Drug Treatment in Medicine and Psychiatry: Papering Over Important Differences


The treatment of mental disorders with drugs is not the same sort of activity as the use of drugs in medicine. Psychiatric drugs do not target underlying disease or symptom-producing mechanisms; they create an altered state of mental functioning that is superimposed on underlying feelings and behaviours. The ethical implications of the two situations are different.

I am grateful to the editor of Epidemiology and Psychiatric Sciences, Corrado Barbui, for publishing my article about the drug-centred model of understanding drug action in psychiatry back in the autumn of last year, and to Carmine Pariante and Catherine Harmer and Phil Cowen who provide commentaries on the article.1 Responding to the commentaries and engaging in a live debate with Professor Pariante has been helpful in sharpening my ideas, and I am also indebted to some persistent questioners at recent talks I have given!

Pariante, Harmer and Cowen have all conducted interesting research that cuts across diagnostic boundaries. It is particularly striking, therefore, that both commentaries confirm the importance of the disease-centred model of drug-action for modern psychiatry. Both commentaries assert that psychiatric drugs work by targeting underlying brain-based abnormalities that are hypothesised to produce psychiatric symptoms. Harmer and Cowen refer to research on the association of dopamine and psychosis, and to work on antidepressant effects on emotional processing. They conclude that “increasing 5-HT [serotonin] function reverses a pathophysiological function central to the experience of depression.”2 Pariante describes experimental work with anti-inflammatory drugs as “targeting a biochemical system in the body, to induce downstream effects in the brain, to eventually affect processes relevant to depression.”

Both Pariante and Harmer and Cowen argue that drug action in psychiatry is essentially the same as it is in the rest of medicine. Pariante pleads “why can’t we accept that psychotropic medications are like all other drugs in medicine.”3

Pariante correctly points out (as have I) that most medical drugs do not target the ultimate cause of the diseases they are used to treat. Anti-inflammatory drugs do not treat the cause of an infection, for example, but may be useful in reducing the swelling, pain and irritation that is produced by the body’s inflammatory response to an infective agent. Anti-asthma drugs, like salbutamol, do not address the biological mechanisms that cause asthma in the first place, but they relieve the symptom of breathlessness by reversing airways constriction.

But the point is that in psychiatry, despite what these authors argue, we have no idea what mechanisms are behind the patterns of feelings and behaviours we call symptoms, and no evidence that the drugs we use act on these mechanisms. We have no idea what biological processes are even associated with depression, schizophrenia or any other mental disorder, let alone evidence of any causative processes. Even if we did, this would not be sufficient to enable us to ignore the general effects that psychiatric drugs exert on mental activity.

Harman and Cowen argue that there is sufficient research implicating dopamine dysfunction as the basis of psychosis. I have already provided a comprehensive critique of this research.4 Just to repeat a few points: some antipsychotics like clozapine have relatively weak effects on the dopamine system, and stronger effects on other neurochemical systems; we do not know the neurochemical basis of the psychosis-inducing effects of amphetamine, and amphetamine affects a range of neurotransmitters, not just dopamine; most tests of dopamine activity show no differences between people with psychosis or schizophrenia and those without; tests which do show differences have not controlled for the many other things that are known to affect dopamine activity including stress, movement, smoking and in many studies the residual effects of current or prior antipsychotic treatment.

Harmer and Cowen also cite their work on the effects of antidepressants on emotion processing. This work is interesting, and Harmer and Cowen are to be congratulated on trying to investigate the way antidepressants alter ‘normal’ mental functioning, and for considering the impact of subjective alterations like sedation, but the results are not consistent or convincing.

Take one typical example.5 The researchers gave 24 volunteers a single dose of the antidepressant duloxetine or a placebo, and measured their responses to pictures of emotional expressions and their ability to correctly classify and recall words representing agreeable and disagreeable personality characteristics 6 hours later. The strongest finding was that people who took the duoloxetine were more likely to recognise the expression of disgust than those who took placebo (p=0.002). They were also slightly more likely to recognise a happy expression (p=0.05). There were no differences in recognition of anger, fear, sadness, surprise or neutral expressions. There were no differences in classification or correct recall of personality characteristics, but people taking duoloxetine were slightly more likely to falsely recall ‘positive’ personality descriptors than those taking placebo (p=0.04). Duoloxetine made people feel dizzy, anxious, nauseous and sad, and they reported impaired mood and energy levels. The authors concluded that the experiment demonstrated rapid effects on emotional processing that are independent of the subjective alterations reported, but results do not support the hypothesis that duloxetine reduces negative thinking (bias) or increases positive thinking, especially as there was no correction for multiple testing.

Even if we had evidence that dopamine activity causes psychosis, or low serotonin causes depression, we still have to account for the fact that changing the brain through drugs, surgery, injury or disease alters the nature of our subjective experience and behaviour in various ways. Changing the brain can reset the substrate of our mental life, by superimposing a new and altered state of brain functioning. This new state interacts with and can override pre-existing mental states and their associated behaviours, including those we refer to as depression, anxiety, psychosis etc., without necessarily having any specific impact on the neurological processes that may be associated with or productive of these states.

We all know this if we think about the effects of alcohol. We talk about using alcohol to “drown our sorrows,” without implying that we think alcohol is specifically targeting the mechanism of that sorrow. The phrase refers to the fact that the altered state produced by alcohol is superimposed on underlying feelings, temporarily overriding them. All drugs that have what we might call ‘psychoactive effects’ can superimpose the alterations they produce on existing emotions, cognitive functions and behaviour, and this includes all the drugs commonly prescribed for mental health problems.

Psychoactive drugs affect normal mental activities including thinking, perception, emotion and behaviour in characteristic ways. We are familiar with the sort of alterations produced by recreational drugs, but have paid less attention to those produced by other drugs prescribed for mental disorders, and some drugs prescribed for physical disorders (steroids, for example). Nevertheless, like alcohol, opiates and cannabis, drugs like antipsychotics, antidepressants and lithium produce particular mental alterations, that are linked with some of the physical alterations they produce (see the Table of psychoactive effects in this 2015 paper, p. 2316). The point is that unless we discount the impact of these alterations in some way, we cannot conclude that a particular drug achieves its effects through targeting a particular brain mechanism.

Pariante correctly points out that medical drugs change the whole body in various ways too. Chemotherapy for cancer alters cell reproduction processes in general, and is not restricted to effects in cancer cells, hence its debilitating and sometimes dangerous adverse effects. Nevertheless, it impacts on cancer cells through inhibiting their tendency for uncontrolled reproduction. It acts specifically, therefore, on the abnormal biological mechanism that produces cancer. If chemotherapy did not act on mechanisms relevant to the production of cancer it would not work. It is not its general effects that are useful, these are in fact harmful; its benefits result from its specific effects on the processes that drive cancer. In contrast, with psychoactive drugs their general impact on normal mental and behavioural functioning can, in and of itself, account for their impact on symptoms of mental disorders. There is no need to postulate action on particular ‘disease’ or symptom-producing mechanisms at a cellular, chemical or physiological level.

The mechanism of effects of psychoactive drugs on symptoms of mental disorder differs therefore from the manner in which most medical drugs achieve their effects. Most drugs used in general medicine can be understood to work according to a disease-centred model by acting on physiological mechanisms that produce symptoms, even if they also affect other systems. As I described in my original article, there are a few exceptions which involve the use of psychoactive drugs, such as opiates for pain relief. Unlike some other pain killers, opiates are psychoactive drugs that produce general mental alterations along with their direct effect on pain conduction systems. The emotional indifference produced by opiates means that people sometimes say that they still have some pain, but do not care about it anymore. The state of emotional indifference is superimposed on people’s experience of pain, lessening its impact, and this effect can be distinguished from the ability of opiates and other pain killing drugs to reduce pain sensations directly.

In my theory of drug action, I do not argue that it is impossible to find drugs that target the mechanisms underlying mental disorders, I just point out that we have no evidence that any of our current drugs work in this way. But we will not be able to conclusively demonstrate that any psychoactive drugs have a disease-targeting action unless we can discount the impact of their general psychoactive effects.

The fact that we have failed to pin down the mechanisms of ‘normal’ mental states or of ‘mental disorders’ so far may reflect more general differences between the nature of human beings and their biology. Human behaviour consists of complex, intentional and unpredictable responses to the unique history and circumstances of each individual. Unlike physical systems, including biological ones (i.e. human bodies), it cannot be captured or understood using universal formulae. Human behaviour can be explained and understood, but it is not ‘caused’ by other events in an inevitable fashion, as events follow each other in a mechanical system (see my previous blog on the philosophy of knowledge7). Although there are undoubtedly neurophysiological events taking place when someone feels depressed, for example, it is not clear that we will ever be able to map these precisely and consistently onto the emotional state. Indeed, for all the reams of research conducted on them, we still do not even know the precise functions of neurotransmitters, nor even, for example, the neurochemical basis of something as basic as arousal.

So, more funding for more research into identifying drugs with targeting actions as advocated by Carmine Pariante may just be pouring good money after bad. Instead of hankering after a situation we may never achieve, what I am calling for is a more sophisticated, transparent and appropriately cautious approach to the use of existing mind-altering chemicals.

Take the example of ‘antipsychotics’, and remember that these drugs were initially referred to as ‘neurological inhibitors’ and ‘major tranquilisers’ by people who recognised the alterations they produce. In human volunteers and animal studies they produce a state of reduced activity and responsiveness to the environment, and reduced emotional reactivity, initiative and motivation (albeit with distinctions between individual agents). We can see immediately that this state will impact on someone who is preoccupied with delusional beliefs or internal experiences, reducing the intensity of psychotic symptoms and their emotional salience, along with other aspects of subjective experience, without necessarily having any effect on specific mechanisms underlying psychosis. We can also see that although the effects may be helpful in reducing the intensity of psychotic symptoms, they may have a detrimental effect on an individual’s functioning and quality of life.

Despite Pariante’s desire to align the use of drugs in psychiatry with the rest of medicine, the use of drugs that ‘reset’ normal mental processes to modify feelings and behaviour is a fundamentally different sort of activity from using drugs to target recognised bodily pathologies. There are some points of commonality, of course. In psychiatry, as in medicine, the decision to intervene with drugs or other means depends on a consideration of the relative harms and benefits of doing so. Use of anti-epileptic drugs may cause more harm than good after one or two fits, for example, but when fits are recurrent and life-changing, the adverse effects may be worth putting up with. Similarly, the benefits of using an antipsychotic to suppress psychotic symptoms may outweigh the harms that may be incurred when someone is acutely psychotic, but the balance may be more uncertain after they have recovered.

Assessing the benefits and harms of interventions that change people’s thinking and behaviour is more complicated than weighing up the effects of a drug with purely physical effects, however. Our mental life is what makes us what we are. It is fundamental to our individuality and sense of ourselves. Moreover, people have different views on the desirability of the feelings and behaviours we refer to as ‘mental disorders’. Mental health legislation exists because when people are in states of mental turmoil and confusion, they may not see their situation in the ways that others see it. They may not agree that anything is wrong or that anything is in need of changing. When we use drugs to change people’s behaviour in such situations, we are doing something akin to restraining them. We are using force to prevent behaviour we do not like (possibly for good reason, if that behaviour is dangerous). Some people will thank us when they have recovered, but we know that many do not. Many will never see the world quite as others see it.

Typical medical treatment and the use of drugs in psychiatry have different ethical implications, therefore. There is usually agreement that treating harmful bodily variations is desirable. There is less likely to be agreement about inducing certain mental and behavioural changes. Insisting on equating the two situations obscures these differences and presents the use of drugs for mental distress and disorder as less controversial than it actually is.

  1. Moncrieff J. Research on a ‘drug-centred’ approach to psychiatric drug treatment: assessing the impact of mental and behavioural alterations produced by psychiatric drugs. Epidemiol Psychiatr Sci 2018 Apr;27(2):133-40.
  2. Harmer CJ, Cowen PJ. How do drugs for psychiatric disorders work? Epidemiol Psychiatr Sci2018 Apr;27(2):141-2
  3. Harmer CJ, Cowen PJ. How do drugs for psychiatric disorders work? Epidemiol Psychiatr Sci2018 Apr;27(2):141-2
  4. Moncrieff J. A critique of the dopamine hypothesis of schizophrenia and psychosis. Harv Rev Psychiatry 2009;17(3):214-25
  5. Harmer CJ, Heinzen J, O’Sullivan U, Ayres RA, Cowen PJ. Dissociable effects of acute antidepressant drug administration on subjective and emotional processing measures in healthy volunteers. Psychopharmacology (Berl) 2008 Sep;199(4):495-502
  6. Yeomans D, Moncrieff J, Huws R. Drug-centred psychopharmacology: a non-diagnostic framework for drug treatment. BJPsych Advances 2015;21:229-36
  7. Moncrieff J. Philosophy Part 3: Knowledge of mental states and behaviour – insights from Heidegger and others. 2017 November 1

Article first published on Mad in America, 29 June 2018.


  1. Do drugs in the rest of medicine really target underlying disease mechanisms? I agree that psychiatry and medicine are different in that there is no physical cause for functional mental illness. However, there is empirical use of drugs in medicine, besides psychiatry. I’m not convinced the best way to define the difference between psychiatry and medicine is on the basis of the way drugs are used.

  2. Hi Joanna,

    I’m interested if you have read some of the opposing literature on DSM/ICD classifications of “mental disorder” – or indeed the helpfulness of a concept such as “abnormality” when it comes to looking at how human beings think, feel and behave in any given set of circumstances.

    I understand that the british psychological society are attempting to develop alternatives to the pathology model of human distress and that this work has been brewing for a long time, starting with psychologists actually listening to people and discovering that their problems, thoughts, feelings and behaviours all made perfect sense once they understood enough about the person and their context .

    There is even less evidence for the idea of “abnormality” or “psychopathology” than for believing that psychiatric drugs mend them. In fact, isn’t what is “normal” or “abnormal decided by a committee? Even if the neural activity associated with you call “psychosis” and “mania” is identified, doesn’t that tell us nothing more than that similar thoughts and feelings take place in similar areas of the brain, or as the result of similar brain activity?

    Given that there is ample evidence for the power of suggestion, isn’t it unwise to call certain problems that can arise in life – usually due to trauma or extreme crisis “pathological” or “abnormal”? And what could possibly be gained for a person in so labelling their experience?