Isr J Psychiatry Relat Sci Vol 44 No. 4 (2007) 301–308

Evidence-Based Psychiatric Practice? Long Live the (Individual) Difference

Nirit Soffer, BA,1 and Golan Shahar, PhD1, 2

1 Department of Behavioral Sciences, Ben-Gurion University of the Negev, Beersheba, Israel

2 Department of Psychiatry, School of Medicine, New Haven, Connecticut, U.S.A..

Abstract: Mental health services in Israel are about to undergo a major, and quite controversial, reform. Yet both those advocating these reforms and those opposing them refrain from resorting to empirical science in arguing their respec- tive cases. In the present article, we present findings concerning the central role of patients’ individual differences in the outcome of mental health services. Data is presented from three research projects: 1. The National Institute of Mental Health (NIMH) Treatment of Collaborative Research Program (TDCRP), in particular recent publications by Blatt, Zuroff, Shahar and their collaborators, 2. The Partnership Project in severe mental illness, and 3. The Menninger Research Project (MPRP). Findings from these three projects, largely predicted by Blatt’s theory of interpersonal relatedness and self-definition, suggest that an effective, evidence-based mental health practice is contingent upon practitioners’ sensitivity to pretreatment patient characteristics. Clinical and policy impli- cations of this conclusion are discussed.

Mental health services in Israel are about to undergo cent decades in empirically-validated forms of psy- a major reform. With this reform, responsibility for chotherapy and psychopharmacology for various mental health services will shift from the govern- psychiatric disorders. The knowledge attained mental Department of Health to private health man- through research could have been conducive to re- agement organizations (HMOs). The above- solving at least some of the controversies surround- mentioned reform is meant to decrease hospitaliza- ing the Israeli reform. For example, given the tion rates and reduce stigma. Nevertheless, the re- demonstrated efficacy of Cognitive-Behavioral form is highly controversial. For instance, the Israeli Treatment (CBT) for various anxiety disorders such Psychological Association as well as the clinical psy- as Panic Disorder, Obsessive-Compulsive Disorder, chology division of the Israeli Council of Psycholo- Post-Traumatic Stress Disorder, and others, one gists strongly object to this reform, arguing that it would expect reform designers to ensure adequate will seriously compromise the delivery of psycholog- coverage for CBT in these disorders. Unfortunately, ical interventions (e.g., psychotherapy) which will in the number of psychotherapy sessions to be covered turn impede the quality of mental health treatment. by HMOs under the new reform is even considerably Several social work organizations expressed similar lower than the number suggested by Randomized concerns (1). Conversely, consumer organizations Clinical Trials (RCTs) demonstrating the efficacy of appear to be highly in favor of the upcoming reform, CBT in anxiety disorders. particularly because they believe that the reform is Further complicating this issue is recent scientific likely to reduce stigmatization (2). criticism regarding RCTs in mental health practice Conspicuously absent, however, in the debate (e.g., 3, 4), including (1) these trials’ exclusion of over the upcoming reform is the voice of empirical comorbid cases, thereby severely limiting the ecolog- science. Specifically, both those advocating the re- ical validity of the tested treatments, (2) the trials’ form and those objecting to it do not base their argu- disregard for patients’ treatment preferences, which ments on currently available data on the outcome of has been recently demonstrated to play a key role in various mental health treatments. This is highly un- treatment, and (3) the distinct brevity of manualized fortunate given the important advances made in re- treatment which, at least in the case of unipolar de-

Address for Correspondence: Nirit Soffer, Department of Behavioral Sciences, Ben-Gurion University of the Negev, Beersheba 84105, Israel. E-mail: [email protected] 302 EVIDENCE-BASED PSYCHIATRIC PRACTICE? LONG LIVE THE (INDIVIDUAL) DIFFERENCE pression, impedes outcome (4). In the present arti- uations of patients were conducted throughout the cle, we would like to address another important study (at pretreatment and every 4-week interval), as limitation: RCTs’ disregard for the role of patients’ wellasthreefollow-upmeasurementsat6,12,and pretreatment characteristics in treatment outcome 18 months. This large-scale project, consisting of (4). Such disregard goes counter to an established over 200 seriously depressed patients, and 28 experi- tradition in psychotherapy research, dating back to enced, well-trained clinicians, was carefully imple- the great methodologist, Lee Cronbach (5), who mented. Patient assessments included an extensive urged investigators to be attuned to Treatment-Apti- battery of self-report and interview-based measures tude Effects, i.e., to the effects of patients’ pretreat- of symptoms, personality, attitudes and social con- ment characteristics. Other theoreticians and text variables. investigators have highlighted the important role of Initial results focusing on differences between the patients’ individual differences in the treatment pro- three active treatments were largely disappointing, cess (6, 7), but until recently, these qualifications in that such differences were not detected (8). How- were largely ignored in the zeal to conduct “horse ever, further analyses (for a review, see 4) revealed race” studies comparing the efficacy of several treat- the significance of patient pretreatment personality ments. characteristics in predicting therapeutic outcome. In what follows, we will illustrate the crucial role Using the Dysfunctional Attitudes Scale (DAS)(9), of patients’ pretreatment characteristics in three Blatt and his colleagues were able to isolate patients’ major clinical studies: the Treatment of Depression pretreatment perfectionism and to demonstrate that Collaborative Research Project (TDCRP) sponsored this personality trait predicted treatment outcome by the National Institution of Mental Health, which according to several measures. focused on the treatment for major depression, the Thus, Blatt et al. (10) found that high levels of Partnership Project (TPP), which focused on a part- pretreatment perfectionism significantly impeded nership-based community intervention for people consequent symptom reduction among two-thirds with severe, and largely psychotic, mental illness, of patients in the study. Specifically, it was shown and the now classic Menninger Psychotherapy Re- that during the first half of treatment (initial 8 search Project (MPRP) in which classical psycho- weeks), all patients generally had a decrease in symp- analysis and supportive-expressive psychotherapy tomatic maladjustment, whereas in the second half were compared. After presenting the findings attest- of treatment (9 to 16 weeks), those patients with high ing to the importance of patients’ individual differ- or medium perfectionism ceased to improve signifi- ences in the treatment process, we will discuss cantly in symptom reduction, while low-perfection- implications of these findings to clinical practice and ism patients continued to improve. As argued by mental health policy. Blatt and Zuroff (4), introjective, self-critical, con- trol-focused, perfectionist patients are likely to react negatively to a pre-defined, externally determined The Treatment of Depression treatment termination date and review their degree Collaborative Research Project (TDCRP) of improvement pessimistically in prospect of this The TDCRP is perhaps the most well known Ran- uncontrollable cessation of therapy. domized Clinical Trial (RCT) for the treatment of In addition to symptom reduction, another im- unipolar depression. It was sponsored by the Ameri- portant outcome variable is the quality of the thera- can National Institute of Mental Health (NIMH). peutic alliance. Zuroff et al. (11) found that Data from this project consists of four conditions for pretreatment perfectionism significantly interfered treatment of severe depression. Two conditions are with patients’ participation in the therapeutic alli- active, brief (16-week) outpatient ance during the latter half of treatment, in turn lead- for depression: Cognitive-Behavioral (CBT) and In- ing to failure to respond to this treatment. terpersonal therapy (IPT). The third condition is an- Nevertheless, the path leading from pretreatment tidepressant medication (Imipramine) and the perfectionism to poor response to treatment via the fourth, a double-blind placebo. Comprehensive eval- therapeutic alliance did not account for the entire NIRIT SOFFER AND GOLAN SHAHAR 303 deleterious effect of pretreatment perfectionism on ventory (BDI)(15), the total score on the Hopkins treatment outcome. In an attempt to account for the Symptoms Checklist-90 (SCL-90)(16), the 17-item entire effect, Shahar et al. (12) found another equally Hamilton Rating Depression Scale (HRDS)(17), the important mechanism of the adverse effect of perfec- Global Assessment Scale (GAS)(18), and the sum of tionism: patients’ social network. Specifically, these the global ratings from the Social Adjustment Scale investigators found that patients’ pretreatment per- (SAS)(19). fectionism predicts low satisfaction with their social network, which in turn predicts poor treatment re- Table 1. Regression analyses of factors contributing to sponse. This path, alongside the aforementioned treatment outcome in the TDCRP path involving the therapeutic alliance, fully ac- N=133 â p-level counted for the deleterious effect of pretreatment perfectionism on treatment outcome. Thus, it may Treatment -.12 .12 be said that pretreatment perfectionism interferes Pretreatment symptoms .20* .01 with interpersonal relations both within and outside Pretreatment perfectionism .30*** .0001 treatment. Therapeutic alliance -.21** .009 In a more recent analysis, Shahar et al. (13) exam- Pretreatment social network -.20* .01 ined interactions between pretreatment perfection- 2 ism and pretreatment satisfaction with social Notes: R =.25, F[5,127]=8.45, p<.001. network in predicting treatment outcome. These in- vestigators identified a complex, curvilinear interac- As shown in Table 1, this regression model ac- tion between perfectionism and satisfaction with counted for 22% of the variance of the outcome. Sta- social network, such that satisfaction with social net- tistically significant predictors were: patients’ work buffered against the deleterious effect of per- pretreatment symptoms (â=.20, p<.05, accounting fectionism on treatment outcome. This effect, for 4% of the variance), pretreatment perfectionism however, was present only in the low-to-mid-range (â=.30, p<.001, accounting for 9% of the variance), of perfectionism. Patients with very high levels of the therapeutic alliance in the 3rd session (â=-.20, perfectionism were not able to benefit from social p<.01, accounting for 4% of the variance) and pre- network in the course of treatment. Particularly im- treatment social network (â=-.20, p=.01, accounting portant in these findings is their contribution to un- for 4% of the variance). The treatment variance, con- derstanding the complex role of individual trasting the three active treatments with the placebo differences in treatment outcome. Specifically, these condition, accounted for only 1% of the total vari- findings suggest that not only are patients’ individual ance and did not emerge as a statistically significant differences pertinent to the assessment of outcome, predictor (â=-.12, p=.11). These findings clearly at- but also that these individual differences interact testthat,atleastintheTDCRP,theimportantinflu- with social context in complex ways. ences on treatment outcome were patients’ To further highlight the important role of indi- characteristics and interpersonal relations, rather vidual differences in treatment outcome, we turned than treatment type. back to the TDCRP data and re-analyzed it. We fo- cused on the percent of variance in the outcome vari- The Partnership Project (TPP) able which is accounted for by the following factors: treatment type, patients’ pretreatment perfection- In a large-scale (N=260), longitudinal, both qualita- ism, patients’ satisfaction with their social network, tive and quantitative study among people with Se- and the therapeutic alliance established in the 3rd vere Mental Illness (SMI), the importance of session of the treatment. In Table 1, we present re- individual differences in determining intervention sults of a multiple regression analysis examining this characteristics was well exemplified. This project, a question. The dependent variable in this analysis is randomized, controlled clinical trial, concentrated Blatt et al.’s (14) composite factor which is comprised on improving the social functioning of socially iso- of five outcome measures: the Beck Depression In- lated people with psychiatric disabilities (20). Spe- 304 EVIDENCE-BASED PSYCHIATRIC PRACTICE? LONG LIVE THE (INDIVIDUAL) DIFFERENCE cifically, the focus of the intervention was to increase prove, while those who did not meet with their enjoyable social activities, with or without a partner consumer partner did, in fact, improve. (i.e., with or without receiving “supported socializa- These counterintuitive results become clearer tion”). To that effect, participants received a monthly when considering participants’ expectations and $28 stipend to spend on such activities. Three differ- self-concepts. Qualitative data revealed that meeting ent intervention conditions were utilized: a stipend- with nonconsumer partners enabled participants to only condition, and two stipend-plus-partner condi- feel “normal” and “live outside their illness” (21). In tions. In the first partnership condition, the volun- contrast, meeting with consumer partners facilitated teer partners were recovered consumers of mental the “internalized stigma,” consequently confining health services (i.e., had a history of mental health the participants to the mental health system, and disorders themselves). In the second partnership thus limiting their ability to improve their social condition, partners were nonconsumers (i.e., volun- functioning. It is evident from these results that the teers from the community without histories of men- efficacy of treatment is contingent upon the partici- tal health conditions). The outcome variables pant’s perceptions and expectations. Additional analyses of the data (22) shed light on included psychiatric symptoms, well-being, self-es- the individual differences found in these perceptions teem, level of activity, functional impairment, and and expectations. The investigators assumed that the satisfaction, and were measured at three points in effect of stigma, and the desire of a participant to as- time during the 9-month trial (at baseline, 4 months sociate him or herself with the healthy rather than and 9 months). the mentally ill, would be of particular pertinence For people with SMI, social isolation and lack of among participants with high self-esteem. Spe- friendships pose a major challenge to their re-enter- cifically, they hypothesized that those individuals ing the community (20). This clinical trial was there- who were matched with a consumer partner and ex- fore of great value for understanding the ideal terms hibited initial high self-esteem would later exhibit of intervention by means of social support (i.e., con- low satisfaction regarding treatment. Factor analysis sumer vs. nonconsumer). However, the initial analy- conducted prior to examining the hypothesis re- sis of the data revealed a marked lack of difference vealed that self-esteem was embedded in a larger between the three different conditions, though all construct, composed of self-esteem, sense of pur- three groups showed improvement over time. In pose, and sense of growth, as well as low levels of de- other words, it seemed that contrary to expectations, pression. This larger construct was labeled “morale” notonlywastherenodifferencebetweencon- and was used instead of the initial “self-esteem” vari- sumer/nonconsumer conditions, the two consumer- able. Consistent with predictions, high initial morale based conditions were also not superior to the con- predicted greater satisfaction with services in the trol, stipend-only condition in terms of improve- control and nonconsumer conditions, and lower sat- ment in the outcome variables. isfaction in the consumer condition. Conversely, However, a closer look at the data unveils a more among low-morale participants, morale predicted complex story. The investigators noticed that when greater satisfaction in the consumer group, and they take into account participants’ “adherence,” lower satisfaction in the control and nonconsumer namely the extent to which they actually met with groups. their partner, group differences emerged. Overall, it In yet another analysis of these data (23), individ- seemed that in the nonconsumer condition, meeting ual differences in recovery rates were also found to regularly with the partner predicted improvement be of psychological importance. This study showed (greater than the improvement of the stipend-only that initial depressive symptoms predicted a de- group), while failure to meet with the partner pre- crease in self-esteem. However, this effect was found dicted lack of improvement or deterioration. In only in the first, 4-month period, and not in the sec- sharp contrast, however, the results were reversed in ond, 5-month period. Additional analyses illustrated the consumer condition: generally, those who met that the effect was moderated by treatment improve- regularly with their consumer partner did not im- ment, such that among those who improved in terms NIRIT SOFFER AND GOLAN SHAHAR 305 of social functioning in both the baseline-through-4- consisting of benevolent interpersonal schemas, and months period and the 4-months-through-9- maladaptive representations, consisting of malevo- months period, the adverse effect of depressive lent interpersonal schemas. symptoms on self-esteem was no longer statistically Findings showed that regarding development of significant. This finding indicates that individual dif- adaptive interpersonal schemas, PSA demonstrated ferences in extent of recovery may serve as a buffer better results than SEP,although this was true mainly against the otherwise detrimental effect of depres- in introjective patients. In fact, while SEP led to im- sive symptoms. provement in anaclitic patients, introjective patients in the SEP condition had a non-significant decline in adaptive responses. Regarding the reduction of The Menninger Psychotherapy Research maladaptive interpersonal schemas, another signifi- Project (MPRP) cant treatment-patient interaction was found: PSA In a recent analysis (24) of data from the Menninger led to a reduction of maladaptive representations Psychotherapy Research Project (25), important among introjective patients, but an increase of these findings emerged regarding the significance of pre- representations among anaclitic patients. SEP, how- treatment individual differences when planning and ever, led to a decline in maladaptive representations implementing therapeutic interventions. The data among anaclitic patients, but an increase of them collected in the MPRP was based on 42 individuals, among introjective patients. These interesting find- who were referred either to a classical, five-sessions- ings demonstrate the significance of patients’ pre- a-week, psychoanalytic treatment (PSA), or to a two- treatment characteristics in responding to long-term to-three-sessions-a-week supportive/expressive psy- psychotherapy. chotherapy (SEP). While participants were not ran- Curious as to these mechanisms, the authors con- domized to the two treatment conditions, but were ducted another analysis based on the total number of rather referred based on clinical judgment, no statis- Rorschach responses, an indicator of level of patient tically significant pretreatment differences were de- associational activity. The authors hypothesized that tected between the PSA and SEP groups. The study SEP leads to therapeutic gain among affectively la- variables were derived from extensive clinical case bile anaclitic patients by providing a containing and records, as well as from psychological tests. Mea- supportive environment to enable a reduction of surement points were at pretreatment, termination their associational activity. On the other hand, PSA andavariablefollow-up. leads to therapeutic gain among emotionally de- As initially reported by Wallerstein (25), PSA and tached, interpersonally isolated introjective patients SEP did not differ in terms of treatment improve- by supplying interpretations, which help increase ment. Nevertheless, Blatt (26) introduced the dis- their associational activity. This additional analysis tinction between introjective (independent and self- showed that indeed, among both patient categories, critical) personality and anaclitic (dependent and PSA led to an increase in Rorschach responses, and affectively labile) personality to the MPRP data. This SEPledtoadecreaseinresponses.However,dueto distinction yielded significant differences in treat- the small sample size, these trends did not reach sta- ment effectiveness as a function of patient pretreat- tistical significance. ment personality structure. A further analysis of these data (27) provided yet Blatt and Shahar’s (24) analysis utilized this dis- another indication of the crucial role played by pa- tinction, defining patient’s personality structure tients’ pretreatment personality in the therapeutic based on clinical conclusions of two independent process. The authors measured patients’ benevolent judges. Outcome variables were pretreatment-to- and adaptive interaction representations, as well as termination changes in five psychological constructs their malevolent and maladaptive interaction repre- based on Rorschach protocols, two assessing the sentations, only this time as predictor variables concept of the human figure and three assessing the (measured at baseline) rather than outcome vari- thematic content of interactions. These constructs ables (measured at termination). In accordance with composed two factors: adaptive representations, the authors’ hypothesis, initial interpersonal 306 EVIDENCE-BASED PSYCHIATRIC PRACTICE? LONG LIVE THE (INDIVIDUAL) DIFFERENCE schemas were correlated significantly with clinical see 29). This patient had responded partially to an functioningattermination,inbothtypesoftherapy. SSRI anti-depressant medication, but significant re- Better representations at intake predicted better clin- sidual depressive symptoms remained, particularly ical functioning at termination, even when control- symptoms pertaining to anhedonia, hopelessness, ling for clinical functioning at baseline. These data lack of energy, social isolation, and suicidal ide- suggest that assessment of pretreatment patient ations. Subsequent to intake, an attempt was made to characteristics might contribute to treatment plan- treat BR using behavioral activation, an empirically ning and evaluation of therapeutic outcome. supported therapeutic technique which aims to en- gage sufferers in enjoyable activities. This effort is an attempt to establish a chain of reward-based operant Summary and Conclusion: Bringing conditioning, which will ultimately improve energy Back the Individual Into Evidence-Based and mood, and break social isolation. BR adamantly Treatment refused to engage in this technique. With time, BR’s With respect to a few manualized treatments, partic- refusal became more comprehensible in light of his ularly those addressing anxiety disorders, the case background and psychodynamics, and in the context can be made to the secondary, if not downright neg- of the transference-countertransference matrix. Spe- ligible role of patients’ individual differences in out- cifically, the patient experienced the suggestion that come. However, with respect to many other he actively engage in enjoyable activities as a com- treatments, psychotherapeutic and pharmacological mand made by a judgemental and punitive father alike, the consideration of patients’ pretreatment figure, and his refusal to such an engagement was ex- characteristics in the treatment process and outcome perienced by him as taking an active and autono- appearscrucial.Itissimplythecase,asarguedby mous stance towards this “introject.” When such Blatt and Felsen (7) that “different strokes are suit- feelings were explicated and worked through in the able for different folks.” Patients come to treatments course of treatment, BR gradually changed his mind with different preferences, beliefs, expectations, per- regarding behavioral activation and began to employ sonalities, and social network configurations, and it spontaneously, to his own benefit. This case illus- these individual difference variables impact upon trates that evidence-based techniques offered to pa- patients’ resources to improve in the course of treat- tients are not “neutral” prescriptions. Rather, they ment. Put differently, while the quest for scientific, can and should be construed as “lived objects” in evidence-based psychiatry and clinical is psychoanalytic terms, and patients do indeed react commendable, it simply will not work without to them with varied responses conditioned upon bringing back the patients’ person, his or her subjec- their background and personality. Without taking tivity, into clinical science and practice. these responses into consideration, the efficacy of More importantly, the two ideals of evidence- evidence-based techniques are, at best, highly lim- based treatment and sensitivity to pretreatment indi- ited. At worse, they may be downright destructive. vidual differences are not mutually exclusive. As ar- As a final word, we would like to note that discus- gued by Stricker’s (28) model of the clinical sion of this case, as well as that of individual differ- practitioner as a local scientist, these two consider- ences in the treatment process and outcome, are not ations might be fully integrated by clinicians’ being solely abstract ones. Such discussions “catch us,” as it trained and informed by state-of-the-art, empirically were, in the midst of a turning point regarding this validated assessment and treatment techniques, country’s mental health reform. As we noted above, while applying these techniques judicially and sensi- thisreformdoesappeartobegovernedbygoodin- tively, taking into account patients’ “local” condi- tentions. Unfortunately, we know all too well where tions, namely, their individual differences. Such an such good intentions sometimes lead. In this case, approach might be illustrated by a case treated by the the reform’s architects’ lack of consideration for both second author. BR, a 20-year-old male, sought psy- evidence-based practice and patients’ individual dif- chotherapy for major depression superimposed on a ferences appears to lead to dangerously trimmed chronic dysthymic disorder (i.e., double depression, mental health services provided by untrained staff. NIRIT SOFFER AND GOLAN SHAHAR 307

We fear that the consequences, manifested in terms 13. Shahar G, Blatt S J, Zuroff D C. Satisfaction with social of under-treatment and resultant medical, social and relations buffers against the adverse effect of (mid- economic problems, are likely to be devastating. level) perfectionism on outcome in brief treatment for depression. J Soc Clin Psychol, in press. 14. BlattSJ,ZuroffDC,QuinlanDM,PilkonisP.Interper- References sonal factors in brief treatment for depression: Further analysis of the National Institution of Mental Health 1. KorenD,ShaharG.Anirresponsiblementalhealthre- Treatment of Depression Collaborative Research Pro- form. Haaretz, October 22, 2006, http://www. gram. J Consult Clin Psychol 1996;64:162–171. haaretz.co.il/hasite/spages/777710.html 15. Beck A T, Ward C H, Mendelson M, Mock J, Erbaugh J. 2. Tzur D, Weinberg D, Cohen G. The Israeli Mental An inventory for measuring depression. Arch Gen Psy- Health Reform. Unpublished Manuscript, Ben-Gurion chiatry 1961;4:561–571. University of the Negev, Beersheba, Israel, 2006. 16. Derogatis L R, Lipman R S, Covi M D. SCL-90: An out- 3. Westen D, Novotny C M, Thompson-Brenner H. The patient psychiatric rating scale — preliminary report. empirical status of empirically supported psycho- Psychopharmacol Bull 1973;9:13–20. therapies: Assumptions, findings, and reporting in 17. Hamilton M A. A rating scale for depression. J Neurol controlled clinical trials. Psychol Bull 2004;130:631– Neurosurg Psychiatry 1960;6:56–62. 663. 18. Endicott J, Spitzer R L, Fleiss J L, Cohen J. The global 4. BlattSJ,ZuroffDC.Empiricalevaluationoftheas- assessment scale: A procedure for measuring overall sumptions in identifying evidence based treatments in severity of psychiatric disturbance. Arc Gen Psychiatry mental health. Clin Psychol Rev 2005;25:459–486. 1976;33:766–771. 5. Cronbach L J. Correlations between persons as a re- 19.WeissmanMM,PaykelES.Thedepressedwomen: search tool. In: Mowrer O H, editor. Psychotherapy: Study of social relationships. Chicago: University of Theory and research. New York: Ronald Press, 1953: Chicago, 1974. pp. 376–389. 20. Davidson L, Shahar G, Stayner D A, Chinman M J, 6. Kiesler D J. Some myths of psychotherapy research and Rakfeldt J, Tebes J K. Supported socialization for peo- the search for a paradigm. Psychol Bull 1966;65:110– ple with psychiatric disabilities: Lessons from a ran- 136. domized controlled trial. J Community Psychol 2004; 7. Blatt S J, Felsen I. “Different kind of folks may need dif- 32:453–477. ferent kind of strokes”: The effect of patients’ charac- 21. Davidson L. Living outside of schizophrenia. New teristics on therapeutic process and outcome. York: New York University, 2004. Psychother Res 1993;3:245–259. 22. Shahar G, Kidd S, Styron T H, Davidson L. Role of pa- 8. Elkin I. The NIMH Treatment of Depression Collabo- tients’ morale in satisfaction with partnership-based rative Research Program: Where we began and where mental health services. J Soc Clin Psychol, in press. we are now. In: Bergin A E, Garfield S L, editors. Hand- 23. Shahar G, Davidson L. Depressive symptoms erode book of psychotherapy and behavior change. 4th ed. self-esteem in severe mental illness: A three-wave, New York: Wiley, 1994: pp. 114–135. cross-lagged study. J Consult Clin Psychol 2003;71: 9. Weissman A N, Beck A T. Development and validation 890–900. of the dysfunctional attitudes scale: A preliminary in- 24. Blatt S J, Shahar G. Psychoanalysis: With whom? For vestigation. Paper presented at the 86th annual con- what? And Why? Comparison with psychotherapy. J vention of the American Psychological Association, Am Psychoanal Assoc 2004;52:393–447. Toronto, Canada, 1978. 25. Wallerstein R S. Forty-two lives in treatment: A study 10. BlattSJ,ZuroffDC,BondiCM,SanislowC,PilkonisP of psychoanalysis and psychotherapy. New York: A. When and how perfectionism impedes the brief Guilford, 1986. treatment of depression: Further analyses of the NIMH 26. Blatt S J. The differential effect of psychotherapy and TDCRP. J Consult Clin Psychol 1998;66:423–428. psychoanalysis on anaclitic and introjective patients: 11. ZuroffDC,BlattSJ,SotskySM,KrupnickJL,Martin The Menninger Psychotherapy Research Project revis- D J, Sanislow C A, Simmens S. Relation of therapeutic ited. J Am Psychoanal Assoc 1992;40:691–724. alliance and perfectionism to outcome in brief outpa- 27. Shahar G, Blatt S J. Benevolent interpersonal schemas tient treatment of depression. J Consult Clin Psychol facilitate therapeutic change: Further analysis of the 2000;68:114–124. Menninger psychotherapy research project. 12. ShaharG,BlattSJ,ZuroffDC,KrupnickJ,SotskySM. Psychother Res 2005;15:345–349. Perfectionism impedes social relations and response to 28. Stricker G. The local clinical scientist: A bridge be- brief treatment of depression. J Soc Clin Psychol 2004; tween science and practice. Am Psychol 1995;50:995– 23:140–154. 1002. 308 EVIDENCE-BASED PSYCHIATRIC PRACTICE? LONG LIVE THE (INDIVIDUAL) DIFFERENCE

29. McCulloughJP,KleinDN,KellerMB,HolzerCE,Da- depression and major depression superimposed on vis S M, Kornstein S G, Howland R H, Thase M E, Har- dysthymia (double depression): Validity of the distinc- rison W M. Comparison of DSM-III-R chronic major tion. J Abnorm Psychol 2000;109:419–427.

List of reviewers for Israel Journal of Psychiatry, 2007 The Editors would like to thank the following for their contribution as reviewers of manuscripts during 2007. Moshe Abramowitz Mimi Ajzenstadt George Anderson Katherine Appleton Alan Apter Eitan Avramovich Yoram Barak Dan Bar-On Jonathan Benjamin Haim Belmaker Elliot Berry David Blass Miki Bloch Yuval Bloch Danny Brom John Calamari Anna Chur-Hansen Robert Cohen Julie Cwikel Yaron Dagan Gary Donohoe Adiel Doron Howard Dubowitz Rimona Durst Robin Emsley Shmuel Fennig Tsvi Fischel Robin Friedlander Esti Galili-Weisstub Marc Gelkopf Jeffrey Geller Tsvi Gil Carlos Giloni Saralee Glasser Harvey Gordon RazGross DeborahHasin DeborahHasin Jeremia Heinik Uriel Heresco-Levy Dan Hoofien Iulian Iancu Richard Isralowitz Moshe Kalian Natan Kellermann Ehud Klein Ilana Kremer Antonio Lasalvia Dori Laub Peretz Lavie Arturo Lerner Vladimir Lerner Yaacov Lerner Itzhak Levav Joseph Levine Daphna Levinson Yechiel Levkovitz Pesach Lichtenberg Christine Lochner Kate Loewenthal Ido Lurie Ruth Malkinson John McGrath Shlomo Mendlovic Jonathan Meyer Ann Mortimer Jan-Willem Muntjewerff Eitan Nahshoni Igor Oyffe David Pearlman Alexander Ponizovsky Michael Poyurovsky Ilya Reznick Miriam Schiff Stanley Schneider Ronnen Segman Mujeed Shad Baruch Shapira Leo Sher Roni Shiloh Emi Shufman Efrat Shushan Moshe Halevi Spero Danny Stein Orit Stein Eyal Steiner Israel Strous Abraham Twersky Jerzy Vetulani Jerry Warsh Avi Weizman Eli Witztum Zvi Zemishlany Ari Zivotofsky Ada Zohar