100

6
M
ental oddities and the pills t hat c
au

101

no matter what effects antidepressants unleash, these effects 
do not ‘cure’ us or return us to ‘normality’. Instead, if we 
are to make any sense of their effects at all, we must regard 
them as we do other mind-altering substances – as propel-ling us into an abnormal state of mind.

101

one of the most 
powerful myths embraced by the psychiatric establishment: M ental oddities and the pills that c ause them 9
5that psychiatric drugs are capable of ‘curing’ us and are 
therefore distinct from recreational drugs that merely 
alter our state of mind.

103

103

There was also something else that the pills had 
changed. Before taking them, in those early weeks after his 
wife’s death, he would regularly experience, almost daily, a 
dull ache growing in his chest. This would build and build, M ental oddities and the pills that c ause them 9
7sometimes building all day, until the pressure, unable to be 
contained any more, would finally burst out through a del-uge of tears. But, again, since taking his pills this internal 
pressure was somehow acting differently. Instead of finding 
an outlet through his tears, he said it almost now seemed 
to get stuck in his head, agitating his thoughts, and making 
him act in uncharacteristic ways. One of these unchar-acteristic behaviours involved religiously counting the 
number of lamp-posts he passed as he walked down the 
street. Another involved endlessly tracing the contours of 
distant objects with his index finger. For Tony these acts 
had reached obsessive levels, but they didn’t end there. He’d 
also become addicted to online card games, which he now 
played deep into the night. He’d never played these games 
before, let alone stayed up late on his computer. He had 
always preferred other activities in the late evening, like 
reading. But he rarely read nowadays, as serious concentra-tion was almost impossible. As he spoke to my colleague he 
wondered openly whether this pent-up pressure was gen-erating these ‘obsessive activities’. Of course he couldn’t be 
sure, but he was worried that the pills had somehow cut off 
his emotions and forced him into his head.

104

During the initial assessment my colleague dutifully 
noted all of Toby’s insightful reflections. That was his task. 
That’s how assessments work. What his task didn’t involve 
was exploring if the pills were responsible for the changes in 
Toby’s behaviour. Assessments do not afford the opportun-ity for that kind of enquiry. After all, if you choose to open 
Pandora’s box, you’d better do it in therapy, where you have 
time to manage what may fly out, rather than in a time-restricted assessment, where you don’t.

Note: neglect of the body. An attempt to find happiness changing your thinking without taking into account the body

105

105

One of the first questions Toby might have asked was 
why his emotions had become so flat. Were the pills respon-sible for cutting his emotions dead?
While it’s difficult to be 100 per cent certain what was 
responsible (because these changes might have occurred 
even if Toby hadn’t taken the pills), we do know that anti-depressants have effects. Mostly they have placebo effects 
and side-effects. We also know that for many people they 
can have sedating or numbing effects – perhaps the very 
effects that interfered with Toby’s grieving. For example, it’s 
clear that antidepressants work differently to many drugs 
used in other branches of medicine. Take antibiotics and 
antiviral drugs, for example. When these enter the body 
they attack the viruses or bacteria at the root of your illness. 
In this sense they cure us by literally killing the problem. 
The same can be said for chemotherapy, which is sent into 
the blood to destroy the cancer cells. But antidepressants 
don’t work in that way, because when we’re talking about 
emotional problems there’s rarely an underlying pathol-ogy, virus or disease to be cured (as I’ll discuss in the next 
chapter). Rather, these drugs, when they do have effects, 
work more like substances that temporarily alter our state 
of mind, such as caffeine or cannabis. These pills, in other M ental oddities and the pills that c ause them 9
9words, don’t cure us – they simply change us. They can 
throw us temporarily into a foreign state of mind, into an 
altered version of who we are.

106

From this standpoint, antidepressant medications do 
not return us to health as medical pills aim to do – they 
rather manufacture a new state of mind, and often an 
unnatural state. Was it natural, in other words, for Toby 
to suddenly stop crying after a terrible loss? After all, the 
natural course of grieving is a long one, drifting away only 
slowly and intermittently – it’s rare for it to just stop dead 
right in its tracks, as it did with Toby. Now, of course, it 
could be argued that Toby suddenly stopped grieving 
because his pills simply ‘cured’ him of an excessive emo-tional response. But that would be a dubious argument, not 
least because it assumes that grief is something we need 
to be ‘cured’ from.

Note: well it is an unpleasant feeling. But it should be up to the patient to decide wether or not they want to free them selves from that feeling, not a moralizing external authority who deems that behavior wrong

106

Rather, it makes far more sense to say 
that in Toby’s situation the pills simply numbed his natural 
reaction to a deeply painful event. This interpretation is 
consistent with research that clearly shows that if you give 
Prozac to a group of healthy individuals, after a while about 
half of them will experience emotional blunting.2

309

2 . Recent research has shown that the most commonly reported drug-induced psychoactive effects of the antidepressants Fluoxetine and 
Venlafaxine were sedation, impaired cognition, reduced libido, emo-tional blunting, activation (feelings of arousal, insomnia and agitation) 
and emotional instability. See: Goldsmith, J. and Moncrieff, J. (2011), 
‘The Psychoactive Effects of Antidepressants and their Association with 
Suicidality’. Current Drug Safety, 6(2): 115–21. Also see: Healy, D. (2006), 
Let Them Eat Prozac. New York: New York University Press (Chapter 7).

106

So the 
question is, what’s happening to these healthy people? Well, 
the pills certainly aren’t ‘curing’ them, because these people 
are healthy and have nothing to be cured of. Rather, these 
people have been thrown into a new state of mind manufac-tured by the drugs they’ve taken. This is to say, these drugs 
can sometimes produce effects beyond side-effects and pla-cebo effects, but often not the effects that drug companies 
advertise – not healing, improving or curative effects, as 
they say, but mind-numbing effects.3

309

3 . Simon Sobo’s article makes this point – that drugs don’t heal us but alter 
us. See: simonsobo.com/a-reevaluation-of-the-relationship-between-psychiatric-diagnosis-and-chemical-imbalances

107

1 0

107

numb-ing things isn’t curing things, or even, in the long run, 
helping things. It’s just providing a temporary and super-ficial distraction, and one that may store up problems later 
along the line.

107

although the alcohol has had an effect, it hasn’t uprooted 
the reason you felt insecure in the first place. It’s merely 
altered your state of mind so that you no longer experience 
your insecurity.

108

1
0

108

Perhaps his 
depression was an alarm that signalled he needed to change 
his life. But rather than listening to the alarm, he just turned 
it off with Prozac,

109

1 0

109

the vision of 
antidepressants I have just articulated is not the vision 
embraced by mainstream psychiatry.

109

it flatly contradicts their mainstream view, which is broadly 
consistent with that of the pharmaceutical industry.

109

to 
understand that view, and how it differs from the one above, 
let’s talk for a moment to Dr Joanna Moncrieff, a psychia-trist and senior researcher in the Department of Mental 
Health Science at University College London.

109

Moncrieff is today considered one of the most excit-ing figures in antidepressant research, earning a reputation 
almost as notorious as Irving Kirsch’s for challenging the 
status quo. As I sat with Moncrieff in her university office, 
she was more than eager to explain to me the difference 
between the vision of drugs largely embraced by the psychi-atric establishment (called the ‘disease-centred model’), and 
the vision outlined above (what she calls the ‘drug-centred 
model’). She traced their differences in the following way:

110

1
0

110

‘In the disease-centred model, people are assumed to 
have a mental disease, a problem in their brain. And drugs 
are thought to be effective because they rectify or reverse 
that underlying brain problem in some way. This is the 
dominant model in psychiatry, and the one that best serves 
psychiatric interests. But the drug-centred model offers an 
entirely different way of understanding how these drugs 
work. It rather emphasises that drugs are drugs; they are 
chemical substances that are foreign to the human body but 
which affect the way people think and feel. They have psy-choactive properties, just like recreational drugs do, which 
alter the way the body functions at a physiological level. So 
the drug-centred model does not say that psychiatric drugs 
heal brain problems, like the disease model claims; it rather 
says they alter people’s states of mind in ways that may or 
may not be helpful.’

110

The interesting thing about the drug-centred model is 
that while there are growing numbers of psychiatric nurses, 
psychotherapists and clinical psychologists who accept this 
version of how psychiatric drugs work, at the heart of the 
psychiatric establishment there is still considerable resist-ance to this view: ‘There are many psychiatrists who find 
the drug-centred model extremely challenging and sim-ply don’t want to hear it’, said Moncrieff candidly. ‘This is 
because it fundamentally undermines the notions upon 
which modern psychiatry is based: the idea that mental 
disorders are brain-based diseases which psychiatric drugs 
can remedy in the same way medical drugs remedy physical 
problems.’

110

Moncrieff ’s own academic research has shown that 
this current and widespread resistance to the drug-centred 1 04 C
racked: Why psychiatry is doing more harm than goodmodel has not always been so strong. ‘Prior to the 1950s, 
psychiatric drugs weren’t understood to act upon underly-ing diseases like they are today. They were seen as drugs that 
would pep you up. They were accepted as sticking plasters 
or uplifters that might at best be able to suppress symptoms 
for a period, but never were they seen as reversing a disease 
state.’
But this all changed, Moncrieff explained to me, when 
in the 1950s the drug-centred model began to be discred-ited. ‘This happened because the pills being used at that 
time [tranquillisers like Valium and Librium] were dis-covered to have terrible withdrawal effects. It also became 
obvious that these pills had been doled out to millions of 
people who were unhappy with their lives, particularly 
women who were trapped in miserable marriages. So once 
people started to realise that these pills had been used to 
suppress appropriate emotional responses to unhappy situ-ations, that whole drug-centred model of taking pep pills to 
pick you up fell into disrepute.’

111

Moncrieff therefore sees the rise of the disease-centred 
model as filling a vacuum left by the demise of the drug-centred model. ‘With the growing popularity of the older 
antidepressants during the 1960s, 70s and 80s, the diseased-centred model began to take over’, continued Moncrieff. 
‘This was especially the case in the 1990s when the new 
SSRI [selective serotonin reuptake inhibitor] antidepres-sants came on the scene. The drug companies were trying to 
capture that huge market of people who once took tranquil-lisers. But because the old model of how drugs work had 
been tarnished, they needed a new model to reassert their 
value and necessity. So now these drugs were cast as curing M ental oddities and the pills that c ause them 1
05us rather than changing us. And that’s where the idea of 
the chemical imbalance came in – it was perfect, because it 
implied that these drugs actually corrected a defect in the 
brain. If you have a brain disorder, a chemical imbalance, 
and this pill is going to correct that imbalance, then obvi-ously you must take it. Few questioned whether this new 
way of thinking totally obscured what these drugs really 
do. And this unthinking acceptance of the disease-centred 
view has dominated mainstream psychiatry for the last 20 
or 30 years.’

112

‘I actually think our work is complementary, not con-tradictory’, Moncrieff answered decisively. ‘I agree with 
Kirsch that the majority of benefit from these pills is due 
to the placebo effect. But also remember that Kirsch’s 
research shows that active drugs can sometimes work frac-tionally better than placebos, especially with people who 
are severely depressed. And that’s where the drug-centred 
model comes in. It can explain this small difference in terms 1 06 C
racked: Why psychiatry is doing more harm than goodof the psychoactive effects real drugs have. Either they 
produce uncomfortable side-effects that convince people 
they’re on the real drug (which in turn makes the placebo 
effect greater). Or some of these drugs are so sedative that 
they put people into such a fog that they can no longer feel 
depressed or anything else.’

113

A recent study published in the British Journal of 
Psychiatry by a team of researchers from Oxford University 
strongly confirms the view that most psychoactive effects 
are neither useful nor beneficial.4
To reach this conclusion, 
the researchers assessed 38 patients who had taken SSRI 
antidepressants for periods between three and 48 months 
(the average length being 23 months). The researchers then 
undertook in-depth interviews with all of the patients to 
find out how the pills affected them. The results they uncov-ered are so at odds with the modern myth of the ‘happy pill’ 
that it’s worth giving them almost in full:

113

Most participants described a general reduction in the 
intensity of all the emotions that they experienced, using 
words like ‘dulled’, ‘numbed’, ‘flattened’ or completely 
‘blocked’, to capture how they felt.

113

A few participants described feeling no emotions at all, 
while others reported their emotional experience had 
become more ‘cognitive’ or ‘intellectual’.

113

A few participants described how the emotions that were 
at times present seemed ‘unreal’, ‘fake’ or ‘artificial’. And 
almost all participants, paradoxically, described a reduc-tion in their positive emotions, including a reduction in 
emotions like happiness, enjoyment, excitement, antici-pation, passion, love, affection and enthusiasm.

114

1
0

114

Most participants also described feeling emotionally 
detached or disconnected from their surroundings. 
Most also described this detachment as extending to a 
detachment from other people. Specifically, they felt 
reduced sympathy and empathy, and felt detached dur-ing social interactions. Many participants also described 
an emotional detachment from their friends and family, 
including their partner or children.

114

Almost all participants described not caring about things 
that used to matter to them. They cared less about them-selves, about other people and about the consequences 
of their actions. Not caring could have both helpful and 
unhelpful consequences: it could reduce the sense of 
pressure and stress, but it could also increase the likeli-hood that important tasks were neglected.

114

Many participants described a general feeling of indiffer-ence to things in life that used to matter to them. Some 
felt they just did not care as much about the consequences 
to themselves of their behaviour. A few participants went 
further, mentioning thoughts of self-harm or suicide that 
they related to their emotional detachment and numb-ness. One participant had started to self-harm in an effort 
to feel emotion. Many participants reported not caring 
as much about others, being less sensitive or courteous 
towards other people, having reduced concern for others’ 
feelings, and reduced concern about other people’s opin-ions of them. Some participants described being less 
concerned or even unable to care about responsibilities 
in their everyday lives.

114

All participants experienced a reduction of intensity or 
frequency of negative emotions. Most considered that 
at some stage the reduction in negative emotions was 1 08 C
racked: Why psychiatry is doing more harm than goodbeneficial to them, bringing relief from distressing nega-tive emotions like emotional distress, anger, irritability 
or aggression; and anxiety, worry or fear. Although this 
reduction was usually at some stage a relief, many par-ticipants also reported it impaired their quality of life. 
Participants described the need to be able to feel negative 
emotions when appropriate, such as grief or concern. 
Some were unable to respond with negative emotions, 
such as being unable to cry when this would have been 
appropriate, or respond appropriately to bad news.

115

Some participants felt their personality had changed 
in some way. They felt they were not the person that 
they used to be. Participants also reported that specific 
aspects of their personality, and, in particular, emotional 
aspects, had been changed or lost. Some participants 
believed that at times their antidepressant had made 
them behave out of character. One participant believed 
that the medication had changed their personality per-manently, having a lasting effect beyond finishing their 
medication.

28

2
T
he DSM – a great work o f f

29

29

‘So presumably’, I asked, ‘these disorders had been 
discovered in a biological sense? That’s why they were 
included, right?’
‘No – not at all’, Spitzer said matter-of-factly. ‘There 
are only a handful of mental disorders in the DSM known 
to have a clear biological cause. These are known as the 
organic disorders [things like epilepsy, Alzheimer’s and 
Huntington’s disease]. These are few and far between.’
‘So, let me get this clear’, I pressed, ‘there are no dis-covered biological causes for many of the remaining mental 
disorders in the DSM?’
‘It’s not for many, it’s for any! No biological markers 
have been identified.’

30

2

30

‘Well, it’s important to hear you say this, because this is 
something most people simply don’t know. I didn’t know it 
when I started out training as a psychotherapist. Most of my 
patients don’t know it today. And I suspect for many people 
reading this interview it will come as a surprise too.* ‘So if 
there are no known biological causes’, I continued, ‘on what 
grounds do mental disorders make it into the DSM? What 
other evidence supports their inclusion?’

30

‘Well, psychiatry is unable to depend on biological 
markers to justify including disorders in the DSM. So we 
look for other things – behavioural, psychological; we have 
other procedures.’

30

Before we look at these other procedures, let me explain 
why you are probably surprised to hear that biological 
research did not guide the DSM’s expansion. This may 
sound strange to you because we all expect psychiatry to 
work much like the rest of modern, mainstream medicine. 
In mainstream medicine a name will be given to a disease 
only after its pathological roots have been identified in the 
body. With few exceptions that is how general medicine 
operates: once you have discovered the physical origins of 
a problem, you then give it a name such as cystic fibrosis, 
cancer or Crohn’s disease. But the surprising truth about 
psychiatry is that it largely operates in completely the oppo-site way. Rather, psychiatry first names a so-called mental 
disorder before it has identified any pathological basis in 
the body. So even when there’s no biological evidence that a 
mental disorder exists, that disorder can still enter the DSM 
and become part of our medical culture.

31

31

Of course, the fact that psychiatry operates differently 
does not mean that its procedures are necessarily wrong. 
The only way to decide this is to assess whether psychia-try’s alternative methods are scientifically valid. To find 
out if this is the case, I asked Spitzer to take me through 
the procedures his Taskforce followed when deciding 
whether to include a new disorder. For example, if the 
findings of biology didn’t help the Taskforce to determine 
what disorders to include in DSM-III, then what on earth 
d

31

‘I guess our general principle’, answered Spitzer can-didly, ‘was that if a large enough number of clinicians felt 
that a diagnostic concept was important in their work, then 
we were likely to add it as a new category. That was essen-tially it. It became a question of how much consensus there 
was to recognise and include a particular disorder.’
‘So it was agreement that determined what went into 
the DSM’?
‘That was essentially how it went – right.’

31

What sprung to mind at Spitzer’s revelation was the 
point I made in the previous chapter about agreement not 
constituting proof. If a group of respected theologians all 
agree that God exists, this does not prove that God exists. 
All it proves is that these theologians believe he does. So 
in what sense is psychiatric agreement different? Why, 
when a committee of psychiatrists agree that a collection of 
behaviours and feelings point to the existence of a mental 
disorder, should the rest of us accept they’ve got it right? 
Perhaps, in the absence of biological evidence to convince 
us, they can produce other kinds of evidence to assure us 
that their agreements were justified? In other words, what T he DSM – a great work of fi ction? 2
5was the evidence leading the Taskforce to agree that a new 
disorder should be included in the DSM?

32

Before coming back to Spitzer for an answer, let me first put 
this question to the psychologist Professor Paula J. Caplan, 
currently a Fellow at Harvard University’s Kennedy School, 
and former consultant to two DSM committees. I inter-viewed Paula from my home in London in late April 2012, 
precisely because she has extensively assessed the evidence 
that guided many of the decisions the DSM Taskforce made.

32

One of the disorders she has focused on closely was 
called ‘masochistic personality disorder’. Spitzer’s Taskforce 
wanted to include this new disorder in the DSM for people 
who displayed ‘masochistic traits’ such as those thought to 
invite harsh treatment from others, or those leading people 
to seek out pain for enjoyment. Now, a crucial reason why 
Paula Caplan and other critics objected to these traits being 
called symptoms of a psychiatric disorder was because these 
traits were also said to be typical of women who were vic-tims of violence. So it was thought that this diagnosis was 
very dangerous, not only because it could be used in courts 
of law to suggest that female victims of violence were in fact 
bringing it upon themselves (because they had a ‘masochis-tic personality disorder’), but also because it could be used 
to let perpetrators of violence off the hook – they were sim-ply doing what these women supposedly wanted.

32

So after much opposition from Caplan and other 
psychologists, the committee finally decided to rename 
masochistic personality disorder as ‘self-defeating per-sonality disorder’ – or the neat SDPD. But the critics then 2 6 C
racked: Why psychiatry is doing more harm than goodargued that this change in name still implied that there was 
something ‘self-defeating’ in these victims – something 
compelling them in some way to invite abuse upon them-selves. ‘So the change in name was not really a victory at 
all’, said Caplan to me energetically, ‘since by renaming the 
disorder as SDPD nothing really had changed: the renamed 
disorder could still be used to claim that women victims of 
abuse, well, kind of asked for it.’

33

When Caplan made this point to Spitzer, he remained 
simply unmoved. In fact, his desire to keep SDPD was so 
strong, it would have been understandable had the critics 
retreated. But they didn’t. Rather, at the last hour, Caplan 
devised a simple plan: ‘I decided to scrutinise thoroughly 
the very research used to justify including SDPD in the 
DSM.’ And here’s what she found.

33

Firstly, she discovered only two pieces of research – a 
remarkably small number by anyone’s standards. But as 
surprising as this discovery was, when Caplan actually 
looked at the research she became incredulous. ‘It was so 
methodologically flawed’, said Caplan animatedly, ‘that it 
would fail an undergraduate examination. In fact, it was so 
full of basic errors that I actually decided to use it on an 
undergraduate exam in which I asked students to point out 
every conceivable methodological error, because his study 
had so many.’

33

For example, in Spitzer’s research a group of psychia-trists at only one university, who already accepted SDPD 
existed, were shown some old case studies. All then unani-mously agreed that the patients in them had SDPD. Caplan 
pointed out that just because some psychiatrists at one 
hospital diagnosed their patients with SDPD, that was not T he DSM – a great work of fi ction? 2
7proof that the disorder actually exists. As Caplan said: ‘all 
Spitzer’s research proves is that a group of psychiatrists 
working in the same institution gave the same label – rightly 
or wrongly – to a given set of behaviours.’1
It proves nothing 
more than that.

Note: it shouldnt be a a matter of wether it “exists” or not. Mental illness categories should be subjective adjectives. Not nouns. They are useful criteria that people can use to analyze their own behavior and potentially treat it

34

‘But if you think that first piece of research was weak’, 
continued Caplan, ‘then consider the second piece. This 
involved sending out a questionnaire to a selected num-ber of members of the American Psychiatric Association. 
This asked them whether the diagnosis SDPD should be 
included in the DSM. If they voted “yes” then they were 
asked to describe what they thought the characteristics 
of SDPD were. If they voted “no” then they were asked to 
return the questionnaire, blank, without any clinical data. 
This meant that the only data gathered about the character-istics of SDPD was data obtained from people who believed 
in the existence of SDPD in the first place.’

34

So how many psychiatrists believed SDPD to exist – 
how many voted ‘yes’?
An official report showed that only 11  per cent of 
those who returned the questionnaire described what they 
thought the characteristics of SDPD were.2
So essentially 
only 11  per cent voted ‘yes’, which is surely not a repre-sentative sample of the psychiatric community. But what 
made matters worse is that the questionnaire was also sent 
to many psychiatrists who already supported the diagno-sis and who were deliberately screened into the study. And 
these psychiatrists, we can assume, made up a proportion 
of this 11 per cent.3

34

Caplan has therefore convincingly argued that neither 
piece of research justifies creating a new mental disorder. But 2 8 C
racked: Why psychiatry is doing more harm than goodthat didn’t stop Spitzer, as she said, from proudly reporting 
in the DSM that the nature of SDPD was defined by exam-ining the ‘data’ from a single questionnaire; a questionnaire 
Spitzer claimed had been ‘distributed to several thousand 
members of the APA’. ‘Spitzer does not report the methodo-logical flaws in his research’, said Caplan incredulously, ‘and 
instead leads us to believe the creation of this disorder was 
based upon widespread scientific consultation and study.’

35

I have discussed at length the case of SDPD because it forces 
us to ask whether other disorders were included in the 
DSM on the basis of equally poor scientific evidence. Was 
this just an isolated example, or is it quite representative? 
To try to find out, I decided to read to Spitzer the following 
quotation, which claims that the research backing was not 
just poor for SDPD but for most of the mental disorders 
Spitzer’s team included. This verdict comes from one of the 
leading lights on Spitzer’s Taskforce, Dr Theodore Millon. 
Here’s what he said about the DSM’s construction:
There was very little systematic research, and much of the 
research that existed was really a hodgepodge – scattered, 
inconsistent, and ambiguous. I think the majority of us rec-ognized that the amount of good, solid science upon which 
we were making our decisions was pretty modest.4
Once I’d read this quote to Spitzer, I asked him whether he 
agreed with Millon’s statement. After a short and somewhat 
uncomfortable silence, Spitzer responded in a way I didn’t 
expect:T he DSM – a great work of fi ction? 2
9‘Well, it’s true that for many of the disorders that were 
added, there wasn’t a tremendous amount of research, and 
certainly there wasn’t research on the particular way that 
we defined these disorders. In the case of Millon’s quote, I 
think he is mainly referring to the personality disorders … 
But again, it is certainly true that the amount of research 
validating data on most psychiatric disorders is very limited 
indeed.’

306

4 . Quoted in: Angell, M. (2009), ‘Drug Companies & Doctors: A Story of 
Corruption’, The New York Review of Books, 15 January 2009.

36

Trying not to look shocked, I continued: ‘So you’re say-ing that there was little research not only supporting your 
inclusion of new disorders, but also supporting how these 
disorders should be defined?’
‘There are very few disorders whose definition was a 
result of specific research data’, responded Spitzer. ‘For bor-derline personality disorder there was some research that 
looked at different ways of defining the disorder. And we 
chose the definition that seemed to be the most valid. But 
for the other categories rarely could you say that there was 
research literature supporting the definition’s validity.’

36

Spitzer’s admission so surprised me that I decided to 
check it with other members of his Taskforce. So on a rainy 
English Monday I called Professor Donald Klein in his New 
York office to ask whether he agreed with Spitzer’s account 
of events. Klein had been a leader on the Taskforce, and so 
was at the heart of everything that went on.
‘Sure, we had very little in the way of data’, Klein con-firmed through a crackling phone line, ‘so we were forced to 
rely on clinical consensus, which, admittedly, is a very poor 
way to do things. But it was better than anything else we had.

36

‘So without data to guide you’, I nudged carefully, ‘how 
was this consensus reached?’3 0 C
racked: Why psychiatry is doing more harm than good‘We thrashed it out, basically. We had a three-hour 
argument. There would be about twelve people sitting 
down at the table, usually there was a chairperson and 
there was somebody taking notes. And at the end of each 
meeting there would be a distribution of events. And at the 
next meeting some would agree with the inclusion, and the 
others would continue arguing. If people were still divided, 
the matter would be eventually decided by a vote.’

39

What is striking about the construction of the DSM is that 
the procedures it followed often had very little to do with 
‘science’ as most people understand the term, because, in 
short, the evidence was lacking. The problem here is obvi-ous. When a group of scientists sit down to decide whether 
something is true, they consult the evidence. If the evidence 
points to a clear conclusion, then irrespective of whether T he DSM – a great work of fi ction? 3
3an individual scientist likes it, the result has to be accepted. 
That is how science works. The evidence is king. But when 
you don’t have evidence to decide the issue for you, people’s 
opinions, beliefs, hopes and prejudices begin to intrude. In 
this instance, the scientist who desires a particular conclu-sion suddenly speaks up, argues loudly, and may, through 
sheer force of character, have their preferences accepted. 
And when there is no evidence to guide me, it can easily 
become largely a matter of personal or professional prefer-ence whether I vote this way or that. Voting, in other words, 
is not a scientific activity. It is a cultural activity. People vote 
for class presidents, union leaders, political parties and a 
host of other things. And yes, sometimes their votes are vin-dicated, but often they are not. Votes can disappoint. This 
is because a vote is not a guarantee that the thing voted for 
is real or true or good or certain. Votes are at best informed 
guesses, and at worse punts in the dark, and so when any-thing is voted into existence, whether it be a new leader, a 
political policy, or indeed a new mental disorder, the likeli-hood that we have got it wrong is never far away.

41

41

most people didn’t know 
that the fundamental changes Spitzer brought to global psy-chiatry required only the consensus of an extremely small 
group of people. Indeed, as Spitzer openly confirmed to 
me in our interview: ‘Our team was certainly not typical of 
the psychiatry community, and that was one of the major 
arguments against DSM-III: it allowed a small group with a 
particular viewpoint to take over psychiatry and change it 
in a fundamental way.’T he DSM – a great work of fi ction? 3
5‘What did you make of that criticism?’, I asked.
‘What did I think of that charge? Well, it was absolutely 
true! It was a revolution, that’s what it was. We took over 
because we had the power.’

43

43

the DSM committee did 
not actually discover mental disorders, at least not in any 
traditional scientific sense. Rather, they contrived them,

45

3
T

45

Month on month, year on year, increasing numbers 
of children like Dominic are being diagnosed with men-tal disorders like ADHD. In fact, diagnoses of ADHD 
have risen so sharply in the last ten years that 5.29  per 
cent of the global child population is now thought to suf-fer from the condition (with prevalence rates in North T he medicalisation of misery 3
9America and Europe pretty much equal at around 5  per 
cent).1
This vaulting rise in ADHD is consistent with 
a growth in other childhood psychiatric disorders. 
If we add up the prevalence rates for all childhood disor-ders, for example, it’s estimated that between 14 and 15 per 
cent of children now suffer from a diagnosable mental 
disorder in any given year.2
But high as these figures may 
be, they pale in comparison to those relating to the adult 
population. For example, the National Institute of Mental 
Health in the US now claims that about 26.2 per cent of all 
American adults suffer from at least one of the DSM disor-ders in a given year;3
while the Office for National Statistics 
on Psychiatric Morbidity in the UK reports a similar figure.4
Effectively this means that one in four people is afflicted by 
a mental disorder in a given year each side of the Atlantic; 
a figure made more startling when you consider that in the 
1950s it was more like one in a hundred, and at the begin-ning of the twentieth century a meagre one in a thousand. 
So what can account for this massive surge in mental dis-orders? Why in just a few decades have we apparently all 
become so psychiatrically unwell?

306

1 . Polanczyk, G. et al. (2007), ‘The worldwide prevalence of ADHD: a sys-tematic review and metaregression analysis’. The American Journal of 
Psychiatry,
164(6): 942–8.2 . Waddell, C. (2002), ‘Child Psychiatric Epidemiology and Canadian Pub-lic Policy-Making: the state of the science and the art of the possible’. 
Canadian Journal of Psychiatry, 47(9): 825–32. Also see: Merikangas, 
K.R. et al. (2010), ‘Prevalence and Treatment of Mental Disorders Among 
US Children in the 2001–2004 NHANES’. Pediatrics,
125(1): 75–81.3 . Kessler R.C., (2005), ‘Prevalence, severity, and comorbidity of twelve-month DSM-IV disorders in the National Comorbidity Survey 
Replication (NCS-R)’. Archives of General Psychiatry,

46

There are at least three hypotheses that mental health 
practitioners use to try to account for the escalating rates.

46

The first goes like this: as the pressures of contempo-rary life have increased, so too have our levels of stress and 
strain, leading to an upsurge in poor mental health. Now 
while this explanation seems reasonable enough, as we will 
see later, it’s difficult to ascertain whether contemporary life 4 0 C
racked: Why psychiatry is doing more harm than goodis really so much more stressful than life many decades ago. 
Indeed, as many sociological studies have shown, social 
stress may have actually decreased rather than increased 
in recent years, therefore putting this hypothesis under 
strain.*

47

The second hypothesis is also problematic: it says that 
mental disorders have increased because today’s psychia-trists are better than in the past at recognising psychiatric 
disease. Perhaps advances in technology now allow clini-cians to more readily spot and diagnose disorders that 
once slipped below their radar. While this hypothesis again 
has some obvious appeal, its weakness is that by and large 
diagnostic technology has not improved: there are still no 
objective tests that can confirm the validity of any psy-chiatric diagnosis, a fact supported by the continued low 
diagnostic reliability rates.

47

third possibility: that psychia-try, by progressively lowering the bar for what counts as 
mental disorder, has recast many natural responses to the 
problems of living as mental disorders requiring psychiatric 
treatment? In other words, has psychiatry, by redrawing the 
line between disorder and normality, actually created the
illusion of a pandemic?

47

  • See Chapter 12 on recent well-being rates in Britain.

49

49

medicalisation – namely, the process by which 
more and more of our human characteristics are seen as T he medicalisation of misery 4
3needing medical explanation and treatment. Now, while in 
the Canadian study it’s clear that the effects of medicalisa-tion can be deleterious, this is obviously not the case in all 
instances – indeed medicalisation, at best, has often been a 
force for good. For example, it was right to use medicine to 
tackle biological conditions such as Huntington’s disease or 
epilepsy that were once unhelpfully understood as religious 
problems (to be healed only by prayer or church attend-ance). And yet there are forms of medicalisation that are 
clearly unhelpful, that invasively spread medical authority 
where it was never designed to go. For instance, ‘prob-lems’ such as low achievement, certain kinds of truancy or 
under-performance have attracted medical diagnoses and 
intervention in our children, as have many normal reac-tions to the demands of adult life that are labelled as ‘stress 
disorders’ to be biologically explained and pharmacologi-cally treated.

50

The issue of medicalisation is so crucial because it 
concerns where the very limits of medical intervention 
should be drawn. At what point does medicalisation begin 
to undermine the health of a population? At what point 
does it begin to turn what should be a matter for spiritual, 
philosophical or political understanding and action into an 
issue to be managed by medicine alone? This question has 
particular relevance for psychiatry. For psychiatry, as we 
will soon see, has been accused more often than any other 
medical specialism of incorrectly medicalising our normal 
actions and responses. The question for us right now, then, 
is to what extent this accusation is true …

51

51

In an interview for a BBC documentary in 2007, the film’s 
maker, Adam Curtis, posed this very question to Robert 
Spitzer. He asked Spitzer whether the DSM had committed 
any errors. More precisely, he asked whether when creating 
DSM-III his Taskforce had adequately distinguished between 
human experiences that were disordered, and human 
experiences that were not. In effect, had the Taskforce, when 
creating their list of mental disorders, wrongly labelled many 
normal human feelings of sadness and anxiety as indicators 
of medical disorders that required treatment?

51

Spitzer, with noticeable regret, admitted that this had 
occurred. He then went on to explain why:
‘What happened is that we made estimates of preva-lence of mental disorders totally descriptively, without 
considering that many of these conditions might be nor-mal reactions which are not really disorders. And that’s 
the problem. Because we were not looking at the context 
in which those conditions developed.’ (In other words, 
Spitzer’s DSM only described the symptoms of each dis-order, but never asked whether these so-called symptoms 
could, in some circumstances, actually be normal human 
reactions to difficult life situations.)

51

An incredulous Curtis therefore said to Spitzer: ‘So you 
have effectively medicalised much ordinary human sadness, 
fear, ordinary experiences – you’ve medicalised them?’
‘I think we have to some extent’, responded Spitzer. 
‘How serious a problem it is, is not known. I don’t know 
if it is 20 per cent, 30 per cent. I don’t know. But that is a 
considerable amount if it is 20 or 30 per cent.’5

306

5 . For interview, see Adam Curtis’s BBC documentary Th e Century of the 
Self 

51

In this interview with Adam Curtis, Spitzer admitted 
that his Taskforce was interested only in the experiences that T he medicalisation of misery 4
5characterised the disorder. It was not interested in under-standing the individual patient’s life or why they suffered 
from these experiences. Because these contextual factors 
were overlooked, experiences of sadness, anxiety or unhap-piness were often listed as symptoms of underlying disorders, 
rather than seen as natural and normal human reactions to 
certain life conditions that needed to be changed.6

306

6 . This discussion is paraphrased from my book: The Importance of Suffer-ing (Davies, 2012)

52

You’ll remember that I met Spitzer in his house in leafy 
Princeton. Well, as we sat eating lunch, I took the chance 
to ask him about the Curtis interview. I recounted to him 
his exchange with Curtis, at which he slowly put down his 
spoon and turned his head in my direction. It was immedi-ately clear to me he was unwilling to elaborate on what he 
had previously said. It also seemed clear that he had shifted 
his position since that earlier interview with Curtis. While 
he still agreed that normal reactions were being recast as 
psychiatric illness, he now seemed keener to locate the 
cause of this problem elsewhere: not in how the DSM was 
constructed and written, as he had confessed to Curtis, but 
in how the manual is being used. As Spitzer explained:
‘[In clinical practice] there is often too much emphasis 
placed by some on the diagnostic criteria of the DSM [in 
other words, if a person has ‘this set of symptoms’ then they 
have ‘this disorder’]. This approach ignores other things 
that are important when making an assessment, such as the 
context in which the person became ill. So there has been a 
move towards an over-emphasis on diagnostic criteria, and 
a neglect of assessing the social context in which the person 
is living.’

52

In other words, problems emerge when a psychiatrist 
simply tries to match the patient’s experiences with one of 4 6 C
racked: Why psychiatry is doing more harm than goodthe disorders in the book, without investigating why the 
person is suffering as they are. After all, perhaps they are 
suffering because they have just lost their job, or someone 
dear to them, or because they’re struggling with their iden-tity, with poverty, with failure in love or work – who knows, 
perhaps their lives just hadn’t turned out as they’d hoped.

53

There are countless understandable reasons why a person 
may suddenly start manifesting emotions or behaviours 
that can be easily misread as ‘symptoms’ of ‘major depres-sion’ or ‘anxiety disorder’

Note: No! There isn’t a difference they exist in spectrum. There ain’t sadness and “depression sadness,” it’s just different degrees of sadness! And treatment would be useful for either case

56

‘Can you put a figure on how many people have been 
wrongly medicalised?’

Note: there is no “rightly” medicalized. People shouldn’t be separated into normal and abnormal categories. These symptoms exist on a spectrum and everyone has a degree of them

57

non-compliance with treatment (a diagnosis that can be 
given when the patient resists treatment); conduct dis-order (repetitive or persistent violation of societal norms 
or others’ rights)

57

oppositional defiant disorder (for 
children with irritable mood swings and who overly defy 
authority)

4

4 The depressing truth about happy pills 5
8 5 Dummy pills and the healing power of belief 7

119

7
B

120

1
1

120

They locked him in solitary confinement 
and subjected him to forced drugging.

124

the APA’s final statement conceded there 
are no “discernible pathological lesions or genetic abnor-malities” in mental disorders.3

131

1 2

131

the 
theory isn’t supported. There are three separate reasons for 
this that I heard again and again:

131

Number 1: the main evidence for the theory is arrived 
at deductively. Because antidepressant drugs like SSRIs act 
to increase our serotonin levels, and because antidepres-sants alleviate depression, depression is therefore probably 
caused by some kind of serotonin deficiency.

132

Furthermore, the largest recent meta-analysis of ser-otonin studies, conducted at the University of Amsterdam, 
has shown that low serotonin does not act as a depres-sant because when it was lowered in healthy individuals it 
produced no decrease in mood at all. As the authors sum-marised: ‘Although previously the monoamine systems 
were considered to be responsible for the development 
of MDD [major depressive disorder], the available evi-dence to date does not support a direct causal relationship 
with MDD.’8

309

6 . Coppen, A. (1967), ‘The Biochemistry of Affective Disorders’. British 
Journal of Psychiatry, 

138

1
3

138

Studies of rats have illustrated this point well. Baby 
rats born to mothers that displayed little affection (who 
licked their pups rarely) where given to foster mothers 
who were very affectionate (who licked them a lot). After 
dissection, it was clear that the affectionately-raised rats 
had brain characteristics different to those receiving little 
affection: affectionately-raised rats possessed more recep-tors in the neurons of their hippocampus, receptors that 
are considered crucial stepping-stones in slowing down the 
production of stress hormones. It turns out that a stretch of 
DNA, serving as a switch for a gene related to these neural 
receptors, had been suppressed in the less affectionately-raised rats. This meant that the rats receiving less affection 
(fewer licks) were far more stressed as adults than those 
who had received more. The conclusion is that adult per-sonality differences related to stress weren’t determined by 
genes inherited from their biological mothers, but were an 
outcome of how they were raised as pups.1

310

1 3. Kaffman A., Meaney, M.J. (2007), ‘Neurodevelopmental Sequelae of Post-natal Maternal Care in Rodents: clinical and research implications of 
molecular insights’. J Child Psychol Psychiatry, 48(3–4):224–44. A really 
good journalistic account of this study can be found here: Zimmer ,
C. (2010), ‘The Brain: The Switches That Can Turn Mental Illness On and 
Off ’, Discover Magazine, published online, 16 June 2010.

138

The same groups of researchers performed a related study 
of human beings. This involved analysing the brains of 36 
people post mortem. Twelve of these people had died of nat-ural causes, while the rest (24) had committed suicide. And 
of the 24 suicide victims, twelve of these had been abused as 
children, whereas the other twelve had not. When the brains 
of these three groups were compared, the group that had suf-fered childhood abuse again stuck out. People in this group 
shared the same pattern of fewer receptors linked to stress 
hormones as those found in the non-affectionately-raised 
rats. Their brains, via epigenetic changes, had reacted to the 
environmental abuse – leading them to grow in a direction 
different to brains receiving environmental care.14

310

1 4. McGowan, P.O. et al. (2009), ‘Epigenetic Regulation of the Glucocorti-coid Receptor in Human Brain Associates with Childhood Abuse’. Nat 
Neurosci,

139

many mental health professionals 
regrettably used simplistic genetic reductionism to avoid 
the complexity of thinking through the various social, eco-nomic and environmental factors influencing who we are, 
what we do, and who we become.

140

1
3

140

Most genetic influences 1 34 C
racked: Why psychiatry is doing more harm than goodon disease are greatly more complicated than those early 
pioneers of the genome project could have dreamed. For 
instance, in the realm of psychiatry there is no known gene 
or clear genetic variants for around 97 per cent of all the 
mental disorders contained in the current DSM and ICD. 
And even where genes may be implicated in disorders like 
bipolar disorder and schizophrenia, research now reveals 
such mind-boggling complexity that nothing definitive 
can be said. A study that scanned the genetic sequences 
of 20,000 normal people and then compared them with 
the sequence of 10,000 patients with schizophrenia, for 
example, revealed that over 10,000 different gene variants 
could have a role in the onset of schizophrenia. And this 
study did not take the findings of epigenetics into account 
(the environmentally susceptible molecules that interfere 
with these genetic variants).1

310

1 8. See: Joseph, J., Ratner, C. (no date), ‘The Fruitless Search for Genes in Psy-chiatry and Psychology: Time to Re-examine a Paradigm?’ Website: www.
councilforresponsiblegenetics.org/pageDocuments/1NX6VC0254.pdf 
(accessed September 2011)

141

Of course, there have been periods of great excitement. 
In 2003, a study was published in the journal Science that 
asked why stressful experiences lead to depression in some 
people but not in others. After analysing 847 patients over 
time, it was found that those who had one or two copies 
of a gene variant that interfered with serotonin transport 
were three times as likely to develop depression if subjected 
to certain stressful life events, like losing a job or getting 
divorced. This study was thought to provide evidence of a 
gene-by-environment interaction, in which an individual’s 
response to environmental stresses is moderated by his or 
her genetic make-up.19 This finding generated a great deal of 
enthusiasm until another study, published a few years later, 
tried to replicate these findings by assessing over 14,000 
people via a meta-analysis of fourteen studies. But the B io-babble? 1
35conclusion it reached dampened the previous excitement: 
‘This meta-analysis yielded no evidence that the serotonin 
transporter genotype alone or in interaction with stressful 
life events is associated with an elevated risk of depression 
in men alone, women alone, or in both sexes combined.’2

310

1 9. Caspi, A. et al. (2003), ‘Influence of Life Stress on Depression: Modera-tion by a Polymorphism in the 5-HTT’, Gene Science,
301(5631): 386–9.2 0. Quoted in Carlat, D. (2010), Unhinged: The Trouble with Psychiatry – a 
doctor’s revelations about a profession in crisis. London: Free Press (p. 80).

142

where genetics plays a role in our mental lives, it does so via 
a given and so far undefined constellation of genes that may
predispose a person to an unknown degree of vulnerability 
to developing a given form of mental distress if other social 
or psychological conditions trigger it, and if environmentally 
influenced epigenetic factors permit it.

142

Such tentativeness is 
now slowly trickling through to the mental health establish-ment, as can be seen from the World Health Organisation’s 
recent official statement on the causes of depression:
Depression is a complex disorder which can manifest itself 
under a variety of circumstances and due to a multiplicity of 
factors … biological (genetic and biochemical), sociological 
(stressors) and psychological (development and life experi-ences) factors interact to produce a picture of depression. 
Research during the last fifty years indicates that there is no 
single factor which can explain the cause for depression.
2
1 [italics added]

311

2 1. World Health Organisation, ‘Mental Health and Substance Abuse, 
Facts and Figures Conquering Depression’. Website: www.searo.who.
int/en/Section1174/Section1199/Section1567/Section1826_8101.htm 
(accessed August 2010)

192

1
0W
hen science fails, m arketing w
ork

65

4
T
he depressing truth about h appy p
ill

307

3 . For original graph, see: Kirsch, I. (2009), The Emperor’s New Drugs: 
exploding the antidepressant myth. London: Bodley Head, (p. 10).

68

6

68

‘Instead of doing a brand-new study’, said Kirsch, ‘we 
decided to do what is called a meta-analysis. This worked 
by gathering all the studies we could find that had com-pared the effects of antidepressants to the effects of placebos 
on depressed patients. We then pooled all the results to get 
an overall figure.’ In total, Kirsch’s meta-analysis covered 38 
clinical trials, the results of which, when taken en masse, 
led to a startling conclusion. ‘What we expected to find’, 
said Kirsch lowering his teacup, ‘was that people who took 
the antidepressant would do far better than those taking the 
placebo, the sugar pill. We couldn’t have been more wrong.’ 
And if you look at the graph below you’ll see exactly what 
Kirsch means.3

69

69

The first thing you’ll notice is that all the groups actu-ally get better on the scale of improvement, even those who 
had received no treatment at all. This is because many inci-dences of depression spontaneously reduce by themselves 
after time without being actively treated. You’ll also see 
that both psychotherapy and drug groups get significantly 
better. But, oddly, so does the placebo group. More bizarre 
still, the difference in improvement between placebo and 
antidepressant groups is only about 0.4 points, which was a 
strikingly small amount.

69

So in the following months Kirsch and Sapirstein 
analysed and re-analysed their data. They cut the figures 
this way and that; counted the statistics differently; checked 
what pills were assessed in each trial, and re-examined their 
findings with colleagues. But each time the same results 
D rugP sychotherapyP laceboN
o treatment0
0
.20
.40
.60
.81
1
.21
.41
.61.8T he depressing truth about happy p ills 6
3came out. Either they couldn’t spot the mistake, or there 
was simply no mistake to spot. Eventually there seemed to 
be no alternative other than to take the risk and publish 
their findings that antidepressants, according to their data, 
appeared to be only moderately more effective than sugar 
pills.

70

‘Once our paper appeared’, Kirsch recalled smiling, 
‘there was, well, how can I put it? – controversy. The most 
significant critique was that we had left out many import-ant trials from our meta-analysis. Perhaps an analysis that 
included those studies would lead to a different conclu-sion?’ Indeed, a professor at George Washington University 
called Thomas Moore pointed this out to Kirsch by reveal-ing that his meta-analysis had assessed only the published
trials on whether antidepressants work. Their study had 
therefore failed to include the drug trials left unpublished
by the pharmaceutical companies who conducted them. 
Kirsch and Sapirstein had been unaware that pharmaceuti-cal companies regularly withhold trials from publication. 
When Kirsch looked into how many trials this amounted 
to, he was aghast at what he found: nearly 40 per cent of all 
the trials on antidepressant drugs had not been published – 
a staggering amount.

70

I asked Kirsch what he did next. ‘Well, Moore sug-gested we appeal to the Freedom of Information Act to 
get the unpublished company studies released, and once 
we were successful, we undertook a second meta-analysis, 
which now included all the studies – both published 
and unpublished.’ As the results came in from this sec-ond meta-analysis, Kirsch grew even more alarmed: they 
showed that the results of his first study were plainly wrong. 6 4 C
racked: Why psychiatry is doing more harm than goodAntidepressants didn’t work moderately better than place-bos – they worked almost no better at all.

71

illustrate to you how 
the studies into antidepressants that Kirsch’s meta-analysis 
surveyed were conducted.

72

6

72

Now, imagine that the clinical trial you have just par-ticipated in contained about 500 other patients, all going 
through the same process as you. This of course is a sig-nificant amount, but it’s still only one trial. What Kirsch 
did, you’ll remember, is pool the results of all the trials he 
surveyed – published and unpublished. So in effect Kirsch’s 
second meta-analysis collated the results of many thou-sands of patients, all of whom had been studied in trials like 
the one I have just described. And it was on the basis of this 
second analysis, as I mentioned a moment ago, that Kirsch 
reached the alarming conclusion that antidepressants work 6 6 C
racked: Why psychiatry is doing more harm than goodhardly better than placebos. Here’s what his results looked 
like:

73

As in his first meta-analysis, which looked only at 
the published trials, Kirsch’s second meta-analysis, which 
assessed both published and unpublished trials, revealed 
that both placebo and antidepressant groups got better. 
But his second meta-analysis also revealed that the differ-ence in rates of improvement between the antidepressant 
group and the placebo group was insignificant. And that’s 
the important bit. ‘After surveying all the trials, we discov-ered that the antidepressant group only improved by 1.8 
points on the Hamilton Scale over the placebo group’, stated 
Kirsch. ‘Now, this may not mean much to you. But what if I 
were to tell you that your score can be reduced by a full 6.0 
points if you are merely sleeping better? Well, you’d rightly 
conclude that 1.8 is a tiny difference. And that’s precisely 
why the NICE [National Institute for Clinical Excellence] 
has said there must be a difference of at least 3.0 points for 
the difference to be deemed clinically significant. Yet, the 
difference we found was only 1.8 points – totally clinically 
insignificant.’

71

To do this, imagine yourself in the following scenario. 
You’ve been depressed for at least two weeks (the minimum 
time needed to be classed as depressed, as you’ll recall from 
Chapter  1). So you eventually decide to drag yourself to 
your doctor, who asks you if you would like to participate 
in a clinical trial. The aim of this trial is to test the effective-ness of a new antidepressant drug (which may cause some 
side-effects if you take it). Your doctor then explains how 
the trial will work: before you are given the antidepressant 
your level of depression will be measured on something 
called the Hamilton Scale. This is a scale that runs from 0 to 
51, and the task is to work out where you sit on this scale. 
Your doctor explains that, to work out where you sit, the 
trial researcher will ask you a number of questions about 
yourself, such as whether you’re sleeping well, whether you 
have an appetite, whether you’re suffering from negative 
thoughts and so on. You’ll then be given points for each 
of the answers you give. For example, if you answer that 
you’re sleeping well you’ll be given 1 point, whereas if you 
say you’re hardly sleeping at all you’ll be given 4 points. The 
more points you accumulate the higher you are rated on 
the scale, and the higher you are rated on the scale the more T he depressing truth about happy p ills 6
5likely you are to be classed as depressed. That’s how the 
Hamilton Scale works: if you’re rated at 26 you’re thought to 
be more depressed than if rated at 19. You’ve got the idea …

72

Now, after this initial assessment, the trial researcher 
will place you in one of two groups of patients and prescribe 
a course of pills. But there’s a catch. You’ll be told that only 
one group of patients will be prescribed the real antide-pressant, while the other group will be given a placebo pill 
– a pill that’s made of sugar and that therefore contains no 
active chemical properties. No one will be told which group 
they are in. Nor will anyone know what pill they are tak-ing until their treatment has ended some three months later 
and their levels of depression have again been measured on 
the Hamilton Scale. Your first rating and second rating will 
then be compared. If your rating has gone down after treat-ment it means you have improved (and thus the pill has 
worked), and vice versa. Once the researchers gather the 
ratings for all patients in the trial, they can then compare 
the two groups to assess how superior the antidepressant is 
to the sugar pill in alleviating depression.

73

Indeed, a 1.8 difference on the Hamilton Scale is barely 
noticeable in terms of a person’s actual experience. But what 
was also interesting for Kirsch was that even the tiny 1.8 
difference between the antidepressant and placebo groups 
still didn’t mean antidepressants worked better than pla-cebos. ‘Just remember how the clinical trials work’, Kirsch 
explained to me. ‘You are told by your doctor you’ll be given 
one of two pills. You are also told that the antidepressant pill 
will produce side-effects in most patients [like drowsiness, 
diarrhoea, nausea, forgetfulness, dry mouth and so on]. So T he depressing truth about happy p ills 6
7what happens if after taking the pill, you start to experience 
some of these side-effects? Well, like most people you’ll fig-ure out that you’re on the real drug. And believing you are 
on the real drug you’ll now expect to get better. And it’s this 
increased expectation of recovery that actually helps you 
improve.’
In other words, side-effects increase the placebo effect. 
And this is how Kirsch could account for the tiny 1.8 per 
cent improvement in the antidepressant over the placebo 
group.

74

Kirsch’s second meta-analysis was far more 
dramatic than the first: it concluded that the new wave of 
antidepressants heralded as wonder drugs – Prozac, Seroxat 
(Paxil in the US), Lustral (Zoloft), Dutonin (Serzone), 
Cipramil (Celexa) and Effexor – worked no better than 
dummy pills for the vast majority of patients. There were 
about 10 to 15 per cent of people, the extremely depressed, 
for whom these pills worked in a very minor way (about 
4 points better than placebos on the Hamilton Scale), but 
this meant, as Kirsch pointed out, that ‘85 to 90 per cent of 
people being prescribed antidepressants are not getting any 
clinically meaningful benefit from the drug itself ’.

307

6 . Kirsch, I., (2008), ‘Challenging Received Wisdom: Antidepressants and 
the Placebo Effect’. McGill J Med.

76

6

76

A further criticism, perhaps even more quixotic, was 
that even if the drugs don’t work, it was still wrong of Kirsch 
and his colleagues to have published their results. Patients 
should be protected from findings that could undermine 
their faith in treatment. Kirsch disagreed adamantly: 
‘Without accurate knowledge, patients and physicians can-not make informed treatment decisions, researchers will 
be asking the wrong questions, and policy-makers will 
be implementing misinformed policies. If the antidepres-sant effect is largely a placebo effect, it is important that we 
know this. So that improvement can be obtained without 
reliance on addictive drugs which have potentially serious 
side-effects.’6

76

A final suite of criticisms was aimed at the methods 
Kirsch used. ‘There were a number of papers criticising our 
statistical methods and redoing them’, said Kirsch, ‘some in 
appropriate ways, some in inappropriate ways, and some just 
making careless mistakes. But no matter how these analyses 
were done, nobody ever passed the 3.0 point threshold for 
clinical significance, which only gave additional support to 
our own conclusions.’

76

So Kirsch’s research withstood the criticisms. But were 
there any other studies that actually replicated his findings? 
I put this question to Walter Brown, professor of psychiatry 
at Brown University, who has co-authored two studies ana-lysing the same set of clinical trials that Kirsch surveyed. 7 0 C
racked: Why psychiatry is doing more harm than goodHis answer was unequivocal: ‘We pretty much found the 
same thing as Kirsch. For a small minority of patients (the 
most severely depressed), our studies showed that antide-pressants may have some minor benefits. But for mildly/
moderately depressed patients’, said Brown earnestly, ‘our 
results confirm that antidepressants offer no advantage over 
placebos, alternative therapies, or even moderate exercise.’ 
In other words, Brown’s research confirmed that the vast 
majority of people taking antidepressants do not receive 
any chemical benefit.7
‘There is no question that these drugs 
are over-hyped to the general public’, reiterated Brown. ‘The 
research shows they are not as good as the psychiatric estab-lishment and the pharmaceutical industry claim they are.

307

7 . Khan, A., Redding, N., Brown, W.A. (2008), ‘The persistence of the 
placebo response in antidepressant clinical trials’. Journal of Psychiatric 
Research,

77

While Irving Kirsch and Walter Brown reached the 
same conclusions independently, so too did a major study 
of antidepressants that the NHS commissioned. This NHS 
study also declared that the difference between placebos 
and antidepressants is so modest that for mild to moder-ate depression antidepressants are not worth having at all. 
‘Our results were again like Irving Kirsch’s’, said Dr Tim 
Kendall, lead author of the study published in The Lancet.
‘Our widespread comparative meta-analysis of antidepres-sants showed pretty clearly that the difference between 
the published and unpublished studies of antidepressants 
in children, was that for the published trials, all the drugs 
worked, while for the unpublished trials none of the drugs 
worked. And if you looked at the published and unpub-lished combined, you’d probably recommend the use of 
only one drug for childhood depression.’

78

7

78

Given the results of studies outlined above, why do 
the regulatory agencies that evaluate antidepressants con-tinue to approve these drugs for public use? Well, the key 
to answering this question is to realise that the regulatory 
agencies do not take into account the results of negative tri-als when deciding whether to approve an antidepressant.
For example, the Food and Drugs Administration (FDA) 
in the United States and the Medicines and Healthcare 
Products Regulatory Agency (MHRA) in Britain require 
that a company show in just two clinical trials that their 
drug is more effective than a placebo.8
This is the case even 
if there are five, ten or fifteen clinical trials showing nega-tive results. In other words, regulatory agencies discard the 
negative trials. And they do this no matter how many there 
are. All they require are two positive trials to give the green 
light for public use.*

307

8 . After asking the MHRA about the number of trials needed to approve 
an antidepressant, they responded: ‘As a general rule a minimum of two 
studies is required to prove efficacy of a drug. A single study will have 
to demonstrate very compelling results to be considered sufficient alone 
to demonstrate efficacy.’ In other words, it’s possible for a drug to be 
approved on the basis of only one study.

78

  • But this is not all. For a trial to be considered ‘positive’, the difference 
    between placebo and antidepressant does not have to be clinically significant 
    (e.g. it can be just one or two points on the Hamilton Scale), which essentially 
    means it can be small enough to make no real difference in peoples’ lives.

83

5
D
ummy pills and the healing p
ower