A Conservative Approach to Diagnosis Grandfathers in Weak Links

A DSM critic, Andrew Hinderliter sent this perceptive email questioning the wisdom of the most fundamental decision we made in preparing DSM IV-- ie, our goal of keeping the system stable.

"A problem with your conservative approach to psychiatric nosology is that it grandfathers in the good and the bad alike and provides no way of changing really bad parts of the system. Unquestionably, there are things in DSM that couldn't pass a risk-benefit analysis (the Paraphilia section is one example). In a previous email to me, you gave the following explanation for its continued presence in DSM-IV: "I think they are there only because of history and inertia, but these are powerful forces.' Quite possibly, the single most powerful force of diagnostic inertia in psychiatric nosology in the past half century was the strongly conservative approach to diagnostic change that you yourself chose to implement in DSM-IV. A conservative approach to diagnostic change has much to be said for it, but in all the arguments back and forth about it, I feel like there is an elephant in the room regarding the much-harder-to-justify diagnoses. I have no doubt that you've thought about this before, but I was just curious as to your thoughts on the matter."

Indeed I have thought and blogged about this before. My fullest previous answer to the question can be accessed at http://alien.dowling.edu/~cperring/aapp/bulletin.htm (see particularly commentaries by Pierre, Piasecki, Kinghorn, Waterman, Cerullo, and Porter, and Ghaemi --and my responses to them). But Andrew Hinderliter has raised the issue most pointedly and is quite right in identifying this important paradox and weakness in my arguments for a conservative approach to psychiatric diagnosis.

To keep the diagnostic system from expanding wildly, we established extremely high thresholds for change in DSM-IV. Substantial scientific evidence was required for changes in either direction-- those that would add to the reach of the system, but also those would subtract from it. The rationale was that without clear and high scientific thresholds, changes would be arbitrary, destabilizing, and subject to personal whim.

This requirement did indeed permit the grandfathering of decisions made previously that would not have met the standards for new suggestions. Many of the conditions in DSM-IV have been subjected to too little research study to justify sunsetting, given our high threshold for change in either direction.

We could have taken a different approach-- denying tenure to the already included DSM-IIIR conditions unless they could prove themselves using the new, more stringent, rules. The advantages of having a slim but solid diagnostic system had to be balanced against the unknowable risks and inconveniences of making such a radical break with longstanding diagnostic traditions.

You could argue the case either way. We decided to play it safe and conservative. This left some problematic conditions in DSM IV--especially the Paraphilia section. But disrupting the system by drastic trimming probably would have caused other disturbing consequences. Our attitude was when in doubt, stand pat and do least harm. However, a plausible case can be made that the system needed more pruning and that we were too timid.  

Andrew Hinderliter has carried our dialog one step further.

He writes:

"In his arguments for diagnostic conservatism, I have largely found Dr Frances' case compelling given the historical context of DSM-IV: DSM-III was a watershed for psychiatric diagnosis, but people felt that DSM-III-R was too much change too fast. There was a need for stabilization, and DSM-IV provided this. In light of the very uncertain scientific foundation of psychiatric nosology, this stability helps make data sets easier to compare and promotes communication among clinicians. Where there is great potential of severe unintended consequences, caution is prudent.However, Dr Frances' response to my email is primarily an apology for past decisions. Whether wise or foolish, what is done is done, and what must now be addressed is what to do now.

The DSM-IV approach leaves a major problem unaddressed (and almost unaddressable). Psychiatric diagnosis has the most potential of getting into trouble when it pathologizes social deviance and normality. Dr. Frances' many critiques of problematic DSM-5 proposals deal precisely with these issues, but there is no reason to assume a priori that none of the diagnoses already in the DSM suffer from the same problems.

Yet the DSM-IV conservative approach leaves no mechanism for rectifying situations where inappropriate medicalization has already been built into the system. The question is how to preserve the benefits of diagnostic conservatism while creating a mechanism for dealing with this problem? In terms of avoiding the pathologization of social deviance, certain diagnoses in the DSM are obvious candidates for potential removal from the DSM (or at least need to be taken back to the drawing board). These would include: conduct disorder, antisocial personality disorder, the paraphilias, some of the impulse control disorders, gender identity disorder, and maybe some of the substance use disorders.

Workgroups could be formed to study the specific diagnoses felt to be most deserving of further scrutiny. Their membership should include people from a wide variety of backgrounds (eg, the law, ethics, psychiatry history, philosophy, sociology, etc.) and not just psychiatric experts in the relevant subfield.

The next question is what standards should be used. This gets us into the much debated question of how to decide whether a condition is a mental disorder, which I shall make no pretension to answer in the space available here. Perhaps the biggest problem to be overcome is disentangling mental disorder from (quickly changing) concepts of what is deviant or immoral. Psychiatry has a long history of pathologizing variant sexualities, including ones that are later no longer considered to be immoral or an indication of illness. In the US, homosexuality is the only variant sexuality to be successfully removed from the DSM. While science did play some role in this decision, as did psychiatrists simply feeling it was the right thing to do, it would have never happened without outside pressure. The anachronistic quality of the Paraphilia section illustrates the weakness of diagnostic conservatism: standards of morality and tolerance change faster than has the diagnostic system (at least in the US). In recent years, several of the Scandinavian countries have removed sadomasochism, transvestism, and fetishism from their versions of the ICD. The increasingly tolerant sexual ethics in the US suggests the need for a similar depathologization of sexuality or else DSM-5 will seem quaint and out of date"

My response:

Mr Hinderliter and I have no conceptual disagreement. We do, however, weigh differently the risks of diagnostic stability versus vigorous pruning. He is most worried about the harmful consequences of the weak categories that have been grandfathered in to the DSM system. I share his concerns, but worry more than he does that we have no well established and widely agreed upon grounds to support their deletion. Any attempt at pruning will likely result in its own set of unintended consequences.

One obstacle to deleting diagnoses results from the conflict that sometimes occurs between clinical utility vs validity. When we say a diagnosis has been included in DSM because of "historical tradition," this means it has been included because it has been a focus of clinical attention in the past. The DSM is first and foremost a tool for clinicians to assist them in their clinical practice. There have been very few diagnoses removed from the DSM over the years because removing a category that is being used by clinicians is disruptive. The value in clinical and research work has to be weighed against the harm done in other settings. As one example, the misuse of the DSM IV paraphilia section in forensic determinations may outweigh its current clinical value-- but deletion would be an extremely controversial call with no clear parameters on which to base judgments.

Mr Hinderliter and I agree that there is not likely to be much, if any, self correction from within psychiatry. The DSM system has a strong tendency to expand, little impetus to contract; it has a strong sensitivity to clinical need, less than adequate concern for forensic risk. Experts working on a particular diagnosis are never likely to recommend it be eliminated from the diagnostic system. If diagnoses are to be sunsetted, it would have to be done by a broader diagnostic "supreme court" responding to the fact that these questions are not just clinical or scientific or psychiatric, but also touch closely on law, ethics, morality, and public policy.

Again, for those interested in more on this see: http://alien.dowling.edu/~cperring/aapp/bulletin.htm (particularly commentaries by Pierre, Piasecki, Kinghorn, Waterman, Cerullo, and Porter, and Ghaemi--and my responses to them).