The effects of psychotherapy.. An evaluation - Eysenck, H.J. (1952)

8/7/2019 The effects of psychotherapy.. An evaluation - Eysenck, H.J. (1952)

1/6

The Effects of Psychotherapy: An EvaluationH. J. Eysenck

Institute of Psychiatry, Maudsley HospitalUniversity of London

The recommendation of the Committee onTraining in Clinical Psychology of the Ameri-can Psychological Association regarding thet raining of clinical psychologists in the field ofpsychotherapy has been criticized by the writerin a series of papers [10, 11, 12]. Of thearguments presented in favor of the policy ad-vocated by the Committee, the most cogent oneis perhaps that which refers to the social needfor the sk ills possessed by the psyc hotherapist.In view of the importance of the issues in-volved, it seemed worth while to examine theevidence relating to the actual effects of psy-chotherapy, in an attempt to seek clarificationon a point of fac t .

Base Line and Unit of MeasurementIn the only previous attempt to carry outsuch an evaluation, Landis has pointed out that"before any sort of measurement can be made,it is necessary to establish a base line and acommon unit of measure. The only unit ofmeasure available is the report made by thephysician stating that the patient has recovered,is much improved, is improved or unimproved.This unit is probably as satisfactory as any typeof human subjective judgment, partaking ofboth th e good and bad points of such judg-ments" [26, p. 156.] For a uni t Landissuggests "that of expressing therapeutic re-sults in terms of the number of patients re-covered or improved pe r 100 cases admitted tothe hospital." As an alternative, he suggests"the statement of therapeutic outcome for somegiven group of patients during some stated in-terval of time."

Landis realized quite clearly that in order toevaluate the effectiveness of any f orm oftherapy, data from a control group of nontreat-ed patients would be required in order to com-pare the effects of therapy with the spontaneousremission rate. In the absence of anything bet-

ter, he used the amelioration rate in statemental hospitals for patients diagnosed underthe heading of "neuroses." As he points out:

There are several objections to the use of theconsolidated amelioration rate . . . of the . . .state hospitals ... as a base rate for spontaneousrecovery. The f ac t that psychoneurotic cases arenot usual ly commit ted to state hospitals unless in avery bad c o n d i t i o n ; the rela t ively small numberof voluntary patients in the group; the fact thatsuch pat ients do get some degree of psychotherapyespecial ly in the recep t ion hospi ta ls ; and the prob-ably qui te di f fe r en t economic, educational, and so-cial s tatus of the State Hospi tal group compared tothe patients reported from each of the other hospi-ta l sall argue agains t the acc eptance of [this] f ig-u re ... as a t ru ly sa t i s fac tory base line, but in theabsence of any other better figure this must serve[26, p. 168].

Actually the various figures quoted by Land-is agree very well. The percentage of neuroticpatients discharged annually as recovered orimproved from New York state hospitals is 70( f or the years 1925-1934); for the UnitedStates as a whole it is 68 ( f o r the years 1926to 1933). The percentage of neurotics dis-charged as recovered or improved within oneyear of admission is 66 for the United States(1933) and 68 for New York (1914). Theconsolidated amelioration rate of New Yorkstate hospitals, 1917-1934, is 72 per cent. Asthis is the figure chosen by Landis, we may ac-cept it in preference to the other very similarones quoted. By and large, we may thus saythat of severe neurotics receiving in the maincustodial care, and very little if any psycho-therapy, over two-thirds recovered or improvedto a considerable extent. "Although this is not,strictly speaking, a basic figure for 'spontane-ous' recovery, still any therapeutic methodmust show an appreciably greater size than thisto be serious ly cons idered" [26, p. 160].Another estimate of the required "baseline" is provided by Denker:

319


2/6

320 H. J. EysenckFive h u n d r e d consecutive disability claims due topsychoneurosis , t reated by general pract i t ionersthroughout the country, and not by accredi ted s p e -cial ists or sanatoria, were reviewed. A ll types ofneurosis were inc luded, and no a t tempt made todi f f erent ia te the neurasthenic, anxiety, compulsive,hysteric, or other states, but the grea tes t care w ast aken to eliminate the true psychotic or organiclesions which in the early stages of i l lness so of ten

simulate neurosis. These cases were taken consec-utively f rom the f i les of the Equi tab le Li fe Assur-ance Society of the United States, were from allparts of the country, and al l had been i l l of a neu-rosis for at least three months before claims weresubmi t ted . They, therefore, could be fa i r ly ca l led"severe," since they had been total ly disabled for atleast a three months' period, and rendered u n a b l e tocarry on with any "occupation for remuneration orprofit" for at least that t ime [9, p. 2164].

These patients were regularly seen andtreated by their own physicians with sedatives,tonics, suggestion, and reassurance, but in nocase was any at tempt made at anything butthis most superficial type of "psychotherapy"which has always been th e stock-in-trade of thegeneral prac t it ioner. Repeated s t a t em en ts ,every three months or so by their physicians,as well as independent investigations by the in-surance company, confirmed the fac t that thesepeople actually were not engaged in produc-tive work during the period of their illness.During their disablement, these cases receiveddisability benefits. As Denker points out, "Itis appreciated that this fact of disability incomem ay have actually prolonged the total periodof disability and acted as a barrier to incentivefo r recovery. One would, there for e, not expectth e therapeutic results in such a group of casesto be as favorable as in other grou ps where theeconomic factor might act as an important spurin helping the sick patient adjust to hisneuroticconflict and illness" [9, p. 2165].

The cases were all followed up for at leasta five-year period, an d often as long as tenyears af ter the period of disabili ty ha d begun.The criteria of "recovery" used by Denkerwere as follows: (a ) return to work, andability to carry on well in economic adjust-ments for at least a five-year period; ( ) com-plaint of no further or very slight dif f icul t ies;(e) making of successful social adjustments.Using these criteria, which are very similar tothose usually used by psychiatrists, Denkerfound that 45 per cent of the patients recovered

af ter one year, another 27 per cent after tw oyears, making 72 per cent in all. Another 10per cent, 5 per cent, and 4 per cent recoveredduring the thi rd , fourth, and fif th years, re-spectively, making a total of 90 per cent re -coveries af ter five years.This sample contrasts in many ways withthat used by Landis. The cases on whichDenker reports were probably not quite assevere as those summarized by Landis; theywere all voluntary, nonhospitalized patients,and came from a much higher socioeconomics t ra tum. The majori ty of Denker's patientswere clerical workers, executives, teachers, andprofessional men. In spite of these differences,the recovery f igures for the two samples arealmost identical. The most suitable figure tochoose from those given by Denker is probablythat for the two-year recovery rate, as follow-up studies seldom go beyond two years andthe higher f igures f or three-, four-, and five-year follow-up would overestimate th e e f f i -ciency of this "base line" procedure. Using,therefore, th e two-year recovery figure of 72per cent, we find that Denker's f igure agreesexactly with that given by Landis. We may,therefore, conc lude with some confidence thatou r estimate of some two-thirds of severeneurotics showing recovery or considerable im -provement without the benefit of systematicpsychotherapy is not likely to be very f ar out.

Effec t s of PsychotherapyWe may now turn to the effects of psycho-therapeutic treatment. The results of nineteenstudies reported in the literature, covering over

seven thousand cases, and dealing with bothpsychoanalytic and eclectic types of t reatment,are quoted in detail in Table 1. An attempthas been mad e to report results un der the f ou rheadings : (a) Cured, o r m u c h i mpr o ved; ( b )I m p rov e d ; (c) Slightly i mpr o ved; (d ) N otimproved, died, discontinued t rea tment , etc.It was usua lly easy to reduce addit ional cate-gories given by some writers to these basicf o u r ; some writers give only two or three cate-gories, and in those cases it was, of course, im -possible to s u b d iv id e f u r t he r , and the figuresfor combined categories are given.1 A slight

aln one or two cases where patients who im-proved or improved s l ight ly were combined by theoriginal author, the total f igure has been dividedequally between the two categories.


3/6

The Effects of Psychotherapy 321Table 1Summary of Reports of the Results of Psychotherapy

(A ) Psychoanalytic1. Fenichel [13, pp. 28-40]2. Kessel & H y m a n [24]3. Jones [22, pp. 12-14]4. Alexander [1, pp. 30-43]5. Knight [25]

All cases(B) Eclectic

1. Huddleson [20]2. M a t z [30]3. Maudsley HospitalReport (1931)4. Maudsley HospitalReport (1935)5 . Neus ta t te r [32]6. Luff & Garrod [27]7. Luff & Garrod [27]8. Ross [34]9. Yaskin [40]10. Curran [7] ..11. M a s s e r m a n &Carmichael [29]12. Carmichael &Masserman [4]..13. Schilder [35]14 . Hamilton & Wall [16]

15. Hamilton i al. [15] .16. Landis [26]17. Institute Med. Psychol.(quoted Neus ta t te r )18 . Wilder [39]19. Miles et al. ([31]

All cases

N

4843459

1414276 0

20 077 51721171146500210108910083

507735100100119

27 05453

7293

Cured ;m u c him -proved

104162028g33 5

191028 837 19140

3854729 51

7161132484058

3134661

Im -proved

84584220

7431090 076 5141358430629

20251134

547

1322418

Slightlyim -proved

9942823742 5

803105 33575826542364232

51461717 32551613

2632

Notim -proved ;died;lefttreat-ment

19793487

27145

1519934

182271732

25Ha

%Cured;m u c him -proved ;im -proved396247506744%

4641696450556877586154536366517370505864%

degree of subjectivity inevitably enters into thisprocedure, but it is doubtful if it has causedmuch distortion. A somewhat greater degreeof subjectivity is probably implied in thewriter's judgment as to which disorders anddiagnoses should be considered to fall underthe heading of "neurosis." Schizophrenic,manic-depressive, and paranoid states havebeen excluded; organ neuroses, psychopathicstates, and character disturbances have been in-cluded. The number of cases where there was

genuine doubt is probably too small to makem u c h change in the final figures, regardless ofhow they are allocated.

A number of studies have been excluded be-cause of such factors as excessive inadequacyof follow-up, partial duplication of cases withothers included in our table, failure to indicatetype of treatment used, and other reasonswhich made the results useless from our pointof view. Papers thus rejected are those byThorley & Craske [37], Bennett and Semrad


4/6

322 H. J. Eytenck[2], H. I. Harris [19], Hardcastle [17], A.Harris [18], Jacobson and Wright [21],Friess and Nelson [14], Comroe [5], Wenger[38], Orbison [33], Coon and Raymond [6],Denker [8], and Bond and Braceland [3].Their inclusion would not have altered ourconclusions to any considerable degree, al -though, as Miles et al. point out: "When thevarious studies are compared in terms ofthoroughness, careful planning, s tric tness ofcri teria and objectivity, there is often an in-verse correlat ion between these factors and thepercentage of successful results reported" [31,p. 88].

Certain difficulties have arisen from the in-ability of some writers to make their columnfigures agree with their totals, or to calculatepercentages accurately. Again, the writer hasexercised his ju d g m e nt as to which figures toaccept. In certain cases, writers have givenfigures of cases where there was a recurrenceof the disorder after apparent cure or im-provement , wi thout indicat ing how many pa-tients were affected in these two groups re-spectively. All recurrences of this kind havebeen subtracted from the "cured" and "im-proved" totals, taking half from each. Thetotal number of cases involved in all these ad-jus tments is quite small. Another investigatormaking al l decisions exactly in the oppositedirection to the present writer's would hardlyalter the final percentage figures by more than1 or 2 per cent.

We may now t u rn to the figures as present-ed . Patients treated by means of psychoanaly-sis improve to the extent of 44 per cent; pa-tients treated eclectically improve to the extentof 64 per cent; patients treated only custodiallyor by general practitioners improve to the ex-tent of 72 per cent. There thus appears to bean inverse correlation between recovery andpsychotherapy; the more psychotherapy, thesmaller the recovery rate. This conclusion re-quires certain qualifications.

In our tabulation of psychoanalytic results,w e have classed those w ho stopped treatmenttogether with those not improved. This ap-pears to be reasonable; a patient w ho fails tofinish his t reatment, and is not improved, issurely a therapeutic failure. The same rulehas been followed with the data summarizedunder "eclectic" treatment, except when the

patient who did not f inish treatment was defi-nitely classified as "improved" by the therapist.H o w e v e r , in v ie w of t he p e c u liar it ie s o fFreudian procedures it may appear to somereaders to be more jus t to class those casesseparately, and deal only with the percentageof completed treatments which are successful.Approximately one-third of the psychoanalyticpatients listed broke off treatment, so that thepercentage of successful t reatments of patientsw ho finished their course must be put at ap-proximately 66 per cent. It would appear,then, that when w e discount the risk the pa-t ient runs of s topping treatment al together, hischances of improvement under psychoanalysisare approximately equal to his chances of im-provement und er eclect ic treatment , an d slight-ly worse than his chances under a generalpractitioner or custodial treatment.

Tw o f u r the r points require clarificat ion:(a) A re patients in our "control" groups(L and i s and D e nk e r) as seriously ill as thosein our "experimental" groups? ( b ) A res tandards of recovery perhaps less stringent inour "control" than in our "experimental"groups ? It is difficult to answer these questionsdefinitely, in view of the great divergence ofopinion between psychiatrists. From a closescrutiny of the literature it appears that the"control" patients were probably at least asseriously ill as the "experimental" patients, andpossibly more so. As regards standards of re-covery, those in Denker's s tudy are as s tringentas most of those used by psychoanalysts an declectic psychiatris ts , but those used by theState Hospitals whose figures Landis quotesare very probably more lenient. In the absenceof agreed s tandards of severity of illness, or ofextent of recovery, it is not possible to go fur-ther.

In general, certain conclusions are possiblefrom these data. They fail to prove that psy-chotherapy, Freudian or otherwise, facilitatesthe recovery of neurotic patients. They showthat roughly two-thirds of a group of neuroticpatients will recover or improve to a markedextent within about tw o years of the onset oftheir illness, whether they are treated by meansof psychotherapy or not. This figure appearsto be remarkaby s table from one investigationto anoth er, regardless of type of patient treated,standard of recovery employed, or method of


5/6

The Effects of Psychotherapy 323therapy used. From the point of view of theneurotic, these figures are encouraging; fromthe point of view of the psyc hotherapist, theycan hardly be called very favorable to hisclaims.The figures quoted do not necessarily dis-prove the possibility of therapeutic effective-ness. There are obvious shortcomings in anyactuarial comparison an d these shortcomingsare particularly serious when there is so littleagreement among psychiatrists relating evento the most fundamental concepts and defini-tions. D efin ite proof would require a specialinvestigation, careful ly planned and method-ologically more adequate than these ad hoccomparisons. But even the much more modestconclusions that the figures fail to show anyfavorable effects of psychotherapy should givepause to those who would wish to give an im-portant part in the training of clinical psy-chologists to a skill the existence and effective-ness of which is s t i l l unsupported by any sci-entifically acceptable evidence.

These results and conclusions will no doubtcontradict the s trong feeling of usefulness andtherapeutic success which many psychiatristsand clinical psychologists hold. While it ist ru e that subjective feelings of this type haveno place in science, they are likely to preventan easy acceptance of the general argumentpresented here. This contradict ion betweenobjective f ac t and subjective c ertainty has beenremarked on in other connections by Kelly an dFiske, who found that "One aspect of our find-ings is most disconcert ing to us: the inverserelationship between th e confidence of staffmembers at the t ime of making a prediction an dthe measured validity of that prediction. W hyis it , f or example, that our staff members tend-ed to m ake their best predict ions at a t ime whenthey subjectively felt relatively unacquaintedwith the candidate, when they had constructedno systematic picture of his personality struc-ture? Or conversely, why is it that with in-creasing confidence in clinical j u d g m e n t . ..wefind decreasing validities of predictions?" [23,p. 406].In the absence of agreement between factand belief, there is urgent need for a decreasein the strength of belief, and for an increasein the number of facts available. Until suchfacts as may be discovered in a process of rigor-

ous analysis support the prevalent belief intherapeutic effectiveness of psychological treat-ment, it seems premature to insist on the in-clusion of training in such treatment in thecurriculum of the clinical psychologist.

SummaryA survey was m ad e of reports on the im-provement of neurotic patients af ter psycho-therapy, and the results compared with the bestavailable estimates of recovery without benefitof such therapy. The figures fai l to support

the hypothesis that psychotherapy facilitates re-covery from neurotic disorder. In view of themany difficulties at tending such actuarial com-parisons, no further conclusions could be de-rived from the data whose shortcomings high-light the necessity of properly planned andexecuted experimental studies into this im -portant field.Received January 23 , 1952.

References1. Alexander, F. Five year report of the Chi-cago Institute fo r Psychoanalysis, 1932-1937.2. Bennett, A. E., & Semrad, E. V. Common er-rors in diagnosis and t r e a tm e n t of the psycho-neurot ic pa t ienta s tudy of 100 case histories.Nebr. med. J., 1936, 21, 90-92.3. Bond, E. D., & B r a c e l a n d , F. J. Prognosis inmental disease. Amer. J. Psychiat., 1937, 94,263-274.4. Carmichael , H. T., & M a s s e r m a n , T. H. Re-sults of treatment in a psychiatric outpatients'd e p a r tm e n t . /. Amer. med. Ass., 1939, 113,2292-2298.5. Comroe, B. I. Follow-up s tudy of 100 pat ientsdiagnosed as "neurosis." /. nerv. ment. Dis.,

1936, 83, 679-684.6. Coon, G. P., & R a y m o n d , A. A review of thepsychoneuroses at Stockbridge. Stockbridge,Mass.: Austen Riggs Foundation, Inc., 1940.7. C u r r a n , D. The problem of assessing psychi-atric treatment. Lancet, 1937, II, 1005-1009.8. Denker, P. G. Prognosis and l i fe expectancyin the psychoneuroses. Proc. Ass. Life Insur.med. Dir. Amer., 1937, 24, 179.9. Denker, R. Results of t rea tment of psycho-neuroses by the general practi tioner. A fol low-up s tudy of 500 cases. N. Y. State J. Med.,1946, 46, 2164-2166.

10. Eysenck, H. J. Training in clinical psychol-ogy: an English point of view. Amer. Psychol-ogist, 1949, 4, 173-176.11. Eysenck, H. J. The relation between medicineand psychology in England. In W. Dennis(Ed.), Current trends in the relation of psy-


6/6

324 H. J. Eysenckchology an d medicine. Pit t sburgh: Univer. ofPittsburgh Press, 1950.

12. Eysenck, H. J. Function and t ra ining of theclinical psychologist. J. went. Sci., 1950, 96,1-16.13. Fenic hel, O. Ten years of the Berlin Psycho-analysis Institute. 1920-1930.14. Friess, C., & Nelson, M. J. Psychoneuroticsfive years later. Amer. J. ment. Sci., 1942,203, 539-558.15 . Hamil ton, D. M., Vanney, I. H., & Wall, T.H . Hospi ta l t rea tment of pat ients wi th psy-choneurotic disorder. Amer. J. Psychiat., 1942,99, 243-247.16 . H a m i l to n , D. M., & W a l l , T. H. Hospitalt rea tment of pa t ients w i th psyc honeurot ic d is-order . Amer. J. Psychiat., 1941, 98, 551-557.

17. H a r d c a s t l e , D. H. A follow-up s tu d y of oneh u n d r e d cases made for the Depar tment ofPsychological Medicine, Guy's Hospital. J.ment. Sci., 1934, 90 , 536-549.18. Harris, A. The prognosis of anxiety states.Brit. med. J., 1938, 2, 649-654-.19. Harr is , H. I. Efficient psychotherapy for thelarge out-patient clinic. New England J. Med.,1939, 221, 1-5.20 . Huddleson, J. H. Psychotherapy in 200 casesof psychoneurosis. Mil. Surgeon, 1927, 60 , 161-170.

21. Jacobson, J. R., & Wright , K. W. Review ofa year of group psychotherapy. Psychiat.Quart., 1942, 16, 744-764.22 . Jones, E . Decennial report of the LondonClinic of Psychoanalysis. 1926-1936.23 . Kelly, E. L., & Fiske, D. W. The prediction ofsuccess in the VA training program in clinicalpsychology. Amer. Psychologist, 1950, 5, 395-406.

24. Kessel, L., & Hy m an , H. T. The value of psy-choanalysis as a therapeut ic procedure. J.Amer. med. Ass., 1933, 101, 1612-1615.25 . Knight, R. O. Evaluation of the results ofpsychoanalytic therapy. Amer. J. Psychiat.,1941, 98, 434-446.26 . Landis , C. Statistical evaluation of psycho-therapeutic methods. In S. E . Hinsie ( E d . ) ,

Concepts and problems of psychotherapy. Lon-don: Heineman, 1938. Pp. 155-165.27 . Luff , M . C., & Garrod, M. The a f te r - r e s u l t sof psychotherapy in 500 adult cases. Brit. med.J., 1935, 2, 54-59.28. Mapother, E. Discussion. Brit. J. med. Psy-chol., 1927, 7, 57.29. Masserman, T. H., & Carmichael , H. T. Diag-nosis and prognosis in psychiatry. J. ment. Sci.,1938, 84, 893-946.30. Matz , P. B. Outcome of hospi ta l t rea tment ofex-service pa t ients wi th nervous and menta ldisease in the U. S. Veteran's Bureau. 17. tf .Vet. Bur. med. Bull, 1929, 5, 829-842.31. Miles , H. H. W., Barrabee, E . L., & Finesinger,J . E. E v a l u a t i o n of psychotherapy. Psychosom.Med., 1951, 13, 83-105.

32 . N e u s t a t t e r , W. L. The resul ts of fifty casestrea ted by psychotherapy. Lancet, 1935, I, 796-799.33. Orbison, T. J. The psychoneuroses: psychas-thenia, neuras thenia and hysteria, with spe-cial re ference to a cer ta in method of trea tment .Calif, west. Med., 1925, 23, 1132-1136.34. Ross, T. A. An enquiry into prognosis in theneuroses. Lon don : Cam bridge Un iver. Press,1936.35 . Schilder, P, Results an d problems of grouppsychotherapy in severe neuroses. Ment. Hyg.,N. Y., 1939, 23, 87-98.36. Skottowe, I., & Lockwood, M. R. The fate of15 0 psychia tr ic ou tpa t ients . J. ment. Sci., 1935,81, 502-508.37. Thorley, A. S., & Craske, N. Comparison andestimate of group an d individual method oftreatment. Brit. med. J., 1950, 1, 97-100.38. Wenger, P. Uber weitere Ergebnisse der Psy-chotherapie in Rahmen einer MedizinischenPoliklinik. Wien. med. Wschr., 1934, 84, 320-325.39. Wilder, J. Facts and figures on psychother-apy, /. din. Psychopath., 1945, 7, 311-347.

40. Yaskin, J. C. The psychoneuroses and neu-roses. A review of 100 cases wit h special ref-erence to t rea tment and results. Amer. J. Psy-chiat., 1936, 93, 107-125.

The effects of psychotherapy.. An evaluation - Eysenck, H.J. (1952)

Documents

Transcript of The effects of psychotherapy.. An evaluation - Eysenck, H.J. (1952)