Rethinking Models of Psychotropic Drug ActionJoanna Moncrieff; David CohenPsyc:lOtherapy and Psychosomatics; 2005; 74, 3; ProQuest Psychology Journalspg.145

Innovations

Psychother Psychosom 2005;74:145-153001: 10.1159/000083999

Rethinking Models of PsychotropicDrug Action

J. MoncrieW D. Cohenb

'Department of Mental Health Sciences. University College London. London. UK; "School of Social Work.College of Health and Urban Affairs. Florida International University. Miami, Fla., USA

Key WordsPsychopharmacology' Psychotropic drugs'Disease-centred model' Drug-centred model

AbstractTheoretical assumptions about how psychotropic drugs'work' are rarely discussed explicitly. In a 'disease-cen­tred model: drugs are believed to work by acting on adisease process. In contrast, in a 'drug-centred model:the characteristic physiological, behavioural and subjec­tive effects of drugs are used to define drug action. Thetherapeutic value of a drug stems from the usefulness ofthese effects in clinical situations. The disease-centredmodel appears dominant but has weaknesses: (1) it can­not logically justify the use of drugs since major patho­physiological hypotheses were derived from selectivelyobserved actions of drugs; (2) comparisons betweendrugs believed to have specific effects in certain condi­tions and drugs thought to have non-specific effects failto support it; (3) outcome measures for various disordersinclude items responsive to non-specific drug effects;(4) studies with healthy volunteers describe characteris­tic drug-induced states independently of a psychiatricdiagnosis; (5) animal tests show effects with agents notusually thought of as specific treatments for the condi­tions modelled by tests. This article offers suggestionsto develop a drug-centred model and discusses its po­tential impact on clinical practice.

Copynght © 2006 S. K.rgor AG. B..ol

Modern psychopharmacology, or the study, classifica­tion and clinical use of psychotropic dmgs, developed inthe 1950s alongside the introduction of new drug treat­ments in psychiatry. However, in contrast to numerousempirical descriptions of dmg effects on aspects of psy­chopathology, there are few discussions oftheoretical as­sumptions about dmg action. Such assumptions nonethe­less exist and influence the treatment ofmental disorders.Indeed, psychiatric drug treatment is currently predicat­ed on a 'disease-centred' modcl ofdrug action, which pro­poses that most psychiatric. drugs act as specific treat­ments for specific conditions. Current nomenclature em­bodies this position with names like antipsychotics,antidepressants, anxiolytics, antimallics and mood stabi­lisers. Models describe general principles to help achievea deeper understanding of natural and social processes.Models guide scientific inquiry and produce therapeuticadvances that may, in turn, lead to the development ofmore complex models. In this article we outline the dis­ease-centred model of drug action and critically evaluatethe different lines of evidence that support it. We alsodescribe an alternative, 'drug-centred' approach andsome of its treatment implications.

KARGERFax +41 61 306 12 ~4

&Mail kargtr@ka!i:(r,chwWYo'.ktlrger.~om

© 2005 S. K.a'iOf AG, B.",1OO~~3190if)51074~145$22,1JOIO

Accessible onlJne ;:at:WYt'w.k~rger.com/~lJX;

Dr. Jc,aona MQDcrieffMaSl.~(jJJ:j Patk HLI~pita)

Maocalls l.f1fit\ Brcntw'c)odEssn eM 14 5lIQ (UK)Tel. +44 1277302695. Fa1\' +4-41277 3026S16: E-M~j)j.m ..mr;ricfr@nd.ac.IJk

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Models of Drug Action in PsychiatricDisorders

The Disease-Centred jI"fode/The disease-centred model underlies orthodox psy­

chopharmacology. Its core assumption is that psychotro­pic drugs help to correct a biochemical abnoffilality thatrepresents a biological substrate of a specific disease pro­cess. This notion, borrowed from certain paradigmatictreatments such as the usc of insulin in insulin-dependentdiabetes, is sometimes called the 'chemical imbalance'theory of mental disorders. Although this model is rarelyexplicated nowadays, its influence can. be inferred fromthe classification of psychotropic drugs according to thedisorders they are believed to treat. In turn, efficacy trialsand animal research are principally organized around thisclassification. In clinical practice, the disease-centredmodel is often presented to patients as the basis for theirneed to take medication, wit.h many psychiatrists drawinganalogies between mental disorders and diabetes or hy­peliensioIl, for example. Some patient information, pub­lished by professional bodies and pharmaceutical firms,states explicitly that drugs work by correcting biochemi­cal abnormalities [1,2, pp. 181-182].

A variant of this model is the symptom-centred mod­el. Here, drugs are believed to act on the pathologicalprocesses producing the symptoms rather than on the un­derlying disease process. For example, antipsychoticdrugs are hypothesized to disrupt the production of psy­chotic symptoms by blocking dopamine over-activity [3].This might be analogized by the action ofanalgesics suchas opiates, that reduce and inhibit transmission of noci­ceptive stimuli along spino-thalamic pathways, or non­steroidal anti-inflammatory drugs, that inhibit the pro­duction of prostaglandins involved in producing painthat arises from inflammatory reactions [4]. Although thismodel appears more pragmatic, it still rests on notionsthat drugs work by atTecting specific abnormal biochemi­cal or neurophysiological processes that give rise to symp­toms.

In a rare recent discussion of the thl.-'Oretics of psycho­tropic drug action, Hyman and Nestler [5, 6] reject theemphasis 011 synaptic neurotransmitters as the basis oflmderstanding drug action, proposing instead that thera­peutic dfects result from impacts OIl new1l1 circuits. Al­though these authors attempt to address some problemsof conventional views of drug action, their ideas remaingrounded in a disease-centred model that assumes that'pharmacotherapeutic agents produce their clink-allybeneficial effects in an abnormal nervous system' and that

these effects 'counter or compensate for the abnomlalpathophysiology' [6, p. 440].

The Drug-Centred ModelThe disease-centred model of drug action is rejected

or seen as limited by critics, practitioners and scholars ofpsychopharmacology [2, 7-11]. Common threads in theirarguments include: (a) the high degree of integration ofthe central nervous system, such that even drugs withspecific targets necessarily produce non-specific actions.(b) the lack ofvalidation of a disease model of psychopa­thology and (c) the clinical use of similar drugs for differ­ent disorders and the use ofpharmacologically dissimilardrugs for similar disorders. Existing critiques converge tosuggest that the evidence on psychotropic drug effectspoints to the validity of a 'drug-centred' model.

In this approach, mugs are seen to induce characteris­tic physiological and subjective states that may, or maynot, be experienced as useful in certain soc·ial and inter­personal situations. including clinical situations. Unlikethe disease-centred model that assumes that dmgs movean abnormal physiological state towards a more normalone, the drug-centred model suggests that drugs createtheir own characteristic abnormal states or alterations ofnormal states. It is these states or effects that need to bedescribed and understood, and the potential therapeuticvalue of a drug is deduced from this understanding. It istherefore implied that diagnosed patients and normal vol­unteers' basic physiological responses to drugs will differonly insofar as a degree of individual variation in drugresponse (including variation in arousal, set, biologicalsensitivity) always exists.

Historically, an elementary drug-eentred c1a.'lsiflcationofdrug action distinguished drugs on the basis ofprimar­ily sedative (or 'depressant') and stimulant effects. Amore elaborate classification might distinguish betweendifferent tyIX~sofsedative effectsofconventional antipsy­chotics. tricyclic antidepressants, benzodiazepines. bar­biturates and opiates and could start to characterize thesedation from the newer antipsychotics. Similarly, stimu­lating etfects of 'classic' psychostimulants, some types of'antidepressants', and other drugs could be differentiatedmore finely. Such a classification would also need to con­sider drugs that calL')e sedation and agitation simultane­ously, such as some antipsychotics and antidepressants.Other etlects, including psychomotor indifference, akine­sia, akathisia, hallucinogenesis, euphoria and dysphoria,require further elaboration an.d suggest yet more ways tocharacterize drug actions.

146 Psyc!1other Psychosom 2005;74; J45-153 MoncrietT/Cohen

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

The case of alcohol briefly illustrates the implicationsof the model. Alcohol reduces conductivity in the centralnervous system. Ingestion of alcohol gives rise to charac­teristic physiological effects, such as vasodilation andslowed reaction times, and to various characteristic sub­jective experiences and behavioural effects such as eu­phoria, social disinhibition and sedation. These effects- usually dose dependent - have several consequences.They are responsible for the popularity of alcohol as asocial lubricant and rccreational substance; they can leadto a~e.rcssive and reckless behaviour in some circum­stances; they can produce \vithdrawal syndromes afterprolonged use at high doses, and they may help people toovercome some behavioural inhibitions. Alcohol mighttherefore be seen as a possible treatment for 'social pho­bia', not because the substanee corrects an underlyingphysical abnormality in social phobia, but because onetype of effect produced by alcohol might in itselfbe use­ful for people experiencing diffi.culties in some interper­sonal or social situations. In this connection, dramaticbeneticial effects of ethyl alcohol on patients with schizo­phrenia have been described [12, p. 287; 13].

Other examples of drug-centred thinking were provid­ed by early proponents of modem psychopharmacology.Pierre Deniker, credited with the 11rst major psychiatricuse ofchlorpromazine, thought that its useful effects wereattributable to the induction ofan abnormal encepha1itis­like state characterized by 'psychomotor inditIerence'[14, p. 92]. Others described chlorpromazine's particularadvantages as its ability to produce 'a calming effect witha minimum ofdrowsiness and confusion' [15, p. 540], ora 'pathological tranquillity of mind' [16, p. 961].

Development of the Disease-Centred Model

Prior to the 1950s, a 'drug-centred' pragmatic modelguided the use of drugs in psychiatric practice, althoughother theoretical frameworks were sometimes proposed.Drugs were classified crudely into sedatives and stimu­lants, as ex.emplificd by Sargant and Slater's [17] discus·sion of 'chemical sedation and stimulation' in their 1944textbook of physical treatments in psychiatry. These au­thors recommended sedative drugs to induce sleep andto calm acutely agitated patients. Sargant and Slater spe­citically recommended prescribing phenobarbitone as a'basic sedative and not pro re nata' [po 87; italics in orig­inal translate as 'for the thing of origin']. Sargant andSlater did not general1y find stimulants useful in psychi.atric conditions, because stimulant and euphoric effects

rapidly diminished with continuing use. However, theseauthors did consider stimulants to be useful in childrenwith clcctro-cnccphalographic abnom1alitics and hyper­activity, in which symptoms 'may yield to the drug inwhat appears a specific way' [po 96]. Apart from the latterexample, however, drugs were not seen as exerting effectson the wlderlying condition bcing treated. For this, treat­ments such as electroconvulsive therapy or insulin comawere seen as necessary. In 1954. an influential Englishtextbook of psychiatry by Mayer-Gross et al. [18] statedthat 'hypoglycaemic treatment clearly touches the physi­cal basis of schizophrenia more closely than all earliermodes of attack' [po 286]. This was one of many expres­sions ofa desire for a specific therapy in psychiatry beforethe modern drug treatment era [13, 19].

In contrast to earlier ideas about drugs, views thatemerged from the 1950s onwards fairly rapidly came tocharacterize drugs as having specific ell'ects in differentconditions, and drugs began to be classified basically ac­cording to the condition for which they were feit to beeffective [19]. Recent histories of psychiatry and psycho­pha.rmacology [8, 20, 21] suggest numerous factors thatmay have reinforced the adoption of a disease-centredmodel ofdrug action. These factors include: the desire todevelop psychiatric treatments with similar specificity assome other medical treatments; the neo-Kraepeliniantrend toward viewing disorders as discrete entities withspecitic aetiologies [22]; the developmcnt of molecularbiological tools and the resulting focus 011 synaptic hy­potheses of drug action; requirements of drug licensingbodics [9], and the penetration of psychiatric thinking bythe marketing language used by phamlaceutical compa­nies [2].

However, the historical evidence suggests that thespeciticity of the new dmg treatments was assumed or as­selted before authors discussed what it might mean for adrug to be specific for a particular psychiatric condition,and what sort of evidence might be needed to reach thisconclusion. Although the assumption that psyehiatricdrugs are specific still underlies most research in clinicalpsychopharmacology and most professional and populardiscourse, thc case that psychiatric dnlgs are specific ei­ther to diseases or to pathological processes is far fromestablished. In the following section, it is argued thatmuch of the evidcnce that would be needed to justify thisposition is lacking, and the evidence thought to supportit is often inadequate.

Models ofDl1Ig Aciion Psychmher Psychosom 2005;74:! 45-153 147

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Deficiencies of the Disease-Centred Model

Derivation o/Disease Alodelstrom Observations ofDrug E..tJectsThe major justification for a disease-centred model of

a drug's action is ifthc pathophysiology of the disease isdescribed independently ot~ and usually prior to, the uscof the drug treatment. 'Vithin this framework, an etrec­tive drug - such as insulin for diabetes or penicillin for abacterial infection - is one that affects a part ofthis patho­physiology. However, modern pathophysiological hy­potheses in psychiatry either derive largely from theknown or presumed mechanisms of actions ofdrugs usedto treat the conditions, or have been adapted to accom­modate drug action. The most famous, the dopamine hy­pothesis ofschizophren ia. arose directly from the elucida­tion of some pharmacological effects of the first antipsy­chotic. drugs [23]. Subsequent refinements, such asspeculation about the role of serotonin in psychosis, re­sulted from using antipsychotics with actions on a seem­ingly greater number ofneurotransmitter systems includ­ing serotonin. The monoamine hypothesis of depressionwas also elaborated in conjunction with research on theactions of antidepressants [8]. Hypotheses about the neu­robiological basis of anxiety also derive from observa­tions of benzodiazepincs' effects on the gamma-amino­butyric acid receptor complex [24], while the clinical useof selective serotonin reuptake inhibitors has increasedtheorizing on the role of serotonin in anxiety [25]. Mostideas about the physiological and biochemical underpin­nings of psychiatric conditions therefore assume, ratherthan provide compelling support for, the disease-centredmodel of drug action.

Research attempting to find independent evidence ofthe suggested biochemical abnomlalities has not, to date,produced conclusive findings in any mental disorder.Some recent imaging studies, showing increased levels ofdopamine activity in acute psychosis [26], have led to arenewed interest in the dopamine-psychosis relationship[3]. However, the evidence is inconsistent, and it remainsunclear whether abnormalities of dopamine activity arcspecific to schizophrenia or psychosis, or merely featmesof an altered state of arousal or some other aspect of anacute psychiatric condition.

Basing a model of drug action on the observed efficacyof drugs used to treat a given psychiatric condition raisestwo further problems. The l'1r5t is that 'efficacy' in psychi­atric disorders is a historically a.nd even geographicallyrelative construct. What it means for a psychotropic dmgto be considered 'efficacious' is subject to periodic revi-

sion based on changing nosological systems, drug regula­tory requirements, clinical trial methodologies and defi­nitions of relapse. Some medical treatments such as insu­lin for diabetes and antibiotics incontrovertibly returnfunctioning to nonnal or near normal. The effects ofpsy­chiatric treatmen.ts are more subtle, an.d desirable effectsunder one era's standards may not be valued in anotherera. Secondly, observations or inferences about drugs'modes of action are usually selective. Thus, biochemicalhypotheses of depression focus on the synthesis, release.metabolism and/or receptor sites of one or two membersof a single neurotransmitter family, wherea.<> antidepres­sants influence almost all neurotransmitters, most hor­mones and many neuropeptides [27]. Further, the initialsites of action of a drug, where pharmacological activityis more easily measured, are part of a long-lasting chainof adaptive events that usually overwhelm a drug's earlyactivity [5].

Failure to Establish Clear Differences withNon-,Specijrc DrngsDemonstrating that dmgs believed to be specific have

superior, or at least different, clinical effects than drugswith non-specific actions would seem to be a prerequisiteto establish specificity of action. Surprisingly, such stud­ies are rare and most were conducted decades ago. Forexample, evidence suggesting that antipsychotics are su­perior to other sedatives in the treatment of psychosis issparse and inconsistent. Two early trials found that vari­ous phenothiazines were superior to phenobarbital in pa­tients with acute and chronic schizophrenia [28, 29].However, a trial comparing opium and chlorpromazineproduced equivalent improvement over 3 weeks in acuteschizophrenia [30]. Wolkowitz and Pickar [31] fmmd 13double-blind trials comparing benzodiazepines with pla­cebo and/or neuroleptics. Six studies compared a benzo­diazepine and placebo for patients with acute and chron­ic psychotic diagnoses; only the largest study fOWId thebenzodiazepine to be markedly superior to placebo at astatistically significant level. However, in the 6 trials com­paring benzodiazepines with neuroleptics, the outcomeswere equivalent in 3, the benzodiazepine was superior in2, chlorpromazine was superior in I, and in 1 trial thebenz.odiazepine was equivalent to haloperidol but infe­rior to chlorpromazine. In 7 of the 10 studies where psy­chotic symptoms were measured separately, benzodiaz­epines reduced symptoms comparably to neuroleptics orbetter than placebo. Comparable effectiveness betweenbellzodiazepines and neuroleptics was recently observedin treating exacerbation signs in schizophrenia [32].

148 Psychother Psychosom 2005;74: 145-153 MoncI'ieff/Cohen

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Numerous studies of the treatment of depression com­pare agen.ts not primarily regarded as antidepressantswi th either placebo or standard antidepressants [33). S~~v­

eral antipsychotics, some benzodiazepines, barbiturates,opioids and stimulants as well as buprenorphine and bus­pirone, have shown superiority to placebo or equivalencyto antidepressants, However, given the deeply ingrainedassumption that antidepressants are specific treatmentsfor depression, such findings are usually explained by sug­gestions that these other agents may have 'antidepressantproperties' [34]. One altemative explanation suggeststhat antidepressants have non-specific effects. They maywork by causing sedation, which reduces agitation associ­ated with depression, induces sleep and may mask de­pressive feelings. Antidepressants may also work by en­hancing the placebo eft"l'X:t, as when. physiological reac­tions to a drug reveal or confirm to patients that they aretaking an active medication [35). In this case, almost any­thing might tum out to have antidepressant properties,and the literature inde~~d suggests this might be so.

Lithium has long been designated as a specific treat­ment for bipolar disorder. However, studies of treatmentof acute mania have not shO\\11 that lithium is superiorto neurolcptics, and it has been found to be inferior forthe treatment of highly overactive patients [36, 37]. Inaddition, studies comparing patients with manic, schizo­affective and schizophrenic psychoses have not shownthat lithium ditTers from antipsychotics in its effects indifferent diagnostic groups [38, 39]. The specificity oflithium as a prophylactic treatment has also not been es­tablished, since careful comparisons with other drugswith strong sedative etlects, excepting anticonvulsants,have not been done. In addition, although the efficacy oflithium as a prophylactic treatment is generally accepted,concerns have been raised about the methodological va­lidity of the placebo-controlled trials that established this[40). In particular, evidence that lithium withdrawal in.­creases risk of relapse over and above the untreated risk[41,42] suggests that the results of these trials may havebeen confounded by lithium withdrawal effects. Althoughthe reality ofa lithium withdrawal phcnomena is not uni­versally accepted [43), recent data provide further evi­dence tor its existence [44).

lnclusion ofNon-Specific Items in Dmg Rating ScalesCommon scales used tor rating outcomes in psycho­

pharmacology trials contain items describing behaviourthat would not nornlally be considered a specific part ofthe disorder being treated and would be expected to re­spond to non-specific efiects ofmedication. For example,

the 17-item Hamilton Rating Scale for Depression con­tains 7 items on sleep disruption and anxiety, all ofwhichcan be expected to respond to sedatiw cffects of drugs.Psychosis rating scales also contain items that relate toarousal and would also respond to sedative effects non­specifically. TIle Brief Psychiatric Rating Scale containsitems on tension, uncoopcrativcncss. excitement audhostility. Similarly, of 7 items on positive symptoms inthe Positive and Negative Symptom Scale, 2 items con­cern excitement and hostility. Tn dmg efficacy studies em­ploying such instruments that partly rate non-specific ef­fects, changes in the global scores do not necessarily sig­nify that a drug has a specific ctl't~ct on a particulardisorder.

Similar Effects in Diagnosed Patients and HealthyVolunteersThe disease-centred model suggests that drugs are like­

ly to have different effects in people with the conditionfor which the drugs are indicated compared with peoplewithout the condition [6]. However, surprisingly fewstudies describe the subjective and behavioural etlects ofpsychotropic drugs on human volunteers or patients withother conditions, and ctlects seen in volunteers arc usu­ally dismissed as side effects. For antipsychotic drugs,available studies show unam biguous drug effects consist­ing of impaired performance on psychomotor and cogni­tive tasks [45-47) and subjective effects, including seda­tion, dysphoria, akathisia, and feelings variously de­Si.';bed as disengagement, indifference or depersonaliza­tion [45,48). These effects are COil sistcnt with etTects seenin patients and usually described as side etTects. How­ever, it is also apparent that these same efteets might beresponsible for reducing agitation and psychotic symp­toms [49).

EfTects oflithium in volunteers include slowing ofper­formance on cognitive tasks, tiredness, lethargy, dyspho­ria and occasionally confusion [SO, 51] - experiences con­sistent with electroencephalogram changes showing in­creased slow wave activity [52]. This pattern of effectscould explain lithium's action in acute mania.

The few available studies ofantidepressants in VOIWl­

teers suggest that different antidepressants show differenteffects. This is consistent with the fact that antidepres­sants come from a variety of ditTcrcnt phamlacologicalclasses. For example, in one study reboxetine appearedto be mildly stimulant and sertraline to be mildly sedat­ing [53]. Amitripyline has bt~cn found to be profoundlysedating and cause cognitive impairment and electroen­cephalographic changes similar to those of chlorproma-

Models of Drug Action Psychother Psychosom 2005;74:145-: 5.~ 149

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

zine [54]. Effects on affect and mood in volunteers haverarely been demonstrated and are not clear cut. Conven­t.ionally, this is taken as cvidcnce that ant.idepressant ef­fects are only apparent in depressed patients. However,a recent meta-analysis found that overall therapeutic ef­fects in patients - as measured by the Hamilton RatingScale for Depression - were also small, and could easilyhave been achieved by non-specific effects such as seda­t.ion [55]. The current confu..'lion and lack of infonnationabout the physiological, behavioural and subjective ef­fects ofmost antidepressants illustrates a limitation ofthedisease-centred model.

Limitations ofAnimal Screening ResearchNumerous animal sereening tests are meant to iden­

tify compounds that might have specific effects on psy­chiatric disorders by using animal assay or homologousmodels of these disorders [56]. The limitations ofanimalmodels in use are widely acknowledged [57], but it is rare­ly pointed out that they often fail to discriminate betweensupposedly specific drugs and non-specific ones. For ex­ample, in all animal models of depression, responses areobtained with drugs that. are not generally considered tohave antidepressant activity in humans [58]. In the forcedswim test, one ofmatry tests used to screen for antidepres­sants. positive results have been obtained with amphet­amines, opiates, antihistamines, some ant.ipsychotics, at­ropine, pentobarbital as well as zinc and antibiotics [58,59]. In line with the underlying assumption that 'antide­pressant activity' can be specifically identified or isolated,some authors label these results as 'false-positives' [58].Conversely, the selective serotonin reuptake inhibitors,\videly considered to be specific antidepressants, typie-al­Iy fail to be detected by the forced swim test [60].

Developing a Drug-Centred Model of DrugAction

Jacobs and Cohen [61] have pointed out how little isknown about the 'psychological alterations' produced bypsychotropic dnlgs. These authors described the limita­tions ofthe randomized controlled trial for evaluating thefuJI range of a drug's effects and the spurious nature ofthe distinction between therapeutic and adverse effects.Other authors have also eriticized various aspects ofcur­rent. methodology [62, 63] and some have emphasized theimportance of finding new methods to explore 'treat­ment-specific efficacy' [64]. The dmg-centred model out­lined here suggests a programme for developing a tuBer

understanding of the effects of psychotropic drugs whichcould include the following features:• A priority on detailed investigations of what d(fferent

types ofdrug-induced experiencesconsistof. This wouldinvolve more studies with volunteers and patients ­over durations more closely approximating actual clin­ical treatment - that focus on the nature of the subjec­tive experience as well as physiological and behav­iouml measures of 'drug respom;e'.

• Developing outcome measures addressing particularbehaviours rather than disorders. Use ofoutcome mea·sures designed to mea<;ure disorders should be replacedby measures that address particular behaviours thatpatients or others desire to be modified.

• Constructing a new vocabulary ofdrug-induced effects.Forexample, different types ofsedative, stimulant andother drug-induced effects could be characterized.Such a vocabulary could provide the basis to groupdrugs according to similarities in the effects they pro­duce.

• Integrating the literature on adverse e.ffects with that ontherapeutic effects. According to a drug-centred modelofdrug action, the distinction between therapeutic ef­fects and adverse etfects is arbitrary. Research undersuch a model would aim to obtain a complete pictureof the range of a psychotropic drug's action.

• Investigating in more detail the potential bem:fits ofus­ing non-spec(fie dntgs tllat are better tolerated by pa­tients, such as bellzodiazepines, as the main treatmentfor acute psychiatric syndromes. This would indudecomparisons with standard psychiatric drug treat­ments.

• Evaluatingpatients' comparative preferences./or differ­ent types ofdrugs in various situations.

• Obtaining patients' post-treatment ratings ofdrug eifeets. The evaluation of a psychotropic drug may beconsidered incomplete until the user has had a chanceto look back on the drug-taking experience from adrug-free standpoint.

Implications of a Drug-Centred Model forPsychiatric Practice

Clink.al practice suggests that a drug-centred approachto psychiatric dmg treatment may not have been totallyabandoned with the psychophannacologic.al revolutionofthe 1950s. Psychotropic drug handbooks list numerOll5different. classes ofdmgs as appropriate for a given condi­tion. Nearly every single class of psychotropics, for ex-

150 PSl'ch"t.her Psychosom 2005;74: 145-153 Moncrieff/Cohen

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

ample, has some indication for the treatment ofpsychosis[65]. Benzodiazepines are widely used to treat acute andchronic psychosis and other behavioural disturbances[66]. Clinicians may thus appear to practise something ofa drug-centred approach while alleging a disease--centredrationale for prescribing. However, on its mvn terms, thedrug-centred approach appears to deserve more seriousengagement by practitioners and researchers.

A drug-centred model would require a tmnsformationin clinicians' explicit approach to the management ofpsy­chiatric conditions. For example, sedation or tranquiliza­tion could become an. explicit (short-term) treatmentstrat.egy, not merely in emergency situations, but for theduration of acute psychotic episodes or situations wheresomeone was experiencing severe agit.ation. This ap­proach need not induce pessimism about possibilities forclinical improvement. As understood by many psychia­trists treating schizophren ic patients in the 1950s and1960s, t.he tranquilizat.ion theory supposed that patientshallucinated less. were less t(~nse, or manifested less pres­sured speech precisely because theywere tranquilized [20,pp. 68-69]. Sedation does not preclude diverse effects onsymptoms, simply a less specific method of short-termpharmacological action.

Psychiatric treatment based on a drug-c,entred modelof drug action has the potential to be a more democraticand user-oriented activity than it currently is. It involvesdetermining whether the effects of different drugs haveutility to lessen the distress associated with various dys­functional emotional and behaviouml states. This re­quires an equitable dialogue between psychiatric serviceusers and professionals, with both parties sharing theirdifferent knowledge and perceptions ofdrugs' action andpotential utility [67, 68]. However, for consensual psy­chiatry, it is ultimately the user's experience that deter­mines a drug's utility and value.

A psychiatry that squarely incorporated the drug-ccn­tred model would focus strongly on the balance betweenthe pros and cons of using drugs in different situations.What this would mean in practice will require substantialelaboration, but can be jJlustratcd by considering themanagement of acute psychosis and depression. The in­duction ofsedation, indifference and akinesia by antipsy­chotic drugs may be perceived as useful in acute psycho­sis, by patients or by others. However, in the long-term,such effects are unlikely to be conducive to a return tononnal functioning. In addition, the reality that takingmost antipsychotic drugs is so often an aversive experi­ence means that even if these dntgs appear uniquely use-­ful at suppressing acute psychotic symptoms, people

might opt for other sorts of treatments given the choice.In the management of depression, some people may fll1dsedative dntgs or stimulants useful in the short tem1.However, these strategies need to be set against possiblenegative effects, such as hindering or delaying the processof self-directed recovery.

It remains to be determined whether a drug-centrcdmodel of drug action, by providing a more balanced viewof the benefits ofdrug treatment but also by creating newways to promote drug usages. would lead to a reductionor an increase in the usc ofprescribed psychotropic drugs.Whatever the outcome, it would demand that mentalhealth professionals become bettcr infomled about thenature of effects of different psychotropic drugs in orderto enter into a constmctive dialogue about their genuineutility with consumers. Similarly, it would require thedevelopment of new directions for psychopharmacologyresearch, much ofwhich would also need to be conductedin collaboration. with consumers. In view of the limita­tions of the disease-centred model outlined in this paper,a drug-centred approach deserves further exploration andelaboration.

Acknowledgement

J.M. ;Icknowledges financial support from North 1-:'ISl LondonMent<ll Health Trust. D.C. acknowledges iinancial support fT<lmlhe College ofHealth and Urban AtTairs, Floridalnl.emationalUni­ve~ity.

Model, ofDmg Action Psychother Psychosom 2005;74: 145- i ~3 151

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

References

Royal College of Psychiatrj~t:;: Antidepres­sallt$ facll; sheet. Lond'Jn, Royal College ofPsy<'11iatriSlS,2002.

2 Val~nstein E: Blaming the br,dn: The 'ntlhabout drugs and mental health. New York,Harper, 1998.

3 Kapur S: Psychosis as a stMe of aberrant sa­lience: A framework linking bie-kogy, phenom­enology and phannacology in schizophrenia.Am J Psychiatry 2003; 160: 13-23.

4 Rang HP, Dale MM., Ritter JM: Pharmacolv­gy, ed 5. Edinburgh, Churchill Livingstone,2003,HYJllan SE, Nestler ill: Initiation and adapta­tior,: A par~digm for understandingpsy<'110tro­pic drug action. Am J Psychiatry 1996;153:t51-162.

6 Hyrllan SE, Ne611er EJ: Drs Hyman & Nestlerreply. Am J Psychiatry I'HI; ',54:440--441.Breggin PR: Brain di~abJingtLr.atments in psy­chiatry: Drugs,electroshlXh. aHd lhero1eoftheFDA. New York, Springer. 1997.Healy [): The antidepressant e·ra. Ne'" York,Harvard University Press, 1999.

9 Healy D: The creation of psychopharmaco10·gy. New York, Harvard University Press,2002.

10 Kaiser D: Against biologi,~al psychiatry. Psy­chiatric Times 1996: 12. URL: http://www.psy­chiatrictimcs.colll/p96 I 242.h!1ll1.

II Ross CA, Pam A: Pseudoscience in biologicalPSychiatry: Blaming the bod)'. New York,Wiley, 1995.

12 Sullivan HS: Schizophrenia 'IS a human pm­cess. New York, Norton, 19G·~.

13 Lehmann H: The introduction ofchlorproma­zine to North America. I'sychiatr J Univ Ott1989: 14;263-265.

1.4 Deniker P: Experimental nC\lrological. syn­dromes and the new drug therapies in psychia­try. Compr Psychiatry 1961; I :92··100.

15 MalitLS, Hoeh PH, less.: S: A two-yearcvalu­ation of chlorpl'Omazine in clini,'al ft~arch

and practice. Am J Psychialry r956;113:540­548.

16 Winkelman NW Jr: An appraisal of chlor­proma:d.lle: GeJleml principles for administra­tion ofchlorproma1.ine based on experience of1090 patients. Am J Psychiatry [957: 114: 161­168.

17 S,lrgam W, Slater E: An introduction to physi­cal mel hods of treatment in psychiatry. Edin­burgh, Livingstone, 1944.

18 Mayer-Gross W, Slater E, Roth M: (1inic.~l

psychiatry, ed I. London, Cassel & Co, 1954.19 Moncrieff J: An investigation into the pre.:e­

dents ofmodem drug treatment in psychiatry.Hist Psychiatry 1999;10:475-490.

20 Gelman S: M.edicating SChizophrcnia: A hL-rIllry. New Brunswick, Rutgers UniversityPres!;, 1999.

21 Shorte.r E: A history of psych ialry: From theag<1 of the asylum t,) tll<1 "g<1 of I'r',".a,'. NewYork, Wiley, 1997.

22 Klerman GL: The evolution of a scientific no­sology; in Shershow JC (ed): SChizophrenia:TheOl)' and Practice. Cambridge, HarvardUniversity Press, J978.

23 Van Rossum JM: The signiiicance of dopa­mine blockade for the llK'Chanisrn of action ofneurolepticdrllf.~.Arch Int Pharmacooyn Ther1966; 160:492-494.

24 Salzman C, Miyawaki EK. Ie Bars P, KerrihardTN: Neurobiologie basis of anxiely and itstreatment. Harv Rev Psychiatry 1993;1:197­206.

15 Suranyi-eadotte BE, BodnolT SR, Welner SA:Antidepresslltll·arudolylic interactions: In­voh'ement of the ben1.odiazepinl>-GABA andserotonin systems. Pmg Neuropsychopharma­col BioI Psychiatry 1990; 14;633-654.

26 Soares JC, Innis RB: Neurochemical brain im­aging investigations ofschi7.Qphrellia. Bioi Psy­chiatry 1999;46:600-615.

27 Khan AU: How do psychotropic medicationsreally work? Psychiatr Timr8 1999; 16 (10).URi.: hltp;/lwww.mhsource.comieduJpsy­timeSlp991012.htmL

28 Casey JF, Bennett IF, Lindley CJ, Hollister LE,Gordon MH, Splinger NN: DlUg therapy inschizophrenia. Arch Ger. Psychiatry 1960;2:210-220.

29 Casey IF, Lasky JL, Klett CJ, Ht,llCiter LE:Treatment of schizophrenia reactions withphenothiazine derivatives. Am Jl'sychiatry1960; 117:97-105.

30 Abse DW, Dahlstrom WG, Tolley AG: Evalu­ation of tranq uilising drugs in the managementof ~cllte mental disturb~nce.Am J PsychiatryJ960; 116:973-980.

31 Wolkowitz OM, Pickar D: Benwdiazepines inthe treatment of schizophrenia. Am J Psychia·try 1991;148:714-726.

32 Carpenter WT, Buchanan RW, Kirlcpatrick B,Breier AF: Diazepam treatment of early signsof exacerbatioo in SChizophrenia. Am J P~'Y­

Chiatry 1999; 156;299-303.33 MoncrieffJ: Areantidepresr.antsover-rated? A

review of methodological problems with anti·depressant trials. J Nerv Ment Dis 2001; 189:288-295.

34 Robertson MM, Trimble MR: Major tran­qui1isers used as antidepressanl~. A review. JAU'cct Dis 1982;4:173-193.

35 Greenberg RP, Fisller S; Suspended judgo­ment. Seeing through the double masked de­sign: A commentary. Control Gin Trials 1994;15:244-246.

36 Prien RF, Caffey EM, Klett 0: Comparison oflithium carbonate and chlorpromazine in thetreatment of mania. Arch Gen Psychialry1972;26: 146-153,

37 Garfmkd PE, Stancer He, l'efSfld h: A com­parL~on of haloperidol, lilhium carbonate andtheiroombinalion in the treatment ofmania. JAI1~t Dis 1980;2:279-288.

38 Braden W, Fink. EB, Qualls CI:\, Ho CK, S~m­uels WOo Lithium and chlorpromazine in psy­chotic inpatients. Psychiatr Res 1982;7:69­81.

39 Johnstone EC, Crow TJ, FrilhCD, OwensDG:The Northwick Park 'functional' psycho~is

study: Diag.nosi~ and treatment response. Lan­cet 1988;ii:1I9-J25.

40 Moncrieff J: Lilhium revisited. A re-e.umina­til'll of the placeb<roontrolled tdals of lithiumprophylaAis in m~nic·depressive disorder. Br JPsychiatry 1995; 167:569-573.

41 Goodwin GM: Re.:urrence ofmania after lith­ium \\~Ihdrawal. Br J Psychiatry 1994; 164:149-152.

42 Perlis RH. Sachs GS, Lnfer B, Otto MW, Faru­one SV, Kane JM, Rose.r.,baum JF: Efiet1 ofabrupt ehange fwm frtandard to low serum Iev­eb oflithium: Reanalysis of double-blind lith­iwn maintenance dala. Am J Psychiatry 2002;159:1155-1159.

43 DavisJM,Janieak I'G, Hogan DM: Mood sta­bilisers in the prevention of recurrent affectivedisorders: A meta-analysis. Acta PsychiatrScand 1999;100:406-417.

44 Cavanagh J, Smyth R, Good",in GM: Relapseinto mania ordepression fonowinglithium di&­continuation: A 7-year follow-up. Acta Psychi­atr Scand 2004; 109:91-95.

4S Ramaekers JG, Louwerens JW, MuntjewerlrNT>, Milius H, de Hie A, Rosenzweig P, et a1:Psychomotor, cognitive, extrapyramidal. andaffective functions of healthy volunteers dur­ing treatment with an atypical (amisulpiride)and aclassic(haloperidol} antipsychotic. J C1inPsychopharmacol 1999;19:209-221.

46 McClelland GR, Cooper SM, Pilgrim AI: Acomparison of the central nervOlls system e(~

fe.:ts of haloperidol, ehlorprom87.ine and sul­piride in uorma1 volunteers. Br J Oill Phamu1­col 1990;30:795-803.

47 Rammsayer T, Gallhofel' B: Remoxipride ver­sus haloperidol iu healthy volunteers: Psycho­metric performance and subje.:tivc toleranceprofiles. Int C1in Psychophannacol 1995; 10:31-37.

48 Healy 0, Farquhar G: Immediate effects ofdroperidol. Hum Psychopharmacol 1998; I :~:

113-120.49 Cohen D: A critique of Ihe usc of ne.urolcplk'

drugs in psychiatry; in Fisher S, Greenberg RG(ed5): From placebo to panacea. New York,Wiley, 1997, pp 173-229.

50 Kropf 0, MlIller-Oerlinghauscn B: Change'S ink.aming, memOI), and mood during lithiwutreatment. Approach to a research strategy.Acta Psychiatr Scand 1979;59:97-124.

51 Judd LL, Hubbard B, JallowsI..-y OS, Hue)' LY,Attewell PA: 'The effect oflithium carbonateon offect, mwd. ond personality of normalsubjects. Arch Om Psychiatry 1977;34:46­51.

52 MuJ.ler-Oerlin~hallsen B, H~mann S, Herr­mann WM, KropfD: Effects oflithium ou vig­ilance, psychomotoric penormanceand mood.Phamlakopsychiatr Neurop~ychophannakol

1979;12:388-396.

152 Psychother p:;ychnsom 2005;74: 145-153 MoncriefflCohen

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

53 Tr:mt~r R. Healy H. Cattell D. Healy D: Func­tional etTects of agents differentially select in'to noradrenergic or scrotonergic systems. Psy­chol Moo 2()(j2;32:5 17-524.

54 Hermann WM, McDonald RI: A multidimen­510nal lest approach lor Ihe description of theeNS nctivityofdrugs in humnn pharmilCology.Pharmakopsychiatr Neurnpsychopharmakol1978;11:247-265.

55 Kirsch I, Moore n, Sroborill A, Nicholls S5:nle empel'Qr.'s new drugs: An analysis ofanti­depressant medication data submitted to theU.S. Food and Drug Adminc.1ration. PrevTre~t 2002;5(artide 23). URL: http:;/w",,,,.jOllmals.apa.org.!prevention/vohl1ne5/pre()050023a.html.

56 Willner P: B~ha~'iouJ"dl models in psychophar­macology; in Willner P (cd): Behavioural mod­els ill psychopharmacology: Theoretical, in­du;,(rial and clinical perspectives. Cambridge,Cambridge University Press, 1991. pp 3-18.

57 Weiss 1M, Kills CD: Animal mDdelsofdcpres­sion and 'lChizophren ia; in Schatzberg AF,NemeroffCB (eds): The American PsychiatricPress Te.~lbook of Psychopharmacology, ed 2.Washington, American Psychiatric Press,1998, pp 89-132.

58 Bourin M, Fiocco AJ, Clenet F: How valuableare animal models in delining antidepressantaC.liyily? Hum Psychopharmacol Gin Exp1001; 16:9-21.

59 farm A: Acommon role for psychotropic med­ications: Memory impairment. MooHyp 2003;60: 133-142.

60 CryWl. IF, Markou A, Lucki I: AsSC5Sillg 8uti­depressant activity in rodent~: Recent develop­ments and future needs. Trend Pharmacol Sci2002;23:238-245.

61 Iacob~ D, Cohen D: Whal is n~ally knownabout p:.ychological alterations produced bypsychiatric drugs? Int I Risk Safety Med 1999;12:37-47.

62 rava GA, Ruini C, Rafanelli C: PsychometricIheory is an obstacle to the progress of clinicalresearch. Psychother Psychosom 2004: 73:145-148.

63 Fava M. Evans AE, Dorer DI. Schoenfeld DA:The problem vfthc placdJo 1~5ponse in clinicaltrials for psychiatrc. disorders: Culprits, possi­ble remedies and a novel study design ar>­prv:l<:h. Psychother Psychosom 2003;72: 115­127.

64 Demyttena,~re K, De fruyl]: Getting what youask for: On the selectivity ofdepression ratingscales. Psy<:h0lher l'sychosom 2003;72:61­70.

65 Bzechlyhnik-Bulter K, Ieffries IJ: Clinic.11handhook of psychotropic drugs, ed 12, re­vised. Seattle, H'.Jgrele & Huber, 2002.

66 Palon C, Hanham S, Whilmore J: Bcnzodiaz­epines in schi,.ophrenia. Is thel~ a trend to­wards long-term prescribing? Psychiatr Buli2000;24: 113.. ! 15.

67 Coht'n D, McC"bbin M, Collin I, PerodC8U G:Medications HS social phenomena. Health2001;5:61-77.

68 Cohen D: The psychiatric medication hislOrj:Context, meaning, and purpose. Soc WorkMenl Kefllt.h 2003; 1:5-28.

Models of Drug Action Psychother Psychosom 2()()5;74: 145- i 5.~ 153

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.