What Do Psychiatric Drugs Do?  Further Reflections on Methodology, Sources of Information, and Meaning

David H. Jacobs, Ph.D.

A fundamental distinction must be made between a drug’s effects per se and a drug’s intended therapeutic effects.  It is only the existence of intended drug effects that makes it possible to speak of side or adverse drug effects.  In medicinal drug research the focus is on the drug’s effects with respect to the condition being treated.  In this context it is never the research task to broadly or comprehensively establish the drug’s effects per se.  In psychopharmacotherapy research the aim is specific symptom reduction or alleviation.  The question of the drug’s impact on mental life broadly considered is not investigated.  In fact, the drug’s impact on mental life broadly considered is never systematically addressed.  The meaning of drug-induced symptom reduction does not speak for itself.  It is not clear on the basis of focused efficacy research just what has happened to the subject, including whether what has happened is or is not salutary.  The present paper addresses this issue.

Two very different answers can be given to the following question: When treatment with a psychiatric drug does result in symptom alleviation, just what has the drug brought about to result in symptom alleviation?  What, in short, has happened?  The favored model in psychiatry is that the drug’s pharmacological activity has ameliorated a pre-existing brain lesion (e.g., “Prozac ameliorates a pre-existing serotonergic deficit” answers the question “Why does treatment with Prozac lift depression?”  Healy, 1999, provides a cogent discussion of the now-dominant lesion model in psychiatry).  A much less favored answer ­­­­­­¾ but actually not that hard to find if one is prepared to look and perhaps read between the lines a bit ­¾ is that the drug’s neurotoxic activity has interfered with the patient’s pre-drugged self-awareness and communications (e.g., is a sedated grief-stricken person less grief-stricken?).  In the favored model emotional pain and so forth is a sign of neuropathology (in the same sense that a seizure is a sign of neuropathology), and drug-induced symptom reduction is desirable, therapeutic, beneficial, salutary, etc.  In the less favored model emotional pain and so forth does not by itself indicate neuropathology[1], and in the absence of evidence of genuine cerebral dysfunction drug-induced symptom relief is seen as a likely indication of iatrogenic brain injury.  I do not think that a neutral candidate answer ¾ some linguistic equivalent of don’t look a gift horse in the mouth ¾ is a responsible position to take, since (as I will argue) the possibility that drug-induced symptom reduction betokens brain injury must be taken seriously.  A psychological understanding of a person’s emotional pain and so on ¾ in contrast to a brain lesion model ¾ leads to a real sense of puzzlement when psychiatric medication brings about reduced distress.  What has happened?  How could the drug’s activity have altered the person’s own view of his/her situation and attendant feelings and behaviors?

            An unusually illuminating presentation of the conceptual, linguistic, and empirical ambiguities involved in depicting “antipsychotic” drug effects as therapeutic or injurious is provided in a 1987 paper by Van Putten and Marder.  The paper begins by noting that neuroleptic-induced akinesia[2] can be difficult to recognize and is often undiagnosed (by the prescribers).  The ambiguities I alluded to can be seen in the following passage:

In the treatment of schizophrenic illness with neuroleptics, akinesia, in our experience produces a type of improvement.  The patient talks less of psychotic material, but he talks less of everything; he is less bothered by his hallucinations, but he is less bothered by everything else as well; he is less invested in his delusions, but he is less invested in all else as well.  Many patients with akinesia experience a peculiar absence of emotions, appear emotionally dead, and often state that everything is all right.  This type of improvement, which occurs particularly with higher doses, is one that we should not be proud of. (p.15)

Van Putten and Marder then cite a prospective study of their own involving newly admitted schizophrenic patients treated with haloperidal.  Ratings of these patients (presumably by others) on the Brief Psychiatric Rating Scale (BPRS) schizophrenia factor were inversely correlated with ratings of akinesia (presumably by Van Putten and Marder; r =  -.39, p = .004, pooled within-subject correlations).  In other words, the more the “antipsychotic” drug created brain injury and its behavioral consequences, the less psychosis was seen by those filling out the BPRS.  Notwithstanding the prior statement (quoted above) to the effect that neuroleptic-induced akinesia is not a treatment effect to be proud of, Van Putten and Marder only two paragraphs later state that when intractable positive symptoms are distressing to the patient or to others (italics added) “an akinetic dampening of the entire mental life is desirable” (p. 16).  This statement illustrates the shifting meaning and ambiguity of “therapeutic drug effects” in psychiatry.  The authors do not directly address whether “antipsychotic” drugs actually have antipsychotic properties in the literal sense, that is, separate from the affective, cognitive, and behavioral consequences of akinesia.  They are clear that it is not difficult for psychiatrists to regard neuroleptic-induced akinesia as improvement, therapeutic, beneficial, etc.  The leeway permitted by considering drug effects therapeutic if the distress of others is reduced should not be overlooked.[3]

            Van Puten and Marder raise two issues I intent to develop:  (i) adverse psychological drug effects may be difficult to detect precisely because the drugged person is a patient who is markedly distressed/disturbed (e.g., is akinesia present?  Is it an adverse effect or the desired effect?), (ii) “improvement” or “symptom reduction” does not necessarily mean that the treatment drug has induced a normal mental state, or a mental state that is any more normal than the pre-drugged condition (e.g., neuroleptic-induced akinesia is not a normal mental state.)

            A thoughtful reader of Van Putten and Marder’s paper is likely to wonder why non-recognition of neuroleptic-induced akinesia is common.  Van Putten and Marder emphasize that “the onus of detection is on the physician, for the patient seldom complains.  Indeed, even when asked … the patient often denies any difficulty and is seemingly locked in a peaceful, apathetic remoteness” (p. 16).  A major theme of the present paper is that psychopharmacotherapy researchers make light of the onus of detecting neurotoxic drug effects.  Note that Van Putten and Marder’s observations were made in a clinical context as opposed to a drug treatment efficacy research context.[4]  The quote presented directly above suggests a certain hands-on familiarity with the patients being treated with neuroleptics.  An earlier paper by the same authors (Van Putten et al., 1981) which focused on patients’ subjective reactions to thiothixene described a therapeutic setting wherein newly admitted schizophrenic patients (to the Veterans Administration Medical Center in Los Angeles) were provided “active milieu” treatment for seven to 14 days before being tried on thiothixene if they failed to improve.[5]  The familiarity with patients as individuals suggested by active milieu treatment stands in sharp contrast to the de facto anonymity of subjects as distinct personalities in conventional drug treatment efficacy research.  As long ago as 1960 F.A. Freyhan, a pioneer in psychopharmacotherapy, warned that the large scale (many research subjects) “check list” approach to psychopharmacotherapy research could not avoid producing misleading information about the effects of psychoactive medications:  “ … experimental investigations cannot be substituted for clinical explorations.  It must remain the clinician’s task to find out what it is that the drug does to whom” (p. 197).  As I have discussed in detail elsewhere (Jacobs, 1999; Jacobs and Cohen, 1999), treatment efficacy research is not designed for the purpose of ascertaining what the drug did to whom.  The focus of such research is very narrow ¾ was the investigational drug superior to placebo in terms of a condition-to-be-treated-relevant research instrument (like the Hamilton Rating Scale for Depression, or HAM-D)?  The corresponding side effect detection component of such research is quite superficial (I will develop this point further). 

            From the beginning of the modern era of psychopharmacotherapy (the era ushered in by chlorpromazine in the 1950s) psychiatry has regarded what is treated by drugs and the effects/efficacy of drugs in terms of conventional somatic medicine.[6]  This is the basic “villain of the piece” from which all problems emanate.  From within the paradigm of conventional somatic medicine ambiguities about what is disease and what is “host”[7] are pushed aside along with specifically psychological and psychosocial concerns regarding the full cost or burden to the host of drug-induced symptom reduction.  Only recently and haltingly has conventional somatic medicine in general turned toward the question of how to assess the short and long term consequences of medical drugs on complex cognitive functions and complex behavioral competencies ¾ a question that is immediately confronted by markedly underdeveloped methods of investigation and assessment (Streufert and Gengo, 1993).  Meanwhile ¾ and in stark contrast to somatic medicine ­¾ the disease-host distinction in psychiatry faces extraordinary conceptual and empirical difficulties, if not absurdities[8] (Kleinman, 1988; Widiger and Shea, 1991).  The lesion model both renders pharmacological treatment “medically necessary” and deflects attention from what would otherwise stand out as the obvious working assumption regarding long-term exposure to chemicals which act directly on the brain, namely the assumption that long-term exposure must have serious deleterious consequences (Summerfield, 1978)[9].

The Side Effect Detection Component of Psychopharmacotherapy Research from the Perspective of Clinical Neurological Experience with Frontal Lobe Injury

I will now begin to develop the point that the “side effect” detection component of psychopharmacotherapy research cannot bear examination from the perspective of clinical neurology.  It is first of all necessary to outline relevant features of the conventional randomized placebo controlled clinical trial (RCT) in psychopharmacotherapy research:

            1.  The usual time frame for controlled psychopharmacotherapy research ranges from four to eight weeks (the mode seems to be six weeks).  I include this feature in the present list because psychiatric medications are intended for long-term use and are in fact used for long periods in clinical practice (years, decades), so it is prima facie highly unrealistic to conduct safety and efficacy research over a four, six, or eight week period.  Experience with “tardive” (late appearing, as in tardive dyskinesia) pathological consequences of psychiatric medications over the past forty-five years should have ended such short-term research long ago.  An extensive reading of adverse drug reaction reports (ADRRs) published in psychiatry journals makes it clear that individuals vary greatly in the time course of vulnerability to the neurotoxic effects of psychiatric drugs.  The point is that terminating observations after four, six, or eight weeks does not realistically address what effects the drug may have as it is actually used in clinical practice.

            2.  The researchers and the subject are essentially strangers when drug treatment initiation begins.  At best contact up to the point of drug treatment initiation is limited to screening and diagnosis, itself a formal question and answer affair.  This arrangement dictates that the researcher is likely to only notice gross or blatant drug effects (marked tremor, myoclonic jerking, pronounced sweating, slurred speech, etc.)

            3.  The researcher and the subject only meet together for the purpose of brief and highly structured sessions devoted to drug-relevant symptoms and side effects review.  This restricted interpersonal/diaglogical format (narrowly focused, closed-ended questions) essentially eliminates whatever skill or acumen the interviewer might possess as a clinician in terms of detecting somewhat subtle forms of “behavioral toxicity.”  It cannot be overemphasized that the side effect check list that the researcher goes over with the subject is limited to blatant, mainly somatic adverse drug effects about which the subject can self-report (nausea, diarrhea, vision problems, sleeping problems, etc).

            4.  At least in so far as recording and analyzing data for publication is concerned, the subject’s own depictions of drug effects beyond the (mainly somatic) side effects questionnaire created by the researchers is treated as non-information.

            5.  Only the subject’s self-report on the drug is regarded as information about drug effects.  This at one stroke relegates the entire issue of drug-induced anosognosia (impaired self-awareness/self-monitoring) a non-issue.  Yet it could be argued (see ahead) that drug-induced anosognosia is the core methodological challenge facing psychopharmacotherapy research.

            6.  No attempt is made ¾ no channel for such information is created ¾ to receive reports from people in the subject’s “natural environment” who do know him/her well concerning alterations in the subject’s cognitive performances and social behavior.  This throws the entire adverse/neurotoxic drug reaction detection effort onto the brief weekly meetings between researcher and subject (a highly restricted linguistic, interactive, and social episode).

            7.  No effort is made to present neurotoxicity/neuropathology-sensitive tasks or challenges to the subject during the course of the clinical trial.  This again throws the entire ADR detection effort onto what can be gleaned in a brief and highly restricted interview.  Since the issue of detecting somewhat subtle forms of drug-induced cerebral dysfunction is simply not acknowledged in controlled psychopharmacotherapy research, this area of research has contributed nothing to the problem of detecting chemically-induced cerebral pathology before gross signs of illness emerge[10].

The design of controlled psychopharmacotherapy research does not address itself to detecting somewhat subtle forms of cerebral pathology.  Clinician-initiated reports “from the field” of drug-induced behavioral toxicity do appear in psychiatric journals (mainly in the form of brief Letters to the Editor), but the dominant view of such reports from the perspective of psychopharmacotherapy researchers is that their validity is unknown because they do not derive from controlled studies (e.g., Gram, 1994; I think it is apparent that we are speaking here ¾ referring to controlled psychopharmacotherapy research ¾ of a closed circle).  What is missing both from ADRRs “from the field” and from dismissals of such reports on the part of researchers is a realistic discussion of the built-in insensitivity of the side effect detection component of controlled research.  Such a discussion would have to address the limitations I outlined above.

The paucity (and thus diagnostic invalidity) of what can be gleaned even by an experienced clinical neurologist under restricted observational conditions is clearly depicted by Stuss (1991) in a review chapter on frontal lobe injury.  Stuss attempts to characterize the deficits/impairments which typically arise specifically from frontal lobe injury on the basis of a review of relevant clinical literature and his own extensive clinical experience.  An important source of information concerning specific frontal injury derives from clinical studies of post-lobotomized patients, since in this procedure the aim is to only make incisions in the frontal region.  The application of this procedure in the 1940s and 1950s not only to a wide spectrum of psychiatric conditions but also in the treatment of non-psychiatric conditions (notably chronic pain) greatly facilitates the task of drawing conclusions[11].  Stuss also reviews endogenous neurological conditions and post traumatic (that is, mechanical trauma) neurological injuries to the frontal region.

The disparity between what can be detected by a clinical neurologist during office visits (not detected, that is) and what is dramatically obvious to people who know the patient well in his/her everyday life is illustrated by reviewing the case of R., a patient Stuss personally treated.  R. is described as a highly educated and successful businesswoman who first came to neurological attention due to seizure activity.  A right frontal astrocytoma was surgically removed, followed by irradiation and chemotherapy to eliminate tumor growth.  R. resumed work one year post-surgery, but some time after that returned for neuropsychological assessment at the insistence of her husband.  R. herself initially denied having any problems (this is called anosognosia in clinical neurology).  Neuropsychological examination revealed excellent abilities, even on “frontal” tests (Porteus maze, Wisconsin card sorting, etc.).  If the assessment question had rested only on R.’s test performances and observable conduct in Stuss’ office there would have been little or no reason to suppose that R. had become seriously impaired.  Yet reports concerning R.’s behavior in “natural” settings conveyed to Stuss by her husband and others indicated a very different picture of her overall mental status.

For example, under her own direction it was difficult for R. to get to work before 11 A.M.  This is only one concrete example of a general deficit which Stuss refers to as a disturbance in self-awareness and cognate expressions (self-consciousness, self-reflectiveness, etc.).  There seems to be general agreement that frontal damage results in a disturbance of higher-order “executive” capabilities.  This form of impairment may be essentially invisible unless the person is adequately challenged in the appropriate realm of mental life (recall R.’s excellent performance on the neuropsychological assessment battery).  Many situations in life do not challenge the individual in such a manner, and in such situations the frontally damaged person may appear intact.  Stuss in fact comments that “in many respects R. was totally normal or unimpaired” (p. 77).  But obviously this is not the end of the story, and indeed it could be said that the behavior of frontally damaged persons serves to expose how much of everyday life is not structured by others who are in addition continuously present to provide surrogate self-monitoring.

Fisher (1989) has commented that “unawareness accompanies so many neurological defects that one might invoke anosognosia as a broad principle of cerebral dysfunction in human beings” (p. 127).  Stuss wishes to distinguish frontally-produced unawareness (anosognosia) from anosognosia that accompanies damage to other areas of the brain (e.g., unawareness of paralysis).  I am inclined to think that frontal anosognosia is of special interest to psychiatry.  Stuss specifically states in his concluding remarks that “disorders of self-awareness can be present with normal sensorium and normal of even superior IQ, memory, and so on” (p. 78).  My own chief concern, to recall, is the problem of understanding what has happened when psychiatric medication brings about symptom-reduction (e.g., Prozac-induced relief from depression).

An especially revealing clue to understanding the effects of frontal injury is provided by the employment of procedures (surgery and ECT) which damage the brain for the express purpose of treating otherwise intractable pain associated with various medical conditions (neuralgia, carcinoma, phantom pain, etc.).  Deliberate physician-produced brain damage for pain treatment  is mentioned in Stuss’ (1991) review and also in a chapter on the frontal lobes by Brown (1988), but neither author gives this phenomenon the attention I think it deserves.  Brown (1988) comments that lobotomy and lobectomy used for pain treatment bring about an altered emotional reaction to pain rather than in an alteration of the pain threshold.  It is interesting to try to reconcile the treatment of pain via deliberate frontal damage with the usual generalization that frontal damage results in enhanced “stimulus boundness,” that is, inability to detach from present stimulation.  The apparent contradiction can be conceptually reconciled by focusing on the consequences of disturbed self-awareness for emotional life.  Stuss notes that prominent advocates of lobotomy in the 1940s and 1950s came to think that lobotomy had the effect of dramatically reducing self-concern because it impaired the patient’s capacity to apprehend him/herself historically, that is as simultaneously having a past, present, and future.  From this perspective it follows that the patient’s emotional reaction to pain is altered because the patient’s capacity to grasp the significance of the pain as imbued with implications for the personal future is lacking or substantially reduced.

ECT, Anosognosia, and Symptom Improvement

            The deliberate induction of anosognosia as psychiatric improvement was discussed openly for a period of at least two decades by Max Fink, one of American psychiatry’s most prominent advocates of ECT.  Fink’s insight concerning anosognosia was undoubtedly fostered by his close association with Edwin Weinstein and Robert Kahn, the authors of what has come to be regarded as a seminal work in clinical neurology on the subject of anosognosia (Weinstein and Kahn, 1955).  In their 1955 book Weinstein and Kahn reviewed their experience with ECT-induced anosognosia as a treatment for intractable pain, but beyond this they did not discuss possible implications of physician-created anosognosia for clinical psychiatry.  However, by 1956 Kahn, Fink, and Weinstein jointly published a paper on their clinical experience using ECT in the treatment of strictly psychiatric conditions (depression, schizophrenia, and mania)[12].

In the 1956 paper Kahn, Fink, and Weinstein specifically discuss “improvement” in terms of the broadened concept of anosognosia advanced by Weinstein and Kahn in 1955, i.e., “A concept of anosognosia was advanced which included not only denial of hemiplegia [the paradigmatic case of anosognosia introduced by Babinki in 1914] and blindness but denial of many other aspects of illness and problems of living” (p. 23, italics added).  In their series of 24 patients treated with ECT, only those who were injured enough by the procedure to become anosognosic were regarded by hospital staff as improved:  “If the patient denies he has any problems or that he is troubled by them, or if he cannot recall any, he is rated as improved.  Such patients appear affable and uncomplaining, their manner reinforced by cliches and banalities, themselves adaptive [to brain injury] forms of language” (p. 28).  The authors were equally clear that ECT-induced improvement was only temporary because patients recovered from the brain injury produced by shock treatments (the injury-induced anosognosia remitted, in other words).  This paper, then, openly presented a model of  the second possible scenario I depicted in my introductory statement, that is, of emotional distress etc. not brought abut by neuropathology which is treated by iatrogenic brain injury.  Note that in this scenario neurological recovery brings about psychiatric relapse.  Kahn, Fink, and Weinstein had no illusions as to why this occurs ¾ when the anosognosia clears, the patient finds himself in the same intra and interpersonal predicament, and the treatment (ECT) has definitely not provided him with any additional psychological resources for living (all the patients in this report were men).

            In the course of their commentary on ECT-produced brain injury, anosognosia, and “improvement,” Kahn, Fink, and Weinstein in effect suggest a method for realistically evaluating the consequences of direct physical intervention in brain functioning.  First of all the patient must have sufficient opportunity prior to somatic treatment to discuss his/her personal difficulties (i.e., the patient’s own view of meaning, history, development, situational and interpersonal influences, and so forth).  Secondly, following a course of somatic treatment (or some segment) and “symptomatic improvement,” it is up to clinical researchers[13] to assess through dialogue with the patient and probably others as well the extent to which improvement represents enhanced psychological functioning vs. anosognosia (denial, indifference, inexplicable unconcern about past events and likely future developments, etc.).  The contemporary approach to diagnosis/assessment based on the neuro-lesion model ¾ that is, emotional pain etc. is simply the subjective component of neuropathology ¾ in effect denies the necessity to contextualize and interpret the meaning of symptom reduction (pain, distress, anguish, etc is bad, reduction is good)[14].

SSRI-Induced Frontal Lobe Injury, Anosognosia, and Symptomatic Improvement

            A recent half-hearted ADRR concerning fluvoxamine and fluoxetine in the treatment (respectively) of panic disorder and major depression illustrates many of the issues discussed above.  I characterize this ADRR as “half-hearted” because the authors, Hoehn-Saric et al. (1990), ultimately cannot make up their minds as to whether fluvoxamine and fluoxetine-induced frontal lobe dysfunction should be considered a neurotoxic effect or a therapeutic/salutary drug effect.  Thus they only reduced the dosage of fluoxetine in case number five of their report despite persistent behavioral symptoms of frontal lobe dysfunction because of concomitant improvement in mood.  Their final comment is worth quoting since it rather dramatically portrays the conceptual, linguistic, and evidentiary ambiguities concerning what psychiatric medication brings about that I have tried to bring into focus:

…apathy and indifference may be dose-related toxic effects that are unrelated to the anxiolytic or antidepressant effects of fluvoxamine or fluoxetine.  However, it is also conceivable that the induction of a mild form of indifference or disinhibition is a therapeutically important ingredient of serotonin reuptake blocking medications.  Such action could explain why imipramine, which has serotonin reuptake blocking effects, was more effective in patients with generalized anxiety disorder than alprazolam in attenuating the tendency to worry excessively and why antidepressants with serotonin reuptake blocking properties were found to be superior to other antidepressants in obsessive-compulsive patients. (p. 345)

            Hoehn-Saric et al. begin the paragraph which contains the above quote with the following statement:  “Not all patients on fluvoxamine or fluoxetine react in the manner described above.  Many recover from depression or anxiety disorder without significant side effects” (p. 345).  But the question must immediately be posed: How do they know?  Their own case summaries clearly state that they personally noticed nothing amiss about their patients during the course of office visits.  Their eventual diagnosis of drug-induced frontal lobe dysfunction was based entirely on patients’ reports of untoward developments as drug treatment continued over time.  Further, it was more the case than not that patients eventually pieced together that something was wrong from the reactions and feedback proffered by others rather than from their own self-monitoring/self-evaluation capacities.  For example, regarding case number one:  “Only gradually did he come to realize intellectually the extent of his neglect, but he still could not get upset about it” (p. 343).  When Hoehn-Saric et al. were informed of neglect, apathy, indifference, etc. by this patient, they reduced his dosage of fluvoxamine from 150 mg/day to 100 mg/day.  On the reduced dosage the patient no longer complained of frontal lobe dysfunction symptomatology, which the authors equate with “functioning well.”  But having raised the issue of frontal lobe dysfunction themselves (even citing Stuss and Benson’s 1986 text on the frontal lobes as support for their diagnosis of frontal lobe dysfunction), it should be clear to Hoehn-Saric et al. that the formula “no complaints = functioning well” simply will not do, precisely because the expectation of impaired self-monitoring/self-awareness vitiates relying so heavily on self-report.

            As is customary in ADRRs in psychiatry journals, Hoehn-Saric et al. express no curiosity about why numerous RCTs with fluvoxamine and fluoxetine (prior to and following FDA approval) have failed to detect indications of frontal lobe dysfunction, nor what might be done to render RCTs more sensitive to drug-induced frontal injury.  Nevertheless, it is apparent that both the time frame of RCTs (four, six, or eight weeks) and researcher-controlled restrictions on what subjects can say operate to eliminate the possibility that what Hoehn-Saric et al. detected in clinical practice would be detected in an RCT.  RCTs are drug treatment efficacy investigations, which in practice means comparing the short-term fate of certain (drug-relevant) symptoms in response to drug treatment or placebo treatment.  The purpose of a drug efficacy RCT is certainly not to investigate the effects of the drug on a broad range of cognitive and social-behavioral capacities and performances.  This is a different research question and enterprise.   Among the lay public there is probably an understandable assumption that safety claims concerning a psychiatric medication encompass the question of the drug’s effects on mental and social life broadly considered.  In fact only a narrow range of drug toxicity is investigated by developers of psychiatric medication leading up to FDA approval.  Following FDA approval the issue of efficacy is further explored, but the investigative approach to “side effects” detection does not change.

Strategies for Realistically Assessing the Psychological Effects of Psychiatric Drugs

            The clearest research strategy for realistically assessing the impact of a psychiatric drug on mental life is to systematically investigate its effects in therapeutic dosages and durations on normal volunteers.  The reason for this does not seem to be generally appreciated.  The psychopharmacotherapy pioneer who originated the term “target symptom,” Fritz Freyhan, can also be given credit for propagating a fundamental misunderstanding concerning how to realistically evaluate the impact of psychoactive drugs.  Freyhan (1960) argued reasonably enough that the clinical study of drug effects (that is, treatment) depends on the identification of “targets” on which the drug is presumed to exert its action and upon appropriate patient selection, e.g., a drug which is presumed to alleviate headache can only be assessed in patients who actually have headache.  In the very next sentence (p. 186), apparently to buttress his point, he remarks that an investigation of the antidepressant properties of amphetamine (as opposed to its appetite suppressant properties) cannot be carried out with an obese but emotionally harmonious patient because he/she would not experience much if any mood change.  This is a doubtful proposition, to say the least.  The mistake here is profound: amphetamine is not a specific medication for depression ¾ an “antidepressant” ¾ if in fact it alters mood regardless of whether the person is depressed or not (nor are amphetamines specific medication for “attention deficit” if they act to narrow and focus attention on relatively mindless tasks even for people who do not have difficulty paying attention  ¾ in this regard see the strangely neglected paper by Rappaport et al., 1978).  There is, then, quite a good reason for using normal volunteers to investigate the effects of a medication for headache: the drug may not even have analgesic properties, it may for example actually induce global indifference and unconcern, but it would be much more difficult to determine this if only people with headache were studied.  A general principle of drug effects investigation can be stated: It is more not less difficult to realistically assess a drug’s properties when the drug is used to treat a clinical problem, especially when the drug has psychoactive properties and is used to treat psychiatric problems.

            Since many psychiatric drugs are extremely unpleasant and/or dangerous to take, it is understandably a major challenge to induce non-clinical volunteers to take such drugs in therapeutic dosages and durations.  Physicians, medical students, and normal volunteers who have taken a single dose of a neuroleptic make it abundantly clear why it is probably not practicable to use normal volunteers to test therapeutic dosages and durations of neuroleptics (Anderson, Reker, and Cooper, 1981; Belmaker and Wald, 1977; Healy and Farquhar, 1998; Kendler, 1976).  As part of her dissertation research an anthropology student, Sue Estroff (1981), took a low clinical dosage (5 mg) of fluphenazine (Prolixin) for almost six weeks.  Actually the diary entries she made on Prolixin and which she presents in her 1981 book stop at four weeks.  Estroff does not explain this in her book, but it is reasonable to surmise from her diary entries that after four weeks on Prolixin she was no longer able to make diary entries, e.g., day ten:  “Having a hard time sitting down to do notes and take notes at report” (p. 104).  Her diary entry after 16 days on Prolixin reads in part:  “The feeling of self-containment grows …  As if there were a glass barrier between me and others – emotions subdued and flat if there” (p. 105).  Since Estroff was not being treated with Prolixin for schizophrenia or any other psychiatric condition, there is little ambiguity about whether the effects she reports should be considered toxic/adverse or not.  As I have proposed, greater ambiguity concerning how to depict drug effects is created when the drug is used clinically (is “emotions subdued and flat if there” a toxic or adverse effect when psychotic agitation is the “target”?).  This is why it is invaluable to obtain information about drug effect from non-clinical subjects (commenting on their own experience with a single 5 mg dose of haloperidol, two psychiatric researchers ¾ Belmaker and Wald [1977] ¾ remark as follows:  “Dopamine blocking by neuroleptics may function to restrict cognitive and emotional processes in normals as well as schizophrenics, and thus it is possible that it does not specifically antagonize schizophrenic pathology” [p. 222-223]).

            It is also of great importance (for the same reason) to carefully record the effects of somatic psychiatric interventions when they are used for non-psychiatric purposes (e.g., lobotomy and ECT for pain, discussed above).  Kushner (1999) has reviewed the use of haloperidol in the treatment of Tourette’s syndrome, which at least in America for the past forty years has been seen as a neuropathological disorder.  Kushner reports that the professional literature mainly focuses on the efficacy of haloperidol for suppressing involuntary vocalizations and movements (tics).  But people who have actually been treated with haloperidol and members of their families have told Kushner personally that “zombification” is the main effect of haloperidol from their perspective.  The implication is that haloperidol is not a specific “antipsychotic” drug but rather a drug which suppresses movement, emotion, and thought in many if not all persons who take it.

            Since it is impractical to obtain information concerning the effects of therapeutic dosages and durations of psychiatric drugs on people who are not being treated for a psychiatric diagnosis, it is a fortiori necessary to draw upon persons who have been psychiatrically treated in order to obtain broader information concerning the effects of these drugs than is provided by efficacy research.  One obvious method to advance what is known about the effects of psychiatric drugs is to provide people who have been treated with them an opportunity to talk about their own experience in their own words when they are not “on” the drug.  It is rather fascinating to contemplate just how alien and irrelevant such information seems from the perspective of efficacy RCTs[15].  The importance of off-drug depictions of drug effects is twofold:  (a) to avoid the enormously confounding issue of drug-induced anosognosia, although this throws the whole investigation into the realm of memory, and (b) to capitalize on perspective, that is, the person no longer on the treatment drug can reflect evaluatively on the pros and cons of the treatment he/she received[16].  The importance of encouraging people to talk about drug effects in their own words should be evident by now.  The reader of a published drug treatment efficacy RCT has literally no idea what the treated subjects thought about the drug’s effects (on the drug or later, off the drug).  An efficacy RCT already names and regards the treatment drug in terms of its desired effects ¾ e.g., an “antidepressant.”  A person who is very upset and dysphoric (drawing upon my own clinical experience) may welcome the emotional relief provided by what is called an antidepressant in psychiatry, but nonetheless depict the drug’s effects in words that have little if any obvious connection to how the drug is designated in psychiatry (by, for example, describing the drug’s effects as emotionally numbing, muting, flattening, suppressing, etc.).  A drug that is thought to literally counter or nullify depression implicitly creates one set of expectations concerning its impact on the person as a psycho-social being, while a drug that is depicted as numbing, flattening, etc. emotional life in general creates a distinctly different set of expectations.  The point here is that there is much to learn from drug-treated persons themselves about drug effects and consequences beyond what is provided by specific “target symptoms” research instruments used in treatment efficacy RCTs.

            I do not of course think that an ethnographic investigation which draws upon formerly drug treated psychiatric patients can by itself solve the problem of creating a realistic picture of the effects of psychiatric drugs on mental life (in my usage here a “psychiatric patient” is anyone who has been treated with one or more psychoactive drugs for a psychiatric diagnosis or mental-emotional-behavioral difficulties).  The problem is very complex and must be systematically “triangulated” (meaning that many methods of investigation yielding diverse forms of information must be generated in order to piece together a realistic picture).  An ethnographic inquiry with formerly treated patients does not, to be more specific, obtain information from individuals who have been in a favorable position to observe drug-induced changes in the treated person, nor does it obtain test performance information of any kind.  But, in line with an overall strategy of “triangulation,” encouraging formerly treated persons to reflect upon their experience with psychiatric medications can provide one conspicuously neglected source of information about the effects of psychiatric drugs on mental life.

Concluding Remarks

                The key observation that “symptom alleviation” brought about by the physician via direct action on the patient’s brain requires further investigation as to meaning was made by Kahn, Fink, and Weinstein in their 1956 report on the use of ECT in psychiatric treatment. Their sensitivity to this issue was a natural extension of Weinstein and Kahn’s seminal book concerning anosognosia as a common clinical phenomenon in many forms of neurological disease and injury which occurred both independently of the physician and due to the physician’s deliberate treatment efforts. It is true that Weinstein, Kahn, and Fink for their own part expressed no reluctance to bring about transitory symptom alleviation via brain injury-induced anosognosia, but all the same it was clear to them that this source of “clinical improvement” should not be confused with enhancement of the patient’s adaptive psychological resources (p.28).

            It is obvious that drug treatment research in psychiatry from the 1950s through to the present has declined to investigate the meaning of drug-produced symptom alleviation. How can this be understood? Clearly, market (that is, pecuniary) considerations play a prominent if not overriding role in the overall picture, but I do not wish to dwell on this point here except to note that The New England Journal of Medicine has recently concluded that the “sponsor’s” influence on published research findings in the realm of psychopharmacotherapy is such that it would be naïve to accept published reports at face value (Angell, 2000 ; Bodenheimer, 2000). I cannot see that anything of substance has changed since this view was expressed 18 months ago (as I write this paper) in America’s most prestigious medical journal. But I rather wish to reflect upon the consequences of regarding emotional distress and so on literally as medical disorders for the conduct of psychopharmacotherapy research.

            I will use as a text a well-known 1978 paper by Gerald Klerman, one of America’s most influential psychiatrists and policy makers from the 1970s until his death in 1992 (during the 1980s he served as head of NIMH and head of the Alcohol, Drug, and Mental Health Administration [ADAMHA], and was co-director of Upjohn’s multinational research project leading up to FDA approval of Xanax as the only FDA approved treatment for panic disorder). In this paper Klerman was keenly aware that the torch in American psychiatry was about to be passed from the Meyerian school of thought (the biopsychosocial view) to the self-named neoKraepelinians, or biopsychiatrists, a development in which Klerman himself was a major figure. He was therefore concerned to provide an explicit rendering of the neoKraepelinian framework of thought (he used the word “credo”) and to put the best face on it.

            Klerman introduced his rendition of the neoKraepelinian credo by noting that it rejects the Meyerian emphasis on “personal experience and the uniqueness of the individual in his social context” and replaces it with an emphasis on identifying the biological bases of discrete mental illnesses. It was – of course – difficult for Klerman to provide examples of discrete mental illnesses with known biological causes, but his choices convey the mind-set of biopsychiatry very well: syphilitic disease of the brain and aminoacidurias and chromosomal abnormalities which produce mental deficiencies. Such examples (which, it could be argued, are actually marginal to psychiatry) make it quite clear that neoKraepelinian thought rejects as irrelevant the view that “symptoms” arise from the workings of the personality as a whole as the individual strives to adapt to his/her slice of the social world and his/her own inner (psychological) life. In the 1980s Klerman advanced the same idea with regard to “panic disorder”  (1988; he seems to have recanted in a paper prepared shortly before his death in 1992, i.e., Shear, Cooper, Klerman, et al., 1993). The point that Klerman is making, namely that neoKraepelinians regard discrete diseases and the person as separate entities was elucidated even more clearly in a companion piece appearing in the same volume by Irwin Savodnik, in which Savodnik admitted that neoKraepelinians have no interest in the “host” other than as “a container for the disease process.” This view is entirely consistent with Klearman’s own exemplars of medical psychiatry (above).

            My hope is that elucidating what biopsychiatrists think about the nature of “mental illness” can help answer why psychopharmacotherapy research makes no effort to investigate the meaning of drug-induced symptom alleviation. Psychopharmacotherapy efficacy research is virtually always aimed at a specific Axis I “thing” (clinical entity) which,  it is assumed, arises from its own unique neuropathology (unfortunately not known with certainly), but which in any event is separable from the person who is afflicted  with it ( as in an endogenously produced seizure disorder, or Tourette’s syndrome). The Axis I “it” being treated is defined in terms of the presence of several blatant symptoms which allegedly constitute a syndrome (like panic disorder or major depressive disorder).

            Since the ‘it” being treated in psychopharmacotherapy research is deliberately divorced from biography, ontogeny, and personality (the person’s adaptation to life considered holistically), there is little interest in investigating how the treatment drug influences the patient’s life in terms of just those complex psychological considerations which have been expunged in order to liberate psychiatry from the Meyerian (biopsychosocial) framework and return it to mainstream medicine. In other words, I am proposing that the blatant unconcern about detecting the treatment drug’s effects on “complex cognitive functions and complex behavioral competencies” (after Streufert and Gengo, 1993) that is manifest in  psychopharmacotherapy research is entirely consistent with the neoKraepelinian agenda of resolutely turning away from the “host” as the subject of interest (I should note that Lennard et al. made a similar point in 1971). Thus, psychopharmacotherapy efficacy RCTs (which, again, are almost always about an Axis I disorder) do not introduce side effect possibilities that are more complex than the deliberately medicalized symptoms that define what is being treated. For example, the treatment research paper which apparently acted as the last straw for The New England Journal of Medicine’s editorial committee (Keller et al., 2000) shows the usual symmetry between what is being treated (major depressive disorder) and the treatment drug’s side effects, i.e.:

            The total investigative interest in what treatment with nefazadone brought about was exhausted by a maximum of two administrations of the Hamilton Rating Scale for Depression and a 15 to 20 minute question and answer period once a week devoted to concomitant use of medications, symptoms, side effects, and illnesses between visits (p.1463). In short, the usual approach to drug effects detection (outlined above) was employed. If my own discussion of what would be involved in a realistic attempt to discern drug-induced psychological alterations has merit, it should be evident that the usual approach employed in a psychiatric drug treatment efficacy RCT is tantamount to turning a blind eye towards drug-induced psychological alterations. I particularly want to re-emphasize the brief, highly structured, researcher-dominated question and answer interview devoted to somatic possibilities which carries the burden of side effects detection in psychoparmacotherapy efficacy RCTs. What we as readers really want to examine (via transcripts or even videotapes) is (a) what subjects actually said in the weekly interviews dedicated to symptoms and side effects review, and (b) the nature of the interview, that is, to what extent did the interviewer attempt to create a situation in which subjects were encouraged to reflect upon and talk about drug effects in everyday life? It is imperative to recall that Hoehen-Saric et al. (1990) were only able to detect fluoxetine and fluvoxamine-induced frontal lobe injury in virtue of allowing their patients (who were not participating in an efficacy RCT) to tell stories about their lives on the treatment drug as time wore on. I am pointing to a phenomenon that is entirely familiar to psychotherapists who encourage their clients to speak freely about personal matters: under such circumstances people wind-up revealing information that they barely if at all were cognizant of before they found themselves talking about it ( Wyatt [1986], a psychoanalyst, refers to this as the “knowing through telling paradigm”). Naturally, executing realistic methods for discerning what psychiatric drugs do would be time consuming, labor intensive, and expensive. The alternative is what prevails today: not knowing what psychiatric drugs do. 

 References

Anderson, B.G., Reker, D., and Cooper, T.B.  (1981).  Prolonged adverse effects of haloperidol in normal subjects.  The New England Journal of Medicine, 305, 643-644.

Andrews, G., Steward, G., Morris-Yates, A., Holt, P., and Henderson, S.  (1990).  Evidence for a general neurotic syndrome.  British Journal of Psychiatry, 157, 6-12.

Angell, M. (2000). Is academic medicine for sale? The New England Journal of Medicine, 342, 1516 – 1518.

Belmaker, R. H., and Wald, D.  (1977).  Haloperidol in normals.  British Journal of Psychiatry, 131, 222-223.

Bodenheimer, T. (2000). Uneasy alliance – clinical investigators and the pharmaceutical industry. The New England Journal of Medicine, 342, 1539-1544.

Brown, J.W.  (1988).  The life of the mind.  Hillside, New Jersey: Lawrence Erlbaum Associates.

Carpenter, W.T., McGlashan, T.H., and Strauss, J.S.  (1977).  The treatment of acute schizophrenia without drugs:  An investigation of some current assumptions.  American Journal of Psychiatry, 134, 14-20.

Deniker, P.  (1970).  Introduction of neuroleptic chemotherapy into psychiatry.  In F.J. Ayd, Jr., and B. Blackwell (Eds.), Discoveries in biological psychiatry (pp. 155-164).  Philadelphia:  J.P. Lippincott Co.

Diamond, R.  (1985).  Drugs and the quality of life:  The patient’s point of view.  Journal of Clinical Psychiatry, 46, 290-35.

Easton, K., and Link, I.  (1986).  Do neuroleptics prevent relapse?  Clinical observations in a psychosocial rehabilitation program.  Psychiatric Quarterly, 58, 42-50.

Elkin, I., Shea, M.T., Watkins, J.T., Imber, S.D., Sotsky, S.M., Collins, J.F., Glass, D.R., Dilkonis, D.A., Leber, W.R., Docherty, J.P., Fiester, S.J., and Parloff, M.B.  (1989).  National Institute of Mental Health treatment of depression collaborative research program.  Archives of General Psychiatry, 46, 971-982.

Estroff, S.  (1981).  Making it crazy.  Berkeley:  The University of California Press.

Finn, S.E., Bailey, J.M., Shultz, R.T., and Faber, R.  (1990).  Subjective utility ratings of neuroleptics in treating schizophrenia.  Psychological Medicine, 20, 843-848.

Fisher, C.M.  (1989).  Neurologic fragments II.  Remarks on anosognosia, confabulation, memory, and other topics.  Neurology, 39, 127-132.

Freyhan, F.A.  (1960).  Psychopharmacology and the controversial clinician.  In L. Uhr and T.G. Miller (Eds.), Drugs and behavior (pp. 184-198).  New York:  John Wiley and Sons, Inc.

Gadsby, J.E.  (2000).  Addiction by prescription.  Toronto:  Key Porter Books Limited.

Goodwin, J.  (1985).  Post-traumatic symptoms in incest victims.  In S. Eth and R.S. Pynoos (Eds.), Post-traumatic stress disorder in children (pp. 157-167).  Washington:  American Psychiatric Association Press, Inc.

Gram, L.F.  (1994).  Fluoxetine.  The New England Journal of Medicine, 331,1354-1361.

Healey, D.  (1999).  The three faces of antidepressants.  The Journal of Nervous and Mental Disease, 187, 174-180.

Healey, D., and Farquhar, G.  (1998).  Immediate effects of Droperidol.  Human Psychopharmacology, 13, 113-120.

Hoehn-Saric, R., Lipsey, J.R., and McLeod, D.A., (1990).  Apathy and indifference in patients on fluvoxamine and fluoxetine.  Journal of Clinical Psychopharmacology, 10, 343-345.

Jacobs, D.H.,  (1995).  Psychiatric drugging: forty years of pseudo-science, self-interest, and indifference to harm.  The Journal of Mind and Behavior, 16, 421-470.

Jacobs, D.H.  (1999).  A close and critical examination of how psychopharmacotherapy research is conducted.  The Journal of Mind and Behavior, 20, 311-350.

Jacobs, D.H., and Cohen, D.  (1999).  What is really known about psychological alterations produced by psychiatric drugs?  International Journal of Risk and Safety in Medicine, 12, 37-47.

Kahn, R.L., Fink, M., and Weinstein, E.A.  (1956).  Relation of amobarbital test to clinical improvement in electroshock.  A.M.A. Archives of Neurology and Psychiatry, 76,  23-29.

Keller, M.R., McCullough, J.P., Klein, D.N., Arnow, B., Dunner, D.L., Gelenberg, A.J.,

      Markowitz, J.C., Nemeroff, C.B., Russell, J.M., Thase, M.E., Trivedi, M.H., and Zajeka, J.

      (2000).  A Comparison of nefazadone, the cognitive-behavioral analysis system of psychotherapy, and their combination for the treatment of chronic depression.  The New England Journal of Medicine, 342, 1462-1470.

Kendler, K.S.  (1976).  A medical student’s experience with akathisia.  American Journal of Psychiatry, 133, 454-455.

Kleinman, A.  (1988).  Rethinking psychiatry.  New York: The Free Press.

Klerman, G.L. (1978). The evolution of a scientific nosology. In J.C. Shershow (Ed.), Schizophrenia : Theory and practice. (pp.99-121). Cambridge, Massachusetts : Harvard University Press.

Klerman, G.L. (1988). Overview of the cross-national collaborative panic study. Archives of General Psychiatry, 45, 407-412.

Kushner, H.I.  (1999).  A cursing brain?  The histories of Tourette’s syndrome.  Cambridge Massachusetts: Harvard University Press.

Lue, T.F.  (2000).  Drug therapy: Erectile dysfunction.  The New England Journal of Medicine, 342, 1802-1813.

Lennard, H.L., Epstein, L.J., Bernstein, A., and Ranson, D.C. (1971). Mystification and drug  misuse. New York : Harper and Row.

Mishler, E.  (1986).  Research interviewing:  Context and narrative.  Cambridge, Massachusetts:  Harvard University Press.

Mosher, L.R., and Menn, A.Z.  (1978).  Community residential treatment for schizophrenia: Two-year follow-up.  Hospital and Community Psychiatry, 29, 715-723.

Rapoport, J., Buchsbaum, M.S., Zahn, T.P., Weingartner, H., Ludlow, C., and Mikkelsen, E.J. (1978)  Dextroamphetamine:  Cognitive and behavioral effects in normal prepubertal boys.  Science, 99, 560-563.

Rappaport, M., Hopkins, H.K., Hall, K., Belleza, T., and Silverman, J.  (1978).  Are there schizophrenics for whom drugs may be unnecessary or contraindicated?  International Pharmacopsychiatry, 13, 100-111.

Rifkin, A., Quitkin, F., and Klein, D.F.  (1975).  Akinesia: A poorly recognized drug-induced extrapyramidal behavioral disorder.  Archives of General Psychiatry, 32, 672-674.

Rosenzweig, S.  (1986).  Idiodynamics vis-à-vis psychology.  American Psychologist, 41, 241-245.

Savodnik, I.  (1978).  The manifest and the scientific images.  In J.C. Shershow (Ed.), Schizophrenia: Theory and practice (pp. 21-43).  Cambridge, Massachusetts:  Harvard University Press.

Shear, K.M., Cooper, A.M., Klerman, G.L., Busch, F.N., and Shapiro, T.  (1993).  A psychodynamic model of panic disorder.  American Journal of Psychiatry, 150, 859-866.

Spiers, P.A., Schomer, D.L., Blume, H.W., and Hochanadel, G.S.  (1992).  Behavioral alterations in temperolimbic epilepsy.  In T. L. Bennett (Ed.), The neuropsychology of epilepsy (pp. 97-138).  New York: Plenum Press.

Strauss, J.S., Hafez, H., Leiberman, P., and Harding, C.M.  (1985).  The course of psychiatric disorder III: Longitudinal principles.  American Journal of Psychiatry, 142, 289-296.

Streufert, S., and Gengo, F.M.  (1993).  Drugs and behavior: An introduction.  In S. Streufert and F.M. Gengo. (Eds.), Effects of drugs on human functioning (pp. 1-12).  New York: Karger.

Stuss, D.T.  (1991).  Disturbance of self-awareness after frontal system damage.  In G.P. Prigatano and D.L. Schacter (Eds.), Awareness of deficit after brain injury (pp. 63-83).  New York: Oxford University Pres.

Stuss, D.T., and Benson, D.F.  (1986).  The frontal lobes.  New York:  Raven Press.

Summerfield, A.  (1978).  Behavioral toxicity ¾ The psychology of pollution.  Journal of Biosocial Science, 10, 335-345.

Valenstein, E.S.  (1986).  Great and desperate cures.  New York: Basic Books, Inc.

Van Putten, T., and Marder, S.R.  (1987).  Behavioral toxicity of antipsychotic drugs.  Journal of

Clinical Psychiatry, 48, 13-19.

Van Putten, T., May, P.R.A., Marder, S.R., and Wittman, L.A.  (1981).  Subjective response to antipsychotic drugs.  Archives of General Psychiatry, 38, 187-190.

Weinstein, E.A., and Kahn, R.L.  (1955).  Denial of illness:  Symbolic and physiologic aspects.  Springfield, Illinois: Charles C. Thomas.

 Widiger, T.A., and Shea, T.  (1991).  Differentiation of Axis I and Axis II disorders.  Journal of Abnormal Psychology, 100, 399-406.

Wyatt, F. (1986). The narrative in psychoanalysis : Psychoanalytic notes on storytelling, listening, and interviewing. In T.R. Sarbin (Ed.), Narrative psychology (pp.193-210). New York : Praeger.