Clinical biases which affect diagnosis accuracy

Post Reply New Topic

From time to time people post threads questioning the accuracy/inaccuracy of their diagnosis or non-diagnosis, often feeling shocked and bewildered over a recent one-off clinical encounter. Members fairly typically respond to these OPs with replies outlining criteria for diagnosis, often citing the latest version of DSM criteria (and/or earlier ones), or the importance of parent testimony etc, or (occasionally) with interesting observations about gender bias affecting diagnostic validity.

What is usually missing from those discussions/threads - interesting as they are - is any mention of the clinical heuristics which are known to affect clinical diagnostic outcomes, yet this is probably as pertinent, if not more pertinent, than the things which are usually mentioned.

I suspect the word 'heuristics' may baffle some but put very simplistically, you could say that it represents thinking in terms of probabilities rather than responding to a specific case. In the general topic of diagnosis, it's an important concept, which I hope will be of some interest - and this link outlines how diagnostic errors due to heuristical perception can occur (it's not specific to autism, it applies generally to clinical decisions):

http://www.msdmanuals.com/professional/ ... ion-making

Thank you for bringing it up. As a new and very confused member on WP, I've been reading a lot of threads about diagnosis troubles. I won't even go into the topics, as this would just start endless discussions off topic, but I really don't understand why people have so much respect for clinical assessment.

In the words of a researcher friend: "Doctors don't understand statistics!" This was after a statistics course with several medical researchers attending.

I have had a lot to do with doctors for physical health reasons. They generally have too little time to handle complex, vague or unusual problems. In one memorable case, the doctor in question gave me a 10-minute speech on how my problem was caused by a certain type of bacteria - even though I'd taken four blood tests that found no trace of said bacteria!

People tend to overestimate how rational they are. This goes for people with AS as well.

I've always believed that people in clinical settings have a lot of biases. Since I'm diagnosed with autism it's the only thing people really see. My other issues are just seen as comorbids. My "comorbids" are actually way worse than any of my autistic traits. I'm not diagnosed with ADHD, but I would be surprised if I just had ADHD with autistic traits instead of autism.

One of the clinical issues that causes trouble diagnostically in medicine arises from the almost pigheaded adherence to "Occam's Razor", which presupposes that there is always only one right diagnosis to be found, probably the simplest - (too bad if the patient has several or more than one disease simultaneously and one of them is a rare and obscure disease).

As an example, say the presenting symptom is a swollen tongue. This could be due to something relatively minor and transient - an allergic reaction or accidental biting during sleep. However in some cases it represents very serious diseases, which are hard to diagnose because they are not commonly looked for (like multiple myeloma, or amyloidosis). Or it could be mixed causation: a blood cancer and accidental biting during sleep.

Why do people put such faith in clinical diagnosis? Surely one reason is that the illusion of certainty is often a more comfortable state than fear of the unknown; and the illusion of certainty may reduce the uncomfortable state of cognitive dissonance. Yet also culturally (I think) there are fundamental issues of structural power, its uses and misuses.

Underwater, I agree with your conclusion that both the neurodiverse and neuronormative populations are similarly affected; although sometimes I wonder if the neurodiverse - because perhaps of a greater need for certainty arising from past trauma, misdiagnosis, bullying, rejection et al - may be more accepting/unquestioning of clinical bias..even perhaps including some of the HFAs who have generally very good pattern detection skills in other disciplines.

DevilKisses, nice to hear from you again! I totally get the point you make about your autism being the only thing they can see..

People diagnosing autism which involves observation and subjective and controversial diagnostic criteria are a lot more susceptible to bias that those diagnosing a physical condition.

Tony Attwood has discusses Aspies having "intellectual crushes" on very smart people. For people who want and order and strict rules, having a rule that experts have the answer and thier conclusions should be always chosen over laymen's personal experiences can become a rule.

NT's with thier generally higher body language reading ability would be more likely pick up an expert that is lying or not confident in thier conclusions.

This is why I really wish there could be some kind of physical test -- a blood test, a brain scan, a gene identified, or some other medically-based marker that proves to be accurate in showing autism, whatever that marker would be.

I think that whenever and wherever you have only human opinion to go on -- even the opinion of "trained professionals" -- you can have doubt, bias, inaccuracy, in either direction.

I would be happy to actually obtain more than one assessment from more than one professional if it were practical or affordable, and if there were ever to be found a biologically-based diagnostic method I'll be the first in line.

I do sometimes doubt even my positive diagnosis and I would like more assurance yea or nay.

"i havent been on here long enough to see how its generally handled but on the rest of the internet and in real life i see way too many people get yelled at for saying "i have ___" when its self diagnosed. and YOU CAN'T SELF DIAGNOSE THAT!! !! so people are probably likely to find a doctor without seeing how full of s**t they are first and believe what they say the problem is because its better than confessing to the internet that its just a feeling you've had all these years"

End quote.

One thing that the formally and informally diagnosed have in common is a lot of unhealed shaming in the pre-diagnosis past. (NB: I don't want to debate lexicology in this thread and have no intention of doing so yet again). Shame is the most difficult past, stored trauma to heal because generally, people are ashamed of having been shamed and too ashamed to reveal the depth and intensity of those powerful and heavy feelings. Any kind of validity which mitigates those internal feelings and memories of past shaming is likely to be fiercely asserted and defended.

The sad thing is that when people bury unhealed wounds of great intensity, they tend to do to others precisely what was done to them, though the topic or context may be different from the original. I think you could fairly safely say that nearly all people on the spectrum whatever their status have experienced very significant levels of past shaming compared to neuronormative populations, with one exception perhaps being children who grew up in daily punitive shaming religious families who believe that children are born evil and it must be purged out of them from the start. Kindness, validation, acceptance and understanding heals old shame. I wonder what the effect of past shaming is on clinical encounters and the outcomes of these.

_________________
I like the skull smiley

That's true -- there is a lot of witchhunt-like vilification of the self-diagnosed, who are told at the very least that professional diagnosis is the only way to be sure, and at the very worst that they themselves are full of sh!t unless or until they get a professional diagnosis.

While the problem remains that even a professional's opinion (which is all a formal diagnosis really is without a biological marker to back it up) may be full of sh!t too one way or another, for one reason or another.

Clinicians rely on the accuracy of what the client tells them and also on the behaviour presented in a clinical setting being representative of that client's behaviour in other settings.

I don't behave in a clinical setting with one or two professionals in the same way that I do in an unstructured social setting. I have a diagnosis but I am quite sceptical about the accuracy because the clinician wasn't very thorough.

Also the criteria themselves are quite malleable in the right hands. Some professionals have an interest in making more diagnoses because it can be profitable.

People seeking a diagnosis obviously have some sort of serious issues and they might not be able to get similar accommodations, supports or protections for mental illnesses or personality disorders. I like the idea of neurodiversity being applicable to more than just autistic people.

I also wish we made more room for acceptance of all human beings. And since it is a spectrum, and without intending to spark discord, but clearly a spectrum means there will be people near the cutoff who just are or just aren't and those borders tend to change, yet we're not really making room well to accommodate that reality and instead seek to say everyone is or isn't on the spectrum. Does not work well for many people that the lines blur and shift and I don't like the ignoring of those near that boundary.

Thanks for the nteresting and pertinent inputs to this discussion... another totally hidden source of bias, of which both parties to the clinical encounter are unaware, is unconscious bias on the part of the clinician. There is a large body of relatively recent research on this, and how widespread it is at decision-making/leadership levels, though not specifically re ASD diagnostics. We can probably expand the picture only by looking at decision making studies in other fields and disciplines.

A recent large well-designed study found that decision makers genuinely considered themselves to be rationally even-handed and unaffected by the 'common' biases - and often those working for them saw the decision makers like this too (as they tended to appoint people who like themselves, and to unconsciously create rational justifications for this). One of the questions the researcher was exploring was the failure to progress female candidates who were objectively as qualified (sometimes more) than the males into leadership roles. The leaders were sure that they were not biased and explained why at length, how they gave equal weight to CVs and equality of opportunity... however (! !) when they were asked instead "what does a potential leader look like, what qualities do you look for?" the whole picture changed: the qualities they unconsciously selected for were stereotypical 'male' qualities.

Another source of bias is all too conscious: prejudice. The clinician - for whatever reason - may not like the particular client and/or the particular group, for reasons that have nothing to do with the criteria, training, and heuristical studies suggest that clinical training is no guarantee of objectivity. You see this very clearly in medicine (internalised gender biases misdiagnose women with heart disease as anxiety disordered; internalised cultural biases diagnose illegal drug injectors with infect skin ulcers as just 'drug seekers'). Or you may look like/sound like/present like/think like/talk like someone in the clinician's life whom they personally detest - and they are triggered by the similarity.

Jezebel in your case, you may have been prejudicially judged as 'just' seeking attention, accommodations as an advantage over your peers, a free ride, a 'fashionable' diagnosis, or in some other way this clinician may have been threatened by the extent of your knowledge compared to hers. The extensive notes you took to the encounter (I understand why) may have been totally interpreted as 'simply' anxiety driven.

Anyway the overall point is that there are so many of these prejudicial factors impacting on clinical encounters and we tend to overlook them all most of the time, professional supervision is supposed to keep biases out of the picture but in truth, either there is no supervision or it addresses only issues the supervised clinician wants to disclose and address (often the ones that show them in the best light for future performance reviews - yet another source of bias....

Despite all these minefields, there are some good clinicians/good evaluations/good diagnoses/good outcomes. I readily acknowledge that. The point of the thread is that over-reliance on clinical objectivity and infallibility is a really dangerous road and a new client can never know all the variables.

APOM I was intrigued by the 'intellectual crush' comment (do you have the original of it?) - I can relate to that in my student past!