The Man Who Couldn’t Stop (27 page)

DSM-III had 265 diagnosable disorders. DSM-IV followed in 1994, with 297 conditions across 886 pages. The new DSM-5 (the scrapping of Roman numerals was one of its less controversial changes) has about the same. In the DSM-5 obsessive-compulsive and related disorders have their own chapter, which gives the new grouping the same status as depression and schizophrenia. The DSM change implies that OCD and its related disorders have something in common. So, therefore, must the people who suffer from them. I have a new family.

On one level, it is irrelevant whether I suffer from an anxiety disorder or an obsessive-compulsive and related disorder. Labels to some extent are just that, labels. They can be peeled off and swapped around as fashion dictates. Some psychiatrists talk of ‘treatment nihilism': the diagnosis – the official label – is irrelevant. They have drugs and treatments and they can try each of them in different doses and combinations until something works. That's a direct progression from the end-justifies-the-means approach of the behaviourists, and it's as far as you can get from the probing, causal analysis of Freudian psychodynamics. In some patients it works. But to work, it needs drugs and treatments to be available. And to work with more people, it requires new and better treatments and drugs to emerge. And here's the problem: the drugs aren't coming.

Just as public and political awareness of psychiatric illness is rising and more people are being encouraged to acknowledge their own mental problems and seek help, the drug industry we rely on to offer some of that help is in full retreat. Mental illness is too difficult. Drugs against it are too expensive to develop, because so many fail. Just as the APA has extended the OCD family, the industry has given up on us. Since 2011, the major drug firms GlaxoSmithKline, AstraZeneca, Pfizer, Merck and Sanofi have all ended or scaled back their research to develop new drugs to treat brain disorders. It now takes more than thirteen years and more than a billion dollars to deliver a new psychiatric drug to market. Even nihilists have to pay the bills.

At present, it simply makes more financial sense for these firms to invest in areas of medicine where the chances of success and profit are higher. One reason for this retreat from mental illness is that drug companies and scientists are starting to realize that modern psychiatry is a castle built on sand.

*   *   *

The first two DSM books were based on Freudian psychodynamics, which dominated the field when they were written. Psychiatrists today often airbrush them away when they talk about the history of their discipline. The modern age of psychiatry, they say, began with the DSM-III in 1980.

That was when psychiatry got in touch with its scientific side and started to use evidence and empirical observations of patients to decide who was sick and what illness they had, rather than guesswork over childhood toilet habits. The DSM-III and the two versions that followed looked not to Freud but a different role model, one as far away from Freud as it is possible to imagine.

Emil Kraeplin was born the same year as Freud. They worked in the same field and they both spoke German, but the two men never met and never corresponded. They didn't agree on anything. Kraeplin had some unpleasant views on racial purity, but he was a true psychiatric scientist in the way that Freud was not. Kraeplin looked to biology for the causes of mental disorder, and he looked for evidence of how biology showed itself as common symptoms in his patients. He found plenty of both, and is most famous in psychiatric circles for using the results to identify and describe the syndromes we now call schizophrenia and bipolar disorder. Crucially, Kraeplin said these were separate conditions, with different causes and unique outcomes.

Kraeplin's work was overshadowed by Freud's and lay largely unnoticed in academic journals for decades, until it was rediscovered and given fresh momentum by scientists in the United States in the 1960s. This re-emergence of the need for empirical evidence to identify and group patients set the tone for the DSM-III – and all of modern psychiatry since.

It's called the category approach and it places firm boundaries between different mental disorders, each of which are viewed as discrete conditions to be diagnosed and treated separately. It uses fixed criteria and checklists that psychiatrists compare to the symptoms of a patient. Those criteria and checklists are modified as more or fewer patients seen in the real world seem to correspond to them. A big part of the DSM revision process is to gather the data on how well the various categories seem to fit those who report mental illness – how efficiently sick people can be placed in the categorical boxes.

Unfortunately for Kraeplin – and for the countless millions of patients around the world who still rely on his ideas for help with their mental illness – it now looks like he, just like his great rival Freud, was wrong. Unlike in Kraeplin's day, modern scientists have the tools to probe how accurately the various categories map onto genuine biological differences in patients. And as science has started to peek behind the curtain of psychiatry in this way, it has revealed the category approach to be a giant conjuring trick.

The most fundamental problem with the category approach is that people with a mixture of symptoms are believed to suffer from several different mental disorders at the same time. Psychiatrists call this co-morbidity. Patients call it confusing and demoralizing. At the most extreme, it's possible for a single person – who by definition can only have a single personality – to be diagnosed with three or four separate personality disorders simultaneously.

This problem arises because the underlying biology and pathology of most mental illnesses remain a complete mystery. No other branch of medicine relies so heavily on self-reported symptoms to make a diagnosis. No blood test can detect OCD. No X-ray can find depression. No biopsy can label someone as anxious or bipolar. In this important respect, the science of psychiatry has hardly moved on since its earliest days. It is still based around the simple question: So how do you feel today? Depression, for instance, can be diagnosed with a checklist. Tick five items and you are depressed, tick four and you are not. It really is as arbitrary as that.

To improve the situation, neuroscientists have long searched for biological differences in patients they can use to distinguish the diagnostic categories of the DSM. They look for gene mutations implicated in depression, or a part of the brain that is smaller or larger in OCD, or more and less active in schizophrenia. These so-called biomarkers would be the first step to reliable tests. But they remain elusive.

This could be because there are no such biological differences. Studies show that various mental illnesses, such as schizophrenia and bipolar disorder, have similar genetic signatures. They show that what the DSM describes as separate conditions seem, in fact, to have plenty in common. This has pushed psychiatrists to confront an awkward truth. The categories of the DSM probably don't reflect the situation on the ground. They may not exist at all. They do not, in a phrase they are fond of, carve nature at the joints.

This would be one reason why current treatments for mental illness fail so often. It's because people classified with depression, or OCD, or mania or whatever, and so grouped together by the listings of the DSM, are in fact different from each other, perhaps very different. And those differences mean that the same treatment won't help them in the same way.

That's something biologists are becoming more aware of as they learn about physical illness, where patients with the same disease also respond in radically different ways to identical treatment. A diagnosis of cancer on its own says nothing about a patient's condition, their day-to-day health and their survival chances. It's not even enough to stratify cancer by type – lung or breast for example – or even by the nature of the tumour. There are important differences between and within common types of tumours, and these differences can dictate whether a patient will, in many cases, live or die.

It sounds obvious. Of course two people with breast cancer or depression are different from each other. But without knowing why and how they are different, it's impossible to separate them, and equally impossible to link them to other people with a shared specific form of the illness. This separating and linking are essential to probe the underlying biology and so develop new drugs to target and treat them. It has taken decades of painstaking study of the way tumour cells live, die and interact, and a revolution in genetics, for oncologists to start to stratify their cancer patients. Psychiatrists can't do it as long as their solitary tool remains to ask people how they feel.

That's why the drive to establish links between mental illnesses and the reclassification of OCD as a disorder with something in common with other mental syndromes is important. Even though most psychiatrists acknowledge that the category approach is flawed, nobody is about to tear up the DSM; it's a product of politics as well as science. And existing categories do serve some useful purposes, such as helping to allocate treatment. They are likely to continue to exist for some time. A flawed but familiar system is better than nothing at all, right?

Not for research scientists, the people who patients and psychiatrists and drug companies all rely on to make the discoveries that will lead to better treatments. The category approach ties one hand behind their back. Here's why: To get a drug approved, it has to make sick people better. In psychiatric medicine, those sick people are those who have a DSM-listed disorder. That means clinical trials must be done with those same DSM groups – this drug for schizophrenia, this one for bipolar. And so, for basic research to stand any chance of being converted to useful medicines, it must analyse the problems of those groups as well. Research with brain scans and DNA scans recruits patients with ‘depression' or ‘obsessive-compulsive disorder' as if they are two separate species. It's not just the DSM that is tied to the flawed category approach to mental illness, it's the entire scientific and medical system. Panels that award research grants and promotions to scientists, and journal editors who decide which research to publish, all tacitly support and follow it. They endorse it, even though many of them know it doesn't work and isn't based on science.

Given these shortcomings of the existing system, and the unwillingness to change, it does at least make sense to group together those DSM categories that seem to have the most in common. The conceptual shift in recent years to view autism as a spectrum of related conditions rather than an isolated problem has significantly changed the way it is considered by scientists, patients and the public alike. It may not have brought a revolution in treatment, but at least we understand those affected better now, and can help them to understand themselves. The creation of the OCD spectrum could do something similar.

In practical terms, the DSM shift to obsessive-compulsive and related disorders means that research scientists now have official permission to ask for money to work on those conditions as if they are connected. In their studies they can group together people with OCD and compulsive hair-pulling, for example, and look for things they have in common. It's a small step, but it does go some way towards breaking Kraeplin's stranglehold on psychiatry, just as Freud's was broken before.

*   *   *

Some neuroscientists argue we will never understand the brain and the mind and how they relate in mental illness until we unravel how the 90-odd billion neurons in the brain connect to each other, and how those connections form, break and rearrange over time.

There are plans to draw up such a brain map – called the functional connectome. The connectome project assumes that it is not the performance of individual portions and regional segments of the brain that drives behaviour and motivation, but the careful choreography of many. That could explain why so many MRI studies of OCD brains give results that are all over the place. The brain is an orchestra, and to focus at any one time on the violins or the percussion section alone misses the point. The connectome conducts and the interplay of different regions of the brain, acting at the same time, delivers the symphony.

We see hints of this in the apparent importance of the circuit between the orbitofrontal cortex and the thalamus thought to play a role in OCD. That loop probably does not operate in isolation. More likely it draws input and variance from surrounding brain cells, and whatever affects them.

All this complexity means that, in the long term, the only way to develop better treatments for mental illness could be to scrap the category approach and go back to the drawing board. A different strategy would be to assume that the causes of mental illness are not divided according to the categories of the DSM, but are spread much more evenly through the population. The symptoms that emerge do so only when the causative agents – genes, environment, misfiring brain circuitry, whatever – reach a certain threshold. Diagnosis then becomes a matter of tuning in to the correct signal, rather than blindly punching clunky buttons until a crude approximation of a picture emerges.

This way of thinking about mental illness is called dimensionality. Blood pressure is dimensional, and so is weight. Both operate on a continuum and are best represented by a number. At some point, doctors class those numbers as too high, and they treat the patient for obesity or hypertension. Could something similar be done with mental health? The US National Institute of Mental Health thinks it could. It has launched a new project that aims to find biological traits in mental illness that, just like blood pressure, can be measured in everybody, and which can be classed as too high and then treated.

Anhedonia, for example, is a trait found in many mental illnesses, depression and schizophrenia among them. It's the inability to take pleasure from activities that should be fun – exercise, hobbies, sex or just hanging out with friends. It's a good way to describe the sense in OCD that life is lived on autopilot. Anhedonia has been linked to malfunctions in the normal reward circuits and systems of the brain. In theory, as neuroscientists develop more powerful brain scans and better genetic tests, these malfunctions could be identified, measured and quantified. Anhedonia could be tested for, and it could be treated. People could get to feel better, and all without a diagnosis of depression or schizophrenia or both.