSM can say whatever it wants about how it “should” be used. People will use it in the same way people use most things – they will act in ways that are self-centered and that avoid discomfort. There will be exceptions, of course, but the incentive is there to blame the victim, and the DSM makes it very comfortable to do so. It needs to be banned!

    Report comment

    • “… the incentive is there to blame the victim, and the DSM makes it very comfortable to do so. It needs to be banned!” Absolutely. And the “victims” who are blamed the most are child abuse victims.

      We are actually living through an American psychiatric holocaust of child abuse victims. Germany lived through a psychiatric holocaust of Jews. Bolshevik Russia lived through a psychiatric holocaust of “political dissidents”/Christians who did not want communism. When will the psychiatric holocausts end?

      The DSM is a classification system of the iatrogenic illnesses that can be created with the psychiatric drugs. The ADHD drugs and antidepressants can create the symptoms of “bipolar.” And the “bipolar” and “schizophrenia” drug cocktails and drugs can create what appears to the DSM believers to be both the negative and positive symptoms of “schizophrenia,” or the depression and mania of “bipolar.” Except these symptoms are actually created by the drugs, via neuroleptic induced deficit syndrome and anticholinergic toxidrome. But, of course, these known psychiatric drug induced illnesses are not listed in the DSM billing code “bible,” so they are always misdiagnosed as one of the billable DSM disorders.

      The DSM is not a classification system of real “genetic” illnesses. The DSM needs to be banned! How long will this take? It was declared “invalid” in 2013 by the head of NIMH. It’s 2017, why is it still in use?

      By the way, it is not in the best interest of our society to have a multibillion dollar, iatrogenic illness creating, child abuse covering up system, which is what today’s “mental health” system is, according to their own medical literature. Thank you to the psychiatrists and psychologists for empowering the pedophiles (sarcasm), by turning millions of child abuse victims into the “mentally ill” with the psychiatric drugs. Pedophiles now rule the world.

      But multibillion dollar “dirty little secrets” do die hard, don’t they doctors and religious leaders?

      Report comment

      • There’s a quote from Ethan Watters, from the book “Crazy Like Us, The Globalization of the American Psyche” that struck me back when I read the book, and has stayed with me ever since, that basically says “once you start looking at someone through the lens of a Western Psychiatric Diagnosis, it’s very hard to stop”. That’s basically what has happened, our society has put on Western Psychiatric diagnoses glasses at some point (started earlier, but mainstreamed in the 90s?) and has left them on.

        It’s not just doctors or mental health professionals who do it, either. Your average Jane and Joe are doing it now too, everyone’s become this “expert” on diagnosing all of a sudden. Looking at people as walking clusters of symptoms rather than people.

        Although I do understand the history of psych hospitals and their barbaric practices, and when the psychiatric drugs came on the scene, this jump to diagnose people as really been mainstreamed in recent years. I remember even 15, 20 years ago there were clinicians who wouldn’t diagnose someone with a personality disorder until they were at least 30, since people oftentimes outgrown certain behaviors, and the idea was to work on those behaviors without giving a person a diagnosis that would stigmatize them for life. Or giving substance abusers at least a year before even thinking about diagnosing them with something else, since substance abuse causes changes in behaviors and personality, and you’d want to give them time to get the drugs out of their system, and get things in order with nutrition/ sleep/ social supports, employment etc. Now substance abusers are labeled “dual diagnosis” right off the bat, getting them diagnosed and on those psych meds from the get go. WTH

        And what’s happening to children, I agree, is beyond horrifying. The first groups home I worked in were DCF homes (1998/ 1999). Even at the time, these children, who came from backgrounds of physical, emotion, sexual abuse and neglect, where being drugged out of their minds. In what bizarro world does this make any sense? It’s funny people can see how lobotomies of the 40s and 50s were barbaric and cruel (and people listened to their doctors then too), but chemical lobotomies (which is what is going on), they are somehow okay with? And why is everyone off the hook on this? Everyone plays their part. Parents , psychiatrists, school systems, medical doctors, Pharma companies, our corrupt government (which I really wish we would stop giving so much power to), yet everyone is somehow off the hook when it comes to taking responsibility for this? Why?

        I agree with you guys who say the whole system is corrupt and they should toss out the DSM, but I personally believe it should start with stopping with the diagnosing/ labeling and drugging of children. Work on the behaviors and figure out what’s going on in the child’s life, medically perhaps, if something is seriously “wrong”, but this diagnosing/ drugging business has got to stop. There are alternative practitioners who can help (Naturopathic Physicians, Orthomolecular Psychiatrists), unfortunately probably not covered by insurance (at least not anymore with what’s happened to insurance in the US the past 8 years unless you are lucky), but a better path to take than getting caught up in this poisonous/ toxic mental health system web that’s very difficult to break free of, if at all, once you’re caught up in it.

        Report comment

  2. After reading this post, I flipped through my copy of the DSM-5. There is a brief section (pp. 19-24) titled “Use of the Manual” in which context is briefly mentioned. The manual states that a “case formulation” (which is distinct from a diagnosis) should include assessment of the broader context, and that a diagnosis should not just involve checking off symptoms but requires clinical judgment about the “relative severity and valence of individual criteria and their contribution to a diagnosis”. But this section has no language that supports the researchers’ contention, noted by Peter, that “In the DSM, context is to be considered as providing alternative explanations for distress—that is, diagnoses should be ruled out if a negative context could explain the symptoms.” I found this statement surprising because my well-informed understanding of the DSM indicates that diagnoses are in fact *context-free by design.*

    I had a look at the article Peter reviewed to see what the authors had to say about the supposed contextual basis of DSM diagnoses. Here is the relevant quote:

    “The Diagnostic and Statistical Manual of Mental Disorders (DSM; American Psychiatric Association [APA], 2013) provides some guidance for several specific disorders on how contextual information may inform diagnostic decision making. For example, the DSM suggests that a diagnosis of conduct disorder should not be provided if a symptom displayed by a child could serve as an adaptive reaction to the environment in which the child is living. In this way, if a child lives in a violent neighborhood, the disorder symptom of “getting into fights” could be explained away as a normative reaction to his environment. Likewise, previous DSM editions suggested a diagnosis of major depressive disorder was not warranted if symptoms were in reaction to the loss of a loved one (e.g., APA, 2000). What these examples have in common is a focus on context as a causal explanation for clinical presentations. That is, if a person displays symptoms that appear to be explained by a reaction to a contextual trigger, then a diagnosis is unwarranted.”

    The paragraph above is incredible, for several reasons. First, I’ve just read the entire conduct disorder section in the DSM-5 and the only relevant language is this: “Conduct disorder diagnoses may at times be potentially misapplied to individuals in settings where patterns of disruptive behavior are viewed as near-normative (e.g., in very threatening, high crime areas or war zones). Therefore, the context in which the undesirable behaviors have occurred should be considered.” This is obviously very different from the authors claim that the DSM-5 says the diagnosis should not be given if the behavior in question is an adaptive response to the environment. The DSM-5 in fact says no such thing. It doesn’t even say the diagnosis should not be given, just that the context should be “considered,” whatever that means.

    Second, the authors chose DSM-4 major depressive disorder as an example of the DSM’s sensible use of context because bereavement was included as an exception. Incredibly, the authors didn’t mention the fact that in DSM-5, this exception was eliminated and a top-level decision was made to make the diagnosis context-free!

    The authors of this article, building on the paragraph I quoted above, added this gem:

    “An underlying theme in the approach of the DSM is that a person will think about a diagnostic conclusion and then alter that diagnosis depending on how context could be used to causally explain those symptoms. In these types of cases the use of context seems to be an active choice made in the final steps of diagnosis.”

    I can only presume that the authors don’t actually understand the DSM and simply assumed based on common sense that it must work this way, or deliberately cast it as working this way to create a nice narrative for their study. But in reality, it most emphatically does not work this way. The context-free nature of nearly all DSM diagnoses is ludicrous and studies like this one shouldn’t obscure this important reality.

    Report comment

    • The study on the effect of negative contextual information was flawed. Subjects read about a child. The descriptions included one symptom of panic disorder or ADHD, and some “contextual” information. The contextual information was varied and any effect of the variation on subjects’ tendency to believe that child suffered from panic disorder (study 1) or ADHD (study 2) was captured.

      “Only contextual information, designed to be unrelated to the diagnosis, was changed.”

      The negative contextual information was supposed to be irrelevant to the “disorder” for which a symptom was also included, but it wasn’t.

      The negative contextual information in the panic disorder study was that babysitters often refused to babysit the child a second time. That is not a statement about babysitters, who rely on repeat business, it is a statement about the child. If we can assume that baby sitters’ avoidance of the child was not triggered by his fashion tense or taste in coloring books, it must be about his behavior. It is unusual for a child’s behavior to repel babysitters, so the behavior of this child must be extraordinary. Panic disorder wouldn’t be the first thing subjects would volunteer in a free-response format, but if panic disorder and no disorder whatsoever were the only options, the baby-sitter repelling quality of this child would cause subjects to affirm a diagnosis of panic disorder.

      And this is a mess, too:

      “…participants were much more likely to conclude that the child had ADHD if the child was described as difficult or unlikeable than if the child was described neutrally.”

      Children diagnosed with ADHD are diagnosed because they are, in one way or another, “difficult.” Being a difficult child generally means being hard to like, which means that ADHD-diagnosed children, because they are by definition “difficult,” are also unlikeable. I don’t see how the negative information could considered irrelevant to a layperson’s guess about the presence or absence of ADHD. Then again, I also don’t know why anyone should care about that.

      The supposedly irrelevant information had a greater tendency to cause diagnosis in the ADHD study than in the panic study, but there’s no way to attribute that to the supposedly internalizing nature of panic disorder and the supposedly externalizing nature of ADHD. To do that, the negative contextual statements would have to be the same in both studies. As it stands, it might be the content of the statements alone that increased the odds of diagnosis. If you really cared, and I hope no one does, you could run four conditions:
      panic, babysitter avoidance
      panic, difficult and disliked
      adhd, babysitter avoidance
      adhd, difficult and disliked
      There’s illogic throughout.

      And then this, which means the entire study should be nuked from human consciousness, even if it means we all have to go back in time from before it was conceived and go through whatever tribulations we have endured since that time. It is worth it for a world in which this study was not conducted or published.

      “If parents and teachers tend to base their reports on contextual information that is not relevant to the criteria for the disorder, clinicians will likewise be biased toward diagnosis based on this information.”

      How do you infer from parents’ and teachers’ behaviors that clinicians will be “biased” by irrelevant contextual information? Never mind that the information provided was not irrelevant. If you want to study clinicians, study them. Don’t make statements about their mental processes based on whatever it is you think you learned about parents and teachers.

      Having read and appreciated Brett Deacon’s analytical comment, I assert that the reported study should not have ended with a call for “further research” unless it’s on how to prevent confused and confusing studies about nothing. Less of this sort of thing would be better.

      Report comment

      • Wow. BetterLife, thanks for this excellent peer review of the article. I hadn’t read the entire thing, but your post lays bare the obvious fact that its publication was a failure of peer review. It’s incredible that a study with such obvious methodological flaws was published in a peer reviewed scientific journal.

        Report comment

  3. I have overheard the following microagressions stated about myself. “Patient claims she does not smoke.” This is biased since it is OBVIOUS I don’t smoke. I would stink of it if I did. I conclude that the nurses smoke. That’s why they cannot distinguish who smokes and who doesn’t.

    “Minor chest pains.” Fellow patient had depression diagnosis and was a frequent flier. They didn’t even check her vitals and they sent her home, telling her to call her therapist. She had a heart attack. She lived.

    Report comment

    • Julie, they were probably smoking cigarettes and doobies in the back of the building, and blaming medical things the patients were reporting on “psych symptoms”. And let me guess, no one was held responsible for the negligence that led to the heart attack?

      Report comment

  4. the female heart attack victim was told “you are not good at managing stress.” She was sent to day treatment after her heart attack. So yes, the patient was blamed for the hospital’s gross malpractice. I spoke with her. Sadly, she bought it. Hook, line, and sinker, and the day program reinforced the coverup.

    Report comment

    • Wow, that’s terrible yet not unusual or surprising, sadly. We had one nurse who made mistakes like that constantly. One time she forgot to set up someone’s lithium bloodwork and she was going into lithium toxicity, and by the time they caught it the patient almost died. The nurse blamed someone else and didn’t even feel bad, just hostile and concerned about covering her own butt. So many incidents like that, though, that’s just one drop in the bucket. 🙁

      Report comment