Readers of the Journal of Applied Behavior Analysis (JABA) will find Van Hasselt and Hersen's Sourcebook of Psychological Treatment Manuals for Adult Disorders quite interesting reading, but will likely regard it with the ambivalence of a parent watching one's child doing well in life through slightly nefarious methods. Be prepared, then, for some mixed feelings as you read this highly recommended, very serious work. The list of contributing authors to the Sourcebook reads like a who's who of contemporary cutting-edge clinical psychology, psychiatry, and social work: Alan Bellack, Kelly Brownell, Paul Emmelkamp, Edna Foa, Michel Hersen, Michael Kozak, Robert Liberman, William Marshall, Vincent Van Hasselt, Robert Weiss, and Stephen Wong, to name the most prominent.
The structure of the Sourcebook is a series of chapters that provide fairly operational guidelines for the behavioral treatment of 17 common psychosocial problems: panic disorder and agoraphobia, obsessive-compulsive disorder, depression (two chapters!), schizophrenia, sex offending, sexual dysfunctions, losing weight (two chapters!), marital problems, and insomnia, among others. An introductory chapter by Ron Acierno and the editors is titled "Accountability in Psychological Treatment," and argues well the case that (a) we now do indeed have reasonably effective treatments for many psychosocial problems, (b) clinicians should be ethically obliged to provide these treatments as first-choice options in assisting clients, and (c) we should be expected to evaluate the treatment of clients using single-system research designs. So far, so good. Nor can one quibble about the selection of disorders; they are mostly common ones, important (no snake phobias here, or enuresis), and they all do indeed enjoy credible evidence of being treatable through behavioral methods. Were I to teach a graduate course on psychosocial treatments for adults, I think that this work would now be my first choice as a text. The second edition of Ammerman, Last, and Hersen (1993), which is in press at the time of this writing, will provide a nice complement (covering children and adolescents) to the present work.
Any reservations? Well, when I wear my behavior analyst cap (given to me some years ago by then-JABA editor Scott Geller), I recall that one of the foundations of our field is that of conducting an individual functional analysis of each client's problem. This approach, nicely outlined by Sturmey (1996), Van Houten and Axelrod (1993), and Bailey and Pyles (1989), among others, can be seen as being at odds with the more prescriptive approach taken by this and related books (e.g., Ammerman et al., 1993; Giles, 1993; Thyer & Wodarski, 1998).
Even some behavior therapists have reservations about such prescriptive approaches to treatment: Wolpe ... was increasingly critical of research which employed standardized protocols with people grouped into diagnostic categories. Not only was it different from the way therapy is practiced, threatening the external validity of research findings, but it detracted from the outcome of treatment in that procedures may be employed with individuals for whom they were ineffective or even counterproductive. (Poppen, 1995, p. 183) Take, as a clear case, the recommendation of exposure therapy and response prevention (ETRP) as a first-choice treatment for a client who meets the DSM-IV criteria for obsessive-compulsive disorder (OCD). That this is provided in a caring, respectful manner involving informed consent and a discussion of treatment options goes without saying, because these have always been hallmarks of behavioral practice. The careful recording of data - perhaps using standardized pencil-and-paper measures completed by the client, rating scales filled out by others, direct behavioral measurements - should be undertaken and the results, along with clinical judgment, used to guide treatment (e.g., sessions of more or less intensity, duration, or frequency). The data are fairly clear: ETRP is the best current method to help the greatest number to obtain the maximum benefit from treatment.
However, this result is obtained, to some extent, at the expense of a functional analysis. Suppose a problem like OCD has significant operant elements maintaining it, or panic disorder-like symptoms are due to a difficult-to-diagnose organic disorder such as pheochromocytoma. Is apparent school phobia caused by fear of school, fear of separation from parents, or Mom's history of giving in to the child's incessant pleas to stay home? Prescriptively supported cognitive-behavior therapy for depression focuses more on feelings and thoughts than on what is felt (a life filled with past or present aversive experiences?), and directs change efforts accordingly. At present, prescriptive approaches to treatment provide little latitude for such inquiries, and to that extent they are incompatible with (or at least incomplete as) behavior analysis. Perhaps in the coming decade, standardized treatment protocols will begin to incorporate functional analyses into their assessment and intervention processes. At present this is little evident.
My dilemma is also in part methodological. Group designs usually make use of convenient samples of clients (e.g., those obtained from one hospital or clinic). With a sufficient sample size, appropriate designs, perhaps involving random assignment to various treatment and no-treatment conditions, suitable statistical analysis and outcome measures, and so forth, one may make inferences about statistically significant changes within groups and differences between groups. Using effect sizes, one can make some judgments as to the meaning of any improvements, not just their statistical reliability. Various social validation techniques can help to determine factors such as the acceptability of treatments or their effects on quality of life. Through replication of such studies, one's confidence regarding the generalizability of findings is enhanced (recalling that generalizability through the use of a randomly selected representative sample drawn from the entire population of interest is almost never a practical possibility).
Given a reliable finding that is generalizable, what does this tell us? Given a large group of similar clients, it tells us how to maximize the likelihood that they will be helped. Inferences can be legitimately drawn from the sample back to the population, but never from the sample to an individual. But this latter reasoning is precisely the logic (my wording) behind the prescriptive treatment movement exemplified by the Sourcebook, that is, "We have a treatment that has worked well with a large group of clients with a particular problem. Therefore, when another individual client with this problem walks in the door, we will provide this empirically supported treatment."
Apart from common sense, I know of no statistical rationale for assuming that a treatment validated via group designs may be inferred to be effective with a new individual with a designated problem. Conclusions derived from group designs are logically valid only when applied to another group, not to a single person. Nevertheless, this prescriptive approach is the hallmark of medical research and is rapidly assuming dominance in clinical psychology and psychiatry.
Imagine Professor Henry Higgins (behavior analyst) brooding in his consulting rooms after Eliza Doolittle (now a successful behavior therapist) has left him. He has taught her all that he knows about using rigorous behavioral interventions, evaluation, and theory, and the ungrateful wretch has had the audacity to set up her own practice. Moreover, Eliza is doing supremely well, attracting clients in droves, as she dilutes the conceptual purity of his techniques with cognitive therapy, mental exercises, interpretation, self-report measures, and group research designs, all of which prove to be more attractive to the masses. Fewer and fewer clients consult Professor Higgins, leaving him and Colonel Pickering bickering endlessly about the failure of his practice to fulfill its earlier promise. Will Eliza return to Higgins? Not bloody likely, as long as she does so well on her own.
Perhaps, just perhaps, what would be useful would be some large-scale evaluation programs of behavior-analytic treatment, using both single-subject and group designs and the visual analysis of graphically presented individual data plus nomothetic statistical analysis. Use rigorous direct behavioral measures plus the pencil-and-paper scales so favored by the behavior therapists. Let us examine individually designed behavior-analytic treatment derived from a functional analysis versus so-called cognitive-behavior therapy delivered via a standardized protocol. If these are truly effective treatments, positive effects should be detectable using both nomothetic and idiographic strategies. And comparative evaluations of treatment effectiveness should be possible. If behavior analysis is as powerful as we believe it is, then we should be players in the great game of randomized controlled clinical trials (RCTs). RCTs are the coin of the realm with respect to contemporary evidentiary standards of proof. Perhaps, just perhaps, then, the second or third edition of Van Hasselt and Hersen's Sourcebook of Psychological Treatment Manuals for Adult Disorders will contain more behavior analysis.
A small quibble: Although most of the authors are psychologists, the occasional psychiatrist and social worker are also present. Given the seminal role of clinical researchers from nonpsychological disciplines in the development and validation of many of these treatments and the nature of these problems, I would have preferred the title of the sourcebook to read Psychosocial (rather than Psychological) Treatment Manuals. The term psychosocial is more etiologically and descriptively accurate and avoids the impression of turf protection engendered by using the term psychological.
REFERENCES
Van Hasselt, V., & Hersen, M. (Eds.).
(1996). Sourcebook of psychological treatment manuals for
adult disorders. New York: Plenum. ISBN 0-306-45144-1, 720 pp.
Correspondence may be addressed to B. A. Thyer, School of
Social Work, University of Georgia, Athens, Georgia 30601
(E-mail: Bthyer@uga.cc.uga.edu).