Psychotherapy - MDFT · in psychotherapy outcome research as a means for testing whether...

Psychotherapy Volume 33/Summer 1996/Number 2

TREATMENT ADHERENCE PROCESS RESEARCHIN FAMILY THERAPY: A RATIONALE AND

SOME PRACTICAL GUIDELINES

AARON HOGUETemple University

HOWARD A. LIDDLEUniversity of Miami School of Medicine

CYNTHIA ROWETemple University

Treatment adherence research hasrecently established a permanent nichein psychotherapy outcome research as ameans for testing whether interventionshave been implemented as intended.Advanced-level adherence methods allowinvestigators to move beyond treatmentintegrity questions regarding modelfidelity and toward treatment processquestions regarding therapeutictechnique and intervention dosage.Though still in the developmental stage,treatment adherence process proceduresappear to be congruent with the methods,goals, and theoretical framework thatcharacterize contemporarypsychotherapy process research.Because adherence process research is

Preparation of this article was supported by grants to How-ard A. liddle from the National Institute on Drug Abuse(Grant Nos. P50-DAO7697 and T32-DAO7297) and was com-pleted in conjunction with a NIDA-sponsored postdoctoralfellowship awarded to the first author.

Correspondence regarding this article can be addressed toeither of the first two authors: Aaron Hogue, Center for Re-search on Adolescent Drug Abuse, Temple University, 13thStreet and Cecil B. Moore Avenue, 3rd floor Weiss Hall (TU265-66), Philadelphia, PA 19122, internet address [email protected]. Howard Liddle, Center for Fam-ily Studies, Department of Psychiatry and Behavioral Sci-ences, University of Miami School of Medicine, 142S N.W.10th Avenue, 3rd floor, Miami, FL 33136, internet [email protected].

virtually absent from the family therapyresearch literature, practical guidelinesare presented for conductingobservational-based adherence researchon family therapy models, using theexample of Multidimensional FamilyTherapy.

There is pleasant irony in the fact that LewisCarroll's Dodo bird, whose species is remarkablefor having been haplessly unfit for survival, isthe unlikely progenitor of a robust legacy in psy-chotherapy research. Over two decades ago Lu-borsky, Singer, and Luborsky (1975) challengedthe psychotherapy community by concluding thatdifferent models of psychotherapy had demon-strated little or no differential effectiveness intreating psychological problems. This and subse-quent reviews of psychotherapy outcome studies(Shapiro & Shapiro, 1982; Smith & Glass, 1977;Smith, Glass, & Miller, 1980), particularly re-views of efficacy studies mat directly comparedifferent models of treatment (Lambert & Bergin,1994), revealed that competing models by andlarge produced relatively equivalent gains in cli-ent well-being. In effect, the Dodo bird verdicthad been reached: "Everyone has won and allmust have prizes" (Luborsky et al., 197S). Thisinglorious judgment troubled members of all ther-apy camps in a manner reminiscent of Eysenck's(1952) assertion that psychotherapy fares no bet-ter than spontaneous remission in improving cli-ent symptoms. Eysenck's controversial findingsinitiated the clarion call for a generation of inves-tigators to employ more rigorous, methodologi-cally sound research standards in the cause of

332

Treatment Adherence Process Research

proving that psychotherapy has substantive bene-ficial effects (Brown, 1987; Hill & Corbett,1993). In similar fashion, Luborsky's verdict hasmarshalled a groundswell of research effortsgeared toward establishing that therapeutic mod-els have intervention-specific, distinguishable ef-fects on client outcome (Stiles, Shapiro, & El-liott, 1986).

Efforts to demonstrate differential clinical ef-ficacy, and thereby overturn the Dodo verdict,include a diverse range of pursuits: improvedcomparative efficacy research (Basham, 1986;Kazdin, 1986), aptitude-treatment interactionstudies (Beutler, 1991; Shoham-Salomon & Han-nah, 1991), and psychotherapy process studiesattempting to link process and outcome(Greenberg & Pinsof, 1986; Orlinsky, Grawe, &Parks, 1994). Notable among these developmentshas been the growing interest in treatment adher-ence research. Treatment adherence research re-fers to the methodological strategies used to docu-ment that a given therapy has genuinely beencarried out in accordance with essential theoreti-cal and procedural aspects of the model. As theprimary means for ascertaining whether therapistsimplement therapeutic models as intended, adher-ence research is an essential element of clinicalefficacy research (Kazdin & Bass, 1989). Al-though it would seem prerequisite to prove thata therapy has occurred before making statementsabout its effects, treatment adherence is a relativenewcomer to the efficacy research scene. Kazdin,Bass, Ayers, and Rodgers (1990) completed anexhaustive review of 233 studies of child andadolescent psychotherapy and found that only19% reported efforts to monitor treatment imple-mentation, provide ongoing supervision, or evalu-ate therapist performance. Reviews of treatmentoutcome studies on other populations have uncov-ered similarly infrequent reports of treatment ad-herence (Moncher & Prinz, 1991; Yeaton &Sechrest, 1981).

Despite this slow start, treatment adherence hasestablished a permanent niche within treatmentoutcome research. To this point, adherence proce-dures have been utilized in the service of tworelated agendas. First, they verify the level oftreatment fidelity for clinical efficacy trials(Moncher & Prinz, 1991). Adherence researchprovides a manipulation check of the independentvariable, testing whether treatments are imple-mented as intended and whether competing treat-ments are sufficiently different from one another.

Second, adherence procedures are invariably uti-lized to monitor whether manualized therapymodels are executed in accordance with docu-mented protocols. In fact, the recent demand foradherence checks can be attributed in large mea-sure to the institution of manualized treatment asthe standard for achieving purity, consistency,and replicability in delivering treatment models(Binder, Strupp, Bongar, Lee, Messer, & Peake,1993; Dobson & Shaw, 1993; Luborsky & Bar-ber, 1993; Luborsky & DeRubeis, 1984). Adher-ence procedures are natural companions to treat-ment manuals in that they are well-suited forverifying the practice of prescribed therapist tech-niques. Whether treatment fidelity or manual veri-fication is the goal, treatment adherence researchoffers a fundamental yardstick of therapeutic ac-countability (Yeaton & Sechrest, 1981). Theseself-imposed "watchdog" methods are fast be-coming required elements of counseling research(Kazdin, 1994). In addition, they represent a stan-dard for monitoring clinical practice that is com-pletely compatible with the quality-of-care per-spective emphasized by managed care healthorganizations (Clarke, 1995; Herron, Eisenstadt,Javier, Primavera, & Schultz, 1994; Newman &Howard, 1986).

With due respect to this established niche, thisarticle focuses on new possibilities emerging onthe horizons of treatment adherence research.First, we review common methods of conductingadherence research and demonstrate mat advanced-level procedures, which yield clinically rich infor-mation about therapist behavior in session, arecongruent with the methods and goals that charac-terize traditional psychotherapy process research.Second, we provide both a theoretical frameworkand some practical recommendations for conduct-ing advanced-level adherence research with fam-ily therapy models. Although there has been arecent surge in outcome research involving familytherapies (Pinsof & Wynne, 1995a), there hasbeen comparatively little utilization of adher-ence procedures in family therapy research and(to our knowledge) no published works that fo-cus on adherence methodology for family-based models.

Methods of Measuring Treatment AdherenceThere is presently no widely accepted tech-

nique for conducting treatment adherence re-search. On the contrary, there is great hetero-geneity in the methods investigators use to verify

333

A. Hogue, H. A. Liddle & C. Rowe

treatment adherence, and every approach has dis-tinctive advantages and drawbacks (see Waltz,Addis, Koerner, & Jacobson, 1993, for a compre-hensive review). Nevertheless, a sufficient num-ber of studies reporting adherence findings hasaccumulated to identify three general categoriesof common adherence methods. The first categoryof adherence methods consists of "quality con-trol" procedures exercised prior to and concur-rently with ongoing treatment provision. Exam-ples of quality control procedures includeextensive training in the model (e.g., Shelter,Dasberg, & Ben-Shakhar, 199S), weekly supervi-sion with an expert in the model (Simons, Gor-don, Monroe, & Thase, 199S), and requirementsfor completing detailed progress notes (Kazdin,Bass, Siegel, & Thomas, 1989). A more ad-vanced example of quality control is provision ofa training with a treatment manual and regular"therapist drift" booster sessions in manual proto-col (Henggeler, Melton, & Smith, 1992). Qualitycontrol procedures, although sometimes reportedas "integrity checks," are better characterized askey ingredients for conducting psychotherapy inan efficient and effective manner. They are akinto adherence insurance, in that they proactivelystrengthen the likelihood that therapists will ad-here to treatment models.

The second category of common adherencemethods contains procedures that involve cata-loguing therapist-in-session behavior in an effortto verify adherence along specified dimensions.Often, persons other than the therapist are trainedto read case notes and cull data related to identi-fied dimensions of model-congruent and model-discordant therapist behavior. Specific examplesinclude reviewing client contact logs to rate thera-peutic engagement efforts according to a pre-established continuum of intensity (Szapocznik,Perez-Vidal, Brickman, Foote, Santisteban,Hervis, & Kurtines, 1988), and summarizing do-mains of systemic activity for a model that pre-scribed multidomain intervention (Borduin,Mann, Cone, Henggeler, Fucci, Blaske, & Will-iams, 1995). An interesting extension of theseprocedures are therapist self-reports of adherence-related activities (Carroll, Nich, & Rounsaville,in press; Carroll & Nuro, in press). Keeping inmind demonstrated biases in therapist recall ofsession events (Chevron & Rounsaville, 1993;Lambert & Hill, 1994), adherence self-reports arestandardized forms comprised of rating scales for

significant prescribed and proscribed behaviors.The scales are completed after every session, andtherapists are trained and monitored specificallyin post-hoc assessment of the purity of their inter-ventions. In all, cataloguing procedures providean after-the-fact review of certain aspects of treat-ment, and they potentially offer a unique (andcost-efficient) perspective on session events (Car-roll, Nich, & Rounsaville, in press).

The third category of adherence methods hasthe virtue of involving observational review oftherapist conduct during treatment sessions.These procedures capture the best intentions oftreatment adherence research: non-participantconfirmation that therapists have practiced inter-ventions according to the letter and spirit of themodel. That said, a wide variety of approacheshas been used to observe and measure therapistbehavior, and they differ from one another alongseveral dimensions: identity of the raters, numberof observations sampled, method of observation,and so forth. For example, many studies (Ken-dall, 1994; Paivio & Greenberg, 1995) ask clin-ical supervisors to review randomly selected au-diotapes of their supervisees to look for"obvious violations" of the model. Family ther-apy studies commonly use live supervision(Montalvo, 1973) as a form of ongoing adher-ence monitoring. In live supervision, the super-visor observers a session as it happens and inter-venes when the therapist needs help or is offtrack in implementing the approach. Some fam-ily therapy studies also incorporate a treatmentmanual or other training material into the con-text of regular live supervision (Liddle, Becker,& Diamond, in press).

A primary strategy within the category of ob-servational review adherence methods entailstraining non-participant raters to recognize a ros-ter of intervention techniques and then to codeaudio or videotapes of randomly selected ses-sions. Raters are typically kept blind to therapistidentity, session number, and the specific goalsof the study (for representative samples of thisapproach, see DeRubeis, Hollon, Evans, &Bemis, 1982; Henry, Strupp, Butler, & Schacht,& Binder, 1993; Hill, O'Grady, & Elkin, 1992).This strategy introduces the highest levels of rigorand reward in adherence research. Admittedly,this resource-intensive approach demands a greatinvestment of time and energy from researchers.It is difficult to train raters to recognize sophisti-

334


cated clinical interventions, and because the cod-ing process is both arduous and time-consuming,rater attrition can be high (Hill, 1991). In compen-sation, researchers obtain adherence informationthat is highly non-subjective, thorough, anddetail-specific with regard to how therapists actu-ally perform interventions in session.

As evidenced by the diversity of methods tohave appeared on the scene since treatment integ-rity became a routine concern of efficacy studies,adherence research is a progressing mini-sciencethat takes many forms. Treatment adherence pro-cedures were initially developed to monitor treat-ment fidelity, and the core task of adherence re-search remains providing an answer to thequestion: "Did the therapy occur as intended?"Even so, this deceptively straightforward ques-tion invariably opens the door to complicatedjudgments about model specification and "goodenough" model compliance (Kazdin, 1986; Waltzet al., 1993). How much protocol violation issufficient to disqualify a session, a therapist, ora clinical trial? How much prescribed behavior,in what combinations, is required for satisfyingadherence criteria? Issues such as these remindus that more rigorous and complex adherence pro-cedures yield more elaborate information aboutmodel implementation. In the following section,we describe how advanced-level, observationalreview techniques of adherence research offer ameans to stretch beyond the fidelity-based ques-tion "Did the therapy occur as intended?"toward the process-based question "What ex-actly occurred in the sessions?" Observationalreview techniques thereby create the opportu-nity for double vision in adherence researchgoals: a baseline judgment regarding adequatetreatment integrity (general adherence research),and a more comprehensive and clinically-rich perspective on therapeutic interventions andtherapist technique (adherence process research)that is the hallmark of traditional psychotherapyprocess research.1

' In the remaiwler of this article, for the sake of simplicityand focus, we concentrate on observational review techniquesas the primary example of "advanced-level'' adherence andprocess research. However, we recognize that this limitedfocus excludes the highly-developed body of process researchthat involves therapist and client recall and review of sessionevents (e.g., Hill, 1994). Moreover, progress in therapist self-report adherence procedures may eventually produce signifi-

Evolving Congruence between AdherenceProcedures and Process Research

The technology of adherence research is morecomplex man ever before. Advancements havebeen made in the methods used in adherence re-search, and with these advancements have comeopportunities for answering process-based ques-tions about clinical efficacy. Adherence processresearch refers to adherence procedures that uti-lize multivariate, observational review methodsto investigate therapeutic interventions. In thissection we outline three dimensions along whichadherence process research is syntonic with con-ventional psychotherapy process studies: meth-ods, goals, and framework.

First, in determining the methods for a givenadherence study, investigators must make severalresearch design decisions that irrevocably shapethe nature of the data collected and the questionsaddressed. Fortunately, adherence process inves-tigators who select observational-based codingsystems can call upon an extensive body of previ-ous process research that details the advantages,disadvantages, and pitfalls of each designchoice. Specifically:

What will be coded? Selecting circumscribedsegments of therapy sessions rather than wholesessions may preclude observation of meaningfulbehavioral dimensions (Mintz & Luborsky,1971), and several alternatives exist for seg-menting sessions into units of observation (Rus-sell & Staszewski, 1988; Stinson, Milbrath, Re-idbord, & Bucci, 1994).

Who shall code? Non-participant judges repre-sent a distinct, perhaps orthogonal category ofobservation from that of therapists or clients whoare asked to recall their own behavior (Lambert &Hill, 1994). Recent advances in process researchhave included asking therapists and clients to re-view and code tapes of their sessions (Hill, 1994).

What kind of coding system will be used? Op-tions for selecting an adherence coding systemcover the full gamut of observational-based tech-nology: simple occurrence versus non-occurrenceof an intervention, frequency counts, event-by-event prospective coding, or Likert-type quantita-tive ratings of intervention extensiveness (Waltz

cant contributions to treatment development, component anal-ysis, and process-outcome linking research (Carroll, Nich, &Rounsaville, in press).

335


et al., 1993). There is general consensus that mi-croanalytic systems (targeting small units overshort periods with minimal inference) are moresensitive to contingent and situational influences,whereas macroanalytic systems (targeting largerunits over more time with multidimensional infer-ences) better predict cross-situational influencesand global outcomes (Mien, Markman, & Lin-dahl, 1989; Markman, Leber, Cordova, & St.Peters, 1995). In sum, observational-based adher-ence research confronts the same methodologicalchallenges faced by traditional process researchin devising process coding schemes (Alexander,Newell, Robbins, & Turner, 1995; Hoyd, 1989;Hill, 1991).

Second, adherence process research sharesgolds that are identical to those championed bypsychotherapy process researchers. Foremostamong these goals, and the raison d'etre of pro-cess research from its earliest conceptions, is theanalysis of what actually takes place in therapysessions (Hill & Corbett, 1993; Shoham-Salomon, 1990). A handful of recent adherencestudies have admirably responded to the challengeof describing therapist in-session behaviors in adetailed, multivariate manner. For example, Hill,O'Grady, and Elkin (1992) analyzed adherenceratings from the NIMH Treatment of DepressionCollaborative Research Program (TDCRP; Elkin,Parloff, Hadley, & Autry, 1985) in order to dis-cern therapist techniques that were unique to, andshared among, cognitive-behavioral, interper-sonal, and clinical management approaches. Fur-thermore, they presented the manner in whichvarious interventions tended to cluster togetheracross modalities, therapists, clients, treatmentsites, and treatment phases, as well as across vari-ous combinations of these variables. Adherenceresearch generated from the Sheffield Psychother-apy Projects (Shapiro & Startup, 1992; Startup& Shapiro, 1993) compared prescriptive versusexploratory psychotherapies using a rare designfeature: the same pool of therapists delivered bothforms of treatment. This research explored thera-pist intervention patterns by stage and durationof therapy and examined both between-modalityand within-modality adherence effects. TheTDCRP and the Sheffield adherence studies areexemplary in their empirically-grounded explora-tion of in-session techniques utilized in differentcombinations and at different phases of treatment.

The primary goal laid out by contemporary pro-cess researchers is establishing links between

treatment process and outcome (Garfield, 1990;Greenberg & Pinsof, 1986; Mannar, 1990). Here,too, adherence process research can contribute tothe overall process agenda. For example, Lubor-sky, Woody, McLellan, O'Brien, and Auerbach(1985) found a significant positive relation be-tween treatment "purity" (adherence to intendedtechniques in exclusion of other techniques acrossthree treatment modalities) and client outcomeon drug use and psychological status variables.DeRubeis and Feeley (1990) employed itemsfrom the TDCRP cognitive-behavioral adherencescale to compare the success of a cluster of "con-crete" versus "abstract" cognitive interventionsin predicting therapeutic alliance and amount ofsymptom reduction evidenced over a 12-week pe-riod of psychotherapy. This study also providesan important example of how adherence processmethods yield valuable, and perhaps unique, dataon therapy process. Non-participant raters com-pleted macroanalytic coding scales consisting ofLikert-type items designed "to assess the extentor amount of therapist behavior of interest" (De-Rubeis & Feeley, p. 473; emphasis added). Inso doing, the investigators were able to addressquestions related to the quantity, or dosage, oftherapist interventions, and further to examinehow various therapist interventions predicted ormoderated other process and outcome variables.

This particular feature of adherence process re-search, the ability to assess the extent to whichspecific (prescribed) interventions are utilized insession, holds the greatest promise of finding aniche for adherence-focused research within theoverall framework of the psychotherapy processresearch agenda. Specifically, an emergingframework of process analysis on therapist inter-vention strategies might be supported by threepillars: adherence, competence, and alliance re-search.2 Each of these pillars has its own legacyand expanding roster of studies in the processfield that cannot be justly described here. Brieflyand simply stated, in this framework "adherence"alludes to what interventions are used and to whatextent, "competence" refers to how well or inwhat manner they are used, and "alliance" con-cerns what relationship conditions exist withinwhich they are used. In this overall process frame-

2 This recalls and updates Schaffer's (1982) three dimen-sions of therapist behavior for predicting treatment outcome:type, skillfulness, and interpersonal manner.

336


work adherence retains its conventional meaningwhenever (as is typically the case) a specificmodel or prescription of therapist behaviors isintended. But, even more broadly, adherence re-fers to both (1) the techniques or categories ofinterventions implemented by the therapist, whichis a central component of any process study(Greenberg, 1986), and (2) the extensiveness ordosage of these interventions in session (Elkin,Pilkonis, Docherty, & Sotsky, 1988; Sechrest,1994). Competence, in contrast, refers to thequality of the intervention. Therapist competenceis a measure of the skill with which interventionsare delivered, such that relevant aspects of thetherapeutic context are taken into account (Waltzet al., 1993) and the interventions are maximallyeffective in the moment or in the session (Schaf-fer, 1983). Barometers of competence includetherapist knowledge of client issues, the appropri-ateness and timing of interventions, and the de-gree of therapist responsiveness to client in-session behaviors (O'Malley, Foley, Rounsaville,Watkins, Sotsky, Imber, & Elkin, 1988; Shaw &Dobson, 1988; Stiles & Shapiro, 1994; Yeaton& Sechrest, 1981). Finally, therapeutic allianceis a multidimensional construct describing thegoals, tasks, and relational bond that comprisethe working relationship between therapist andclient (Bordin, 1978, 1994; Gaston, 1990; Hor-vath, 199S). Aspects of therapist behavior, clientbehavior, and their interactional dynamics allcontribute to determining the quality of the alli-ance (Horvath & Symonds, 1991; Pinsof, 1994),which in turn, shapes the relationship context ofthe therapy and thereby colors all in-session be-haviors (Horvath & Luborsky, 1993).

Multifaceted, integrative frameworks for as-sessing psychotherapeutic process are requiredfor tackling the notoriously difficult task of Unk-ing process to outcome (Greenberg, 1986; Stiles& Shapiro, 1994). We have argued that treatmentadherence process research can make a valuablecontribution to both pure process and process-outcome research. Next, we describe our effortsto apply adherence process procedures in familytherapy research.

Treatment Adherence Process Research hiFamily Therapy: Process

Despite some of family therapy's most notableearly advancements having occurred in the re-search arena (Pinsof & Wynne, 19956), familytherapy research is, in many ways, a relative new-

comer to the clinical efficacy game (Lebow &Gurman, 1995). With its roots planted in the para-digms of systemic and interactional theories aboutpsychotherapy that flowered less than thirty yearsago (Bateson, 1972; Watzlawick & Weakland,1977), family therapy has only recently plungedfull-blown into the research culture of empiricalspecification and validation (Coyne & Liddle,1992), meta-analytic status review (Hazelrigg,Cooper, & Borduin, 1987; Shadish, Ragsdale,Glaser, & Montgomery, 1995), process-outcomestudies (Diamond & Liddle, 1996; Friedlander,Wildman, Heatherington, & Skowron, 1994),and disorder-specific clinical efficacy trials (Es-trada & Pinsof, 1995; Liddle & Dakof, 1995;Prince & Jacobson, 1995). Moreover, the tran-sition to empirically-focused agendas continuesto meet resistance in some quarters (Liddle,1991; Shields, Wynne, McDaniel, & Gawinski,1994). For these reasons, family therapy re-search on adherence is less advanced than inthe individually oriented therapies, and workto specify guidelines and standards is needed(Mann & Borduin, 1991).

In those instances when adherence has beenaddressed, observational review methods haverarely been attempted. Although it has long beencustomary to describe a given model's theoreticalunderpinnings and training procedures (e.g.,Alexander, Barton, Schiavo, & Parsons, 1976),non-participant review of in-session therapist be-havior remains the exception (e.g., Bry, Conboy,& Bisgay, 1986; Friedman, 1989) rather than therule (e.g., Friedman, Tomko, & Utada, 1991;Henggeler et al., 1992; Joanning, Quinn,Thomas, & Mullen, 1992; Romijn, Platt, &Schippers, 1990). To date, the most rigorous ad-herence research has been produced by Szapocz-nik, Rio, Murray, Cohen, Scopetta, Rivas-Vasquez, Hervis, Posada, and Kurtines (1989) intheir comparison of structural family therapy andpsychodynamic child therapy. They employedtrained raters to view videotapes randomly se-lected from three phases of treatment and, at 10-minute intervals, complete a frequency count offamily-identified and psychodynamic-identifiedinterventions. Their goal was to establish adher-ence thresholds for each modality—in their case,at least 75% of observed interventions being thoseprescribed for that model. The Szapocznik et al.study is one of the first to demonstrate that, likeother paradigms, family therapy research canachieve the highest standards of treatment adher-

337


ence documentation and, further, successfully en-gage in adherence process research.

In our own work with a multisystemic, family-based intervention model, MultidimensionalFamily Therapy (MDFT; Liddle, 1995; Liddle,Dakof, & Diamond, 1991; Liddle & Diamond,1991; Schmidt, Liddle, & Dakof, 1996), we arein the final stages of a five-year comparative effi-cacy trial for treating adolescent substance abuse.Faced with the challenge of demonstrating bothtreatment integrity and treatment differentiation(Moncher & Prinz, 1991) from an individual-based cognitive-behavioral model, we developedan adherence process rating scale (Hogue, Rowe,Liddle, & Turner, 1994a) designed to identifycore clinical interventions within each modality.Based on our experience with this adherence scalewe present guidelines for conducting adherenceprocess research in general and for evaluatingfamily therapy modalities in particular.

First, as with any psychotherapy coding ven-ture, it is imperative that researchers present anexplicit rationale for their unitizing and proce-dural choices (Alexander et al., 1995; Ruseell &Staszewski, 1988). In this case, we were guidedat every juncture by the wish to conductobservational-based, quantitative, minimally bi-ased adherence research that would serve the dualpurpose of establishing integrity thresholds andproviding process-level data on therapist interven-tions. In order to hit these marks, (1) we recruited"naive" undergraduate psychology majors toserve as judges who were kept blind to the natureof the treatment models and the goals of the study,intensively trained over a five-month period toreach adequate reliability, and reconvened on aweekly basis for supportive training and to pre-vent rater drift; (2) we selected a 7-point Likert-type coding system for rating the extensivenessof therapeutic interventions; (3) the scale was con-structed with some items intended to be uniqueto each modality, some common to both, andsome applicable to almost any form of psycho-therapy; and (4) judges viewed videotapes ofwhole therapy sessions.

Second, we closely modeled our scoring man-ual (Hogue, Rowe, Liddle, & Turner, 19946)after the TDCRP Rater's Manual for the Collabo-rative Study Psychotherapy Rating Scale-Form 6(Evans, Piasecki, Kriss, & Hollon, 1984; see alsoHollon, Evans, Auerbach, DeRubeis, Elkin,Lowerry, Kriss, Grove, Tuason, & Piasecki,1988). The general format of our scoring manual,

and many of the recommendations presented be-low, were based on the ground-breaking TDCRPtreatment adherence project.

Our adherence scale contains four theory-driven categories of therapeutic interventions:(1) Items unique to MDFT (e.g., Arranges,coaches, and processes multiparticipant interac-tions; Shapes parenting practices of functionedparents); (2) Items unique to the individual ther-apy model (e.g., Explores tacit beliefs that under-lie and organize client behaviors; Utilizesprotocol-driven behavioral techniques or relax-ation exercises); (3) Items common to both mod-els (e.g., Encourages client to experience andexpress affect in session; Helps client develop afuture-based orientation); (4) "Non-specific"items prescribed by virtually every therapeuticmodel (e.g., Probes for client's unique point ofview; Responds to client in warm and compas-sionate manner).3 One primary goal of our adher-ence study is demonstrating that therapists utilizeappropriate levels of model-unique interventions.However, in keeping with the spirit of adherenceprocess research, we are equally interested in thecommon and non-specific items. Our agenda istwo-fold: Above and beyond a model's tendencyto promote signature interventions, what nuancesin delivering the model give rise to alternative(perhaps unexpected) strategies? And which pat-terns or clusters of interventions collectivelyemerge as most prominent with the given popula-tion? Questions like these are concerned with thespecific ingredients of therapy process that revealsubtleties of how a theory and therapy model areactualized in practice (Elkin et al., 1988; Stiles& Shapiro, 1995). Further, as testimony to thefast pace of development of this specialty area infamily therapy, process-based questions of this

3 In the process of developing this scale we discovered anunforeseen and, in our case, unresolvable confound: Every"family therapy" item involves more than one member of theclient system. That is, all items uniquely prescribed in theMDFT condition require judges to rate therapist behavior thatby definition occurs either in conjoint sessions or in prepara-tion for conjoint sessions. Of course, it is certainly not truethat conjoint sessions are a prerequisite for family therapyinterventions. Many methods are available for conductingfamily therapy through one-person sessions (see Szapocznik,Kurtines, Perez-Vidal, Hervis, & Foote, 1990). Moreover,some individually-focused interventions (e.g., exploringfamily-of-origin issues) are more common in family-basedmodels than in certain individual therapies.

338


nature are receiving great attention in contempo-rary family therapy research (Friedlander, Wild-man, Heatherington, & Skowron, 1994; Heng-geler, Borduin, & Mann, 1993; Pinsof, 1989;Shadish, Montgomery, Wilson, Wilson, Bright,& Okwumabua, 1993).

We now offer eight guidelines for conductinga treatment adherence process research project.These have served as procedural blueprints fordesigning the study, practical instructions fortraining raters, and scoring reminders to compan-ion the rating of videotapes. Note mat theseguidelines apply primarily to constructing adher-ence items; substantially different procedures aretypically utilized for constructing items intendedto measure various forms of rater bias (see Hill,O'Grady, & Price, 1988). The first five guidelinesare general, while the last three pertain specifi-cally to family therapy adherence work:

Rate therapist behavior only. Treatment adher-ence scoring is geared exclusively toward thera-pist behavior, not the behavior of clients. There-fore, judges should code only the statements andactions of the therapist, and only these count asjustification for any particular score. This guide-tine increases accuracy and reliability. Equallyimportant, it inhibits consideration of therapistresponsiveness to client behavior, which falls un-der the domain of therapeutic competence.

Score "is," not ought." Judges should scoreonly what the therapist actually does in session,not what might have been done or should havebeen done. That is, there is no "negative scoring"for missed opportunity, incompleteness, or di-luted effort. A logical extension of the previousguideline, this caveat discourages judges frominferring therapist intention, assessing therapeuticcontext, or weighing therapist responsiveness(again, competence rather than adherence con-cerns).

Extensiveness — Thoroughness + Frequency.Interval-level ratings of adherence are anchoredon a scale that measures the extensiveness towhich a given intervention is utilized in session.Extensiveness constitutes two separate dimen-sions of therapist behavior (1) thoroughness, de-fined as the persistence and intensity with which asingle intervention is executed, and (2)frequency,defined as the number of instances in which agiven intervention occurs. High ratings may begiven for a powerfully focused intervention thattakes place during a brief segment, for an inter-vention that arises consistently throughout the

session but is not pursued at length in any episode,or both. Unfortunately, it is not possible to pro-vide a precise algorithm for blending judgmentsof thoroughness and frequency. Judges must relyon item descriptions in the coding manual, train-ing, familiarity with the adherence scale, and ex-perience in coding tapes in order to determinehow much weight to assign each dimension inscoring extensiveness. For example, in our studyEncourages client to experience and express af-fect in session tends to be a thoroughness-weighteditem, in that therapists from both modalities occa-sionally promote an extended interaction around anemotional theme but do not persistently tie interac-tions to emotional themes throughout a given ses-sion. On the other hand, Probes for clients uniquepoint of view is usually scored according to thefrequency with which a therapist specifically"checks in" for a client's reaction or opinion.

Emphasize item exemplars and item distinc-tions. Each item described in the manual shouldbe accompanied by "exemplar" therapist state-ments. Exemplars are brief therapist speech actsthat help judges recognize interventions by cuingthem to emblematic content and style for a givenitem. Exemplars are not intended to capture itemextensiveness (which is too context-specific andcumbersome to describe in a useful manner); in-stead, they are prototypes of the spectrum of be-haviors represented by each item. In addition,manuals should devote considerable space to itemdistinctions. Item distinctions are narrative de-scriptions that relate how a given item can bedistinguished from other items that contain sim-ilar content or tend to occur in conjunction withthe target item. Distinctions should also includecontrasting exemplars to act as concrete refer-ences for item differentiation.

Buffer raters against halo effects. Halo effectsrefer to situations wherein a rating for a givenitem is biased by a rating awarded to another itemor by a global judgment of the session as a whole.Halo effects come in many shapes and sizes, de-pending on the goals and design of the study.We caution our judges about these in particular,giving high ratings on every item because it isdecided that a therapist is generally competent;allowing one impressive intervention to generate"positive spillover" into ratings of similar items;intentionally deciding that, or unintentionallyscoring as if, two items naturally covary; andforming a negative impression partway through asession and thereby giving insufficient weight to

339


later behavior. In sum, at every sitting judgesmust remember that each item is to be rated as aseparate, independent entity, and its rating shouldnot be influenced by ratings given to any otheritem (unless specified otherwise in the manual).Of course, reliable item independence is possibleonly if the scale and the manual are planfullyconstructed with this in mind.

Score vicarious interventions. Vicarious inter-vention refers to the process by which a therapistintervenes with one participant through another,or intervenes with multiple participants by orches-trating their interaction. For sessions with multi-ple participants, items should be scored not onlywith regard to what therapists do with each per-son, but also with regard to what therapists en-courage participants to do with one another. Forexample, in a session with a mother and son, atherapist might instruct the pair to have an ex-tended conversation about past disappointments,or coach the mother in helping the son plan moreeffective solutions to a recurring problem. Inthese cases, the therapist can receive high marksfor encouraging affect or teaching problem-solving skills, even though the clients are chieflycarrying out the intervention. In short, judgesshould account for the manner in which therapistsencourage, facilitate, and comment upon the in-teractions of others in conjoint sessions.

Incorporate "to whom" into appropriate items.For certain items, the target of the interventiondetermines the classification of the intervention.For example, our scale contains two similar items:Shapes parenting practices of functional parentsand Teaches client new coping skills. When thera-pists address the problem-solving, coping, orcommunication skills of adults specifically in therole of parenting their children, the former isscored. However, when a therapist is workingwith a parent to improve these skills in otherdomains (e.g., communicating with a spouse),the latter is scored, even if the specific skill ad-dressed (e.g., active listening) is identical. Manypopular family therapy interventions (e.g., re-structuring intergenerational coalitions, strength-ening parental alliance) are conditionally identi-fied with the targets) of the behavior.

Keep up the appearance of item generalizabil-ity. In designing adherence scales to differentiatefamily-based treatments from other modalities, itis important mat the content of items does notimmediately betray their "allegiance" to a specificmodel; otherwise, the internal and construct va-

lidity of the study may be compromised. This isa formidable challenge. When comparing familyversus group treatments, items can refer to "multi-ple participants" rather than "family" or "group."In our project's comparison of family-based andindividual-based treatments, we take two precau-tions: (1) judges are not informed that two differ-ent models are being assessed, and (2) judges areinstructed that family involvement varies ac-cording to the contingencies of each case. In ourfamily-based treatment (MDFT), therapists spenda great deal of time working with individualsalone, and individual sessions routinely focus onintrapersonal experiences of the client. Similarly,our individual-based model promotes interven-tions mat focus on concerns related to significantothers (e.g., helping the adolescent process nega-tive experiences with a parent at home). Never-theless, given the experience that judges accumu-late in identifying interventions and recognizingtherapists over the course of an adherence study,some rater insights (hence biases) are unavoid-able.

Treatment Adherence Process Research hiFamily Therapy: Outcome

Contemporary process researchers agree matidentifying the types of interventions used in suc-cessful psychotherapies, particularly interven-tions empirically linked with favorable treatmentoutcome, will be a principal concern of futureprocess studies (Greenberg, 1995; Hill & Corbett,1993). We predict that adherence process re-search will become a staple analytic tool for thesestudies. Nevertheless, adherence process researchmethods are ill-suited for answering some funda-mental process-related questions. For example,at its core adherence research is a hypothesis-testing approach to studying psychotherapy. Itsmain goals are confirming the clinical model andcataloguing prescribed behaviors hi session. Asdescribed in detail elsewhere (Hill, 1990; Mahrer,1988; Shoham-Salomon, 1990), discovery-oriented research, with its emphasis on exploringpsychotherapy process from a hypothesis-generating stance, takes quite the opposite tack.Discovery-oriented research attempts to track un-predictable or unexpected developments in ses-sion, draw from a variety of clinical theories tounderstand enigmatic phenomena, and formulatenew hypotheses when available evidence is weakor contradictory. These goals are better served bytask analysis and change events methods (Green-

340


berg, 1991; Safran, Greenberg, & Rice, 1988)than by adherence methods.

Adherence process research also has limitedpotential for evaluating facets of the therapeuticcontext. Context-sensitive process research is de-signed to investigate the often-subtle manner inwhich the meaning of behavior varies accordingto the context in which that behavior occurs(Greenberg, 1986; Heatherington, 1989; Messer,Tishby, & Spillman, 1992). Client behavior, andtherapist-client interpersonal processes, are keycomponents of the therapeutic context. On the onehand, adherence process research incorporates agiven model's routine expectations for context-sensitive interventions into the development of acoding instrument and training of judges. On theother hand, adherence research concentrates ontherapist behavior and, to a large degree, factorsout differential client behavior as "nuisance vari-ance." As a result, adherence process methodsare insensitive to many variables of great interestto process researchers: client pre-treatment char-acteristics, the nature of the presenting complaint,therapist awareness of client needs, therapist-client relationship characteristics, and subtletiesof culturally-bound or idiosyncratic communica-tion, to name a few.

Finally, treatment adherence process researchclearly belongs to the tradition of psychotherapyprocess research concerned with therapeutic tech-nique and comparative intervention strategies(Lambert & Bergin, 1994). This tradition has re-cently come under fire by some (e.g., Stiles &Shapiro, 1989,1994) who take issue with implicitsubscription to the "drug metaphor" in processresearch, a metaphor mat highlights "strength ofactive ingredients" and "dosage of intervention."According to these critics, the drug metaphor hastwo significant drawbacks: (1) It misrepresentstherapy as an enterprise in supplying titrated dos-ages of a prepared intervention, rather than oneof theory-based responsivity to client in-sessionneeds; and (2) It generates process research meth-ods with little power to detect process-outcomelinks, given their focus on intervention dosagerather than therapeutic responsivity. These criti-cisms support our contention that adherence pro-cess research is but one pillar in the process re-search framework. To bear the full weight ofcomprehensive clinical efficacy and process-outcome research, support from therapist compe-tency and therapeutic alliance research isrequired.

Concluding CommentsGrowing interest in treatment specification and

manualization, comparative clinical efficacy, andtherapeutic accountability has generated in-creased attention to the tasks of ensuring and doc-umenting that therapists implement treatmentmodels as intended. Although the technology forinvestigating treatment adherence is still in thedevelopmental stage, particularly in the field offamily therapy, it is clear that more complexmethods yield more comprehensive, clinicallyrich information. When sophisticated proceduresare used, adherence research becomes a powerfulprocess research tool for examining therapeuticinterventions across sessions, phases, therapists,and modalities. For these reasons, adherence re-search is poised to make significant contributionsto the ongoing project of identifying essential in-gredients and illuminating process-outcome linksin clinically-effective psychotherapies. As thescience of adherence research matures, advancingfrom issues of "What this model is" to "Whatthis model does," treatment adherence processresearch should become a valuable feature of psy-chotherapy process research.

ReferencesALEXANDER, J. F., BARTON, C , SCHIAVO, R.S.,& PARSONS,

B. (1976). Systems behavioral intervention with familiesof delinquents: Therapist characteristics, family behavior,and outcome. Journal of Consulting and Clinical Psychol-ogy, 44, 656-664.

ALEXANDER, J. F., NEWELL, R. M., ROBBINS, M. S., &TURNER, C. W. (1995). Observational coding in familytherapy process research. Journal of Family Psychology,9, 355-365.

BASHAM, R. (1986). Scientific and practical advantages ofcomparative design in psychotherapy outcome research.Journal of Consulting and Clinical Psychology, 54,88-94.

BATESON, G. (1972). Steps to an ecology of mind. NY:Ballantine.

BEUTLER, L. (1991). Have all won and must all have prizes?Revisiting Luborsky et al.'s verdict. Journal of Consultingand Clinical Psychology, 59, 226-232.

BINDER, J. L., STRUPP, H., BONOAR, B., LEE, S., MESSER,S., & PEAKE, T. (1993). Recommendations for improvingpsychotherapy training based on experiences with manual-guided training and research: Epilogue. Psychotherapy, 30,599-600.

BORDIN, E. S. (1979). The generalizability of the psychoana-lytic concept of the working alliance. Psychotherapy, 16,252-259.

BORDIN, E. S. (1994). Theory and research on the therapeuticworking alliance: New directions. In A. Horvath and L.Greenberg (Eds.), The working alliance; Theory, research,and practice (pp. 13-37). New York: John Wiley.

BORDUIN, C , MANN, B., CONE, L., HENCHELER, S., Fuoa,B., BLASKE, D., & WILLIAMS, R. (1995). Multisystemic

341


treatment of serious juvenile offenders: Long-term preven-tion of criminality and violence. Journal of Consulting andClinical Psychology, 63, 569-578.

BROWN, J. (1987). A review of meta-analyses conducted onpsychotherapy outcome research. Clinical Psychology Re-view, 7, 1-23.

BRY, B., CONBOY, C , & BISCAY, K. (1986). Decreasingadolescent drug use and school failure: Long-term effects oftargeted family problem-solving training. Child and FamilyBehavior Therapy, 8, 43-59.

CARROLL, K., NICH, C , & ROUNSAVILLE, B. J. (In press).Use of observer and therapist ratings to monitor deliveryof coping skills treatment for cocaine abusers: Utility oftherapist sessions checklist. Psychotherapy Research.

CARROLL, K. & NURO, K. (In press). Treatment manuals andtreatment compliance. In K. M. Carroll (Ed.), Facilitatingcompliance in alcohol treatment. NIAAA Project MATCHMonograph Series.

CHEVRON, E. S. & ROUNSAVILLE, B. J. (1983). Evaluatingthe clinical skills of psychotherapists: A comparison oftechniques. Archives of General Psychiatry, 40, 1129-1132.

CLARKE, G. (199S). Improving the transition from basic effi-cacy research to effectiveness studies: Methodological is-sues and procedures. Journal of Consulting and ClinicalPsychology, 63, 718-72S.

COYNE, J. & LIDDLE, H. (1992). The future of systems ther-. apy: Shedding myths and facing opportunities. Psychother-

apy, 29, 44-50.DERUBHS, R. J. & FEELEY, M. (1990). Determinants of

change in cognitive therapy for depression. Cognitive Ther-apy and Research, 14, 469-482.

DERUBEIS, R. J., HOLLON, S. D., EVANS, M. D , & BEMIS,K. M. (1982). Can psychotherapies for depression be dis-criminated? A systematic investigation of cognitive therapyand interpersonal therapy. Journal of Consulting and Clini-cal Psychology, 50, 744-756.

DIAMOND, G. & LIDDLE, H. A. (1996). Resolving a therapeu-tic impasse between parents and adolescents in multidimen-sional family therapy. Journal of Consulting and ClinicalPsychology, 64(3), 481-488.

DOBSON, K. S. & SHAW, B. F. (1993). The training of cogni-tive therapists: What have we learned from treatment manu-als? Psychotherapy, 30, 573-577.

ELKIN, I., PARLOFF, M., HADLEY, S., & AUTRY, J. (1985).NIMH Treatment of Depression Collaborative ResearchProgram. Archives of General Psychiatry, 42, 305-316.

ELKIN, I., PILKONIS, P., DOCHERTY, J., & SOTSKY, S. (1988).Conceptual and methodological issues in comparative stud-ies of psychotherapy and pharmacotherapy, I: Active ingre-dients and mechanisms of change. American Journal ofPsychiatry, 145, 909-917.

ESTRADA, A. U. & PINSOF, W. (1995). The effectiveness offamily therapies for selected behavioral disorders of child-hood. Journal of Marital and Family Therapy, 21,403-440.

EVANS, M., PIASECKI, J., KRISS, M., & HOLLON, S. (1984).Rater's Manual for the Collaborative Study PsychotherapyRating Scale- Form 6 (CSPRS-6). Unpublished manuscript.University of Minnesota and St. Paul-Ramsey MedicalCenter.

EYSENCK, H. (1952). The effects of psychotherapy: An evalu-ation: Journal of Consulting Psychology, 16, 319-324.

FLOYD, F. J. (1989). Segmenting interactions: Coding unitsfor assessing marital and family behaviors. Behavioral As-sessment, 11, 13-29.

FRIEDLANDER, M., WILDMAN, J., HEATHERINGTON, L., &

SKOWRON, E. (1994). What we do and don't know about theprocess of family therapy. Journal of Family Psychology, 8,390-416.

FRIEDMAN, A. S. (1989). Family therapy vs. parent groups:Effects on adolescent drug abusers. The American Journalof Family Therapy, 17, 335-347.

FRIEDMAN, A. S., TOMKO, L., & UTADA, A. (1991). Clientand family characteristics that predict better family therapyoutcome for adolescent drug abusers. Family Dynamics ofAddiction Quarterly, 1, 77-93.

GARFIELD, S. L. (1990). Issues and methods in psychotherapyprocess research. Journal of Consulting and Clinical Psy-chology, 58, 273-280.

GASTON, L. (1990). The concept of the alliance and its rolein psychotherapy: Theoretical and empirical considerations.Psychotherapy, 27, 143-153.

GREENBERG, L. S. (1986). Change process research. Journalof Consulting and Clinical Psychology, 54, 4-9.

GREENBERG, L. S. (1991). Research on the process of change.Psychotherapy Research, 1, 3-16.

GREENBERG, L. S. (1995). The use of observational codingin family therapy research: Comment on Alexander et al.(1995). Journal of Family Psychology, 9(4), 366-370.

GREENBERG, L. & PINSOF, W. (1986). Process research: Cur-rent trends and future perspectives. In L. Greenberg and W.Pinsof (Eds.), The psychotherapeutic process: A researchhandbook (pp. 3-20). New York: Guilford.

HAZEUUGG, M. D., COOPER, H. M., & BORDUIN, C. M.(1987). Evaluating the effectiveness of family therapies:An integrative review and analysis. Psychological Bulletin,101, 428-442.

HEATHERINGTON, L. (1989). Toward more meaningful clinicalresearch: Taking context into account in coding psychother-apy interaction. Psychotherapy, 26, 436-447.

HENGGELER, S. W., BORDUIN, C. M., & MANN, B. J. (1993).Advances in family therapy: Empirical foundations. InT. H. Ollendick and R. J. Prinz (Eds.), Advances in clinicalchild psychology (vol. 15, pp. 207-241). New York:Plenum.

HENGGELER, S., MELTON, G., & SMITH, L. (1992). Familypreservation using Multisystemic Therapy: An effective al-ternative to incarcerating serious juvenile offenders. Jour-nal of Consulting and Clinical Psychology 60, 953-961.

HENRY, W. P., STRUPP, H. H., BUTLER, S. F., SCHACHT,T. E., & BINDER, J. L. (1993). Effects of training in time-limited dynamic psychotherapy: Changes in therapist be-havior. Journal of Consulting and Clinical Psychology, 61,434440.

HERRON, W., EISENSTADT, E., JAVIER, R., PRIMAVERA, L., &SCHULTZ, C. (1994). Session effects, comparability, andmanaged care in the psychotherapies. Psychotherapy, 31,346-350.

HILL, C. E. (1990). Exploratory in-session process researchin individual psychotherapy: A review. Journal of Con-sulting and Clinical Psychology, 58, 288-294.

HILL, C. E. (1991). Almost everything you ever wanted toknow about how to do process research on counseling andpsychotherapy but didn't know who to ask. In C. E. Hilland L. J. Schneider (Eds.), Research in counseling (pp.85-118). Hillsdale, NJ: Lawrence Erlbaum.

HILL, C. E. (1994). From an experimental to an exploratorynaturalistic approach to studying psychotherapy process. InR. L. Russell (Ed.), Reassessing psychotherapy research(pp. 144-165). New York: Guilford.

342


HILL, C. E. & CORBETT, M. M. (1993). A perspective onthe history of process and outcome research in counselingpsychology. Journal of Counseling Psychology, 40, 3-24.

Hnx, C. E., O'GRADY, K. E., & ELHN, I. (1992). Applyingthe Collaborative Study Psychotherapy Rating Scale to ratetherapist adherence in cognitive-behavior therapy, interper-sonal therapy, and clinical management. Journal of Con-sulting and Clinical Psychology, 60, 73-79.

HILL, C. E., O'GRADY, K. E., & PRICE, P. (1988). A methodfor investigating sources of rater bias. Journal of Consultingand Clinical Psychology, 35, 346-350.

HOOUE, A., ROWE, C , LIDDLE, H., & TURNER, R. (1994a).Therapist Behavior Rating Scale (TBRS). Unpublishedmanuscript, Center for Research on Adolescent DrugAbuse, Temple University, Philadelphia, PA.

HOOUE, A., ROWE, C , LIDDLE, H., & TURNER, R. (19946).Scoring manual for the Therapist Behavior Rating Scale(TBRS). Unpublished manuscript. Center for Research onAdolescent Drug Abuse, Temple University, Philadel-phia, PA.

HOLLON, S., EVANS, M., AUERBACH, A., DERUBEUS, R.,ELUN, I., LOWERRY, A., KRISS, M., GROVE, W., TUASON,V., & PIASECKI, S. (1988). Development of a system forrating therapies for depression: Differentiating cognitivetherapy, interpersonalpsychotherapy and clinical manage-ment pharmacotherapy. Unpublished manuscript, Vander-bilt University, Nashville, TN.

HORVATH, A. O. (1995). The therapeutic relationship: Fromtransference to alliance. In Session: Psychotherapy in Prac-tice, 1, 7-17.

HORVATH, A. O. & LUBORSKY, L. (1993). The role of thetherapeutic alliance in psychotherapy. Journal of Con-sulting and Clinical Psychology, 61, 561-573.

HORVATH, A. O. & SYMONDS, B. D. (1991). Relation betweenworking alliance and outcome in psychotherapy: A meta-analysis. Journal of Counseling Psychology, 38, 139-149.

JOANNNO, H., QUINN, W., TliOMAS, F. , & MULLEN, R.(1992). Treating adolescent drug abuse: A comparison offamily systems therapy, group therapy, and family drugeducation. Journal of Marital and Family Therapy, 18,345-356.

JUUEN, D., MARKMAN, H. J., & LINDAHL, K. M. (1989). Acomparison of a global and a mkroanalytic coding system:Implications for future trends hi studying interactions. Be-havioral Assessment, 11, 81-100.

KAZXHN, A. E. (1986). Comparative outcome studies of psy-chotherapy: Methodological issues and strategies. Journalof Consulting and Clinical Psychology, 54, 95-105.

KAZTHN, A. E. (1994). Methodology, design, and evaluationin psychotherapy research. In A. Bergin and S. Garfield(Eds.), Handbook of psychotherapy and behavior change(4th ed., pp. 19-71). New York: John Wiley.

KAZUN, A. E. &BASS, D. (1989). Power to detect differencesbetween alternative treatments in comparative psychother-apy outcome research. Journal of Consulting and ClinicalPsychology, 57, 138-147.

KAZMN, A. E., BASS, D., AYERS, W. A., & RODOERS, A.(1990). Empirical and clinical focus of child and adolescentpsychotherapy research. Journal of Consulting and ClinicalPsychology, 58, 729-740.

KAZWN, A. E., BASS, D., SIEGEL, T., & THOMAS, C. (1989).Cognitive-behavioral therapy and relationship therapy inthe treatment of children referred for antisocial behavior.Journal of Consulting and Clinical Psychology, 57, 522-535.

KENDALL, P. C. (1994). Treating anxiety disorders in chil-dren: Results of a randomized clinical trial. Journal of Con-sulting and Clinical Psychology, 62, 100-110.

LAMBERT, M. &. BERGIN, A. (1994). The effectiveness ofpsychotherapy. In A. Bergin and S. Garfield (Eds.), Hand-book of psychotherapy and behavior change (4th ed., pp.143-189). New York: John Wiley.

LAMBERT, M. & HILL, C. (1994). Assessing psychotherapyoutcomes and processes. In A. Bergin and S. Garfield(Eds.), Handbook of psychotherapy and behavior change(4th ed., pp. 72-113). New York: John Wiley.

LEBOW, J. L. A GURMAN, A. S. (1995). Research assessingcouple and family therapy. Annual Review of Psychology,46, 27-57.

LIDDLE, H. A. (1991). Empirical values and the culture offamily therapy. Journal of Marital and Family Therapy,17, 327-348.

LIDDLE, H. A. (1995). Conceptual and clinical dimensions of amnit«tmifngifWHii (imiltisysi^iii^ engagement strategy in fami-ly-based adolescent treatment. Psychotherapy, 32, 39-58.

LIDDLE, H. A., BECKER, D., & DIAMOND, G. (In press).Training and supervision issues in family-based treatment:The case of Multidimensional Family Therapy. In E. Wat-kins (Eds.), Handbook of psychotherapy supervision (2nded.). New York: John Wiley.

LIDDLE, H. A. & DAKOF, G. A. (1995). Efficacy of familytherapy for drug abuse: Promising but not definitive. Jour-nal of Marital and Family Therapy, 21, 511-543.

LIDDLE, H. A., DAKOF, G. A., & DIAMOND, G. (1991). Ado-lescent substance abuse: Multidimensional family therapyin action. In E. Kaufman and P. Kaufman (Eds.), Familytherapy of drug and alcohol abuse (pp. 120-171). Boston:AUyn and Bacon.

LIDDLE, H. A. & DIAMOND, G. (1991). Special problems inworking with adolescent substance abuse: Engaging reluc-tant teenagers in family therapy. Family Dynamics of Ad-dictions Quarterly, 1, 55-69.

LUBORSKY, L. & BARBER, J. (1993). Benefits of adherenceto psychotherapy manuals, and where to get them. In N.Miller, L. Luborsky, J. Barber, and J. Docherty (Eds.),Psychodynamic treatment research: A handbook for clini-cal practice (pp. 211-226). New York: Basic.

LUBORSKY, L. & DERUBEB, R. (1984). The case of psychother-apy treatment manuals: A small revolution in psychotherapyresearch style. Clinical Psychology Review, 4, 5-14.

LUBORSKY, L., SINGER, B., & LUBORSKY, L. (1975). Compar-ative studies of psychotherapy: Is it true that "Everyonehas won and all must have prizes?" Archives of GeneralPsychiatry, 32, 995-1008.

LUBORSKY, L., WOODY, G., MCLELLAN, A., O'BRIEN, C ,& AUERBACH, A. (1985). Therapist success and its determi-nants. Archives of General Psychiatry, 42, 602-611.

MAHRER, A. R. (1988). Discovery-oriented psychotherapyresearch. American Psychologist, 43, 694-702.

MANN, B. J. & BORDUIN, C. M. (1991). A critical reviewof psychotherapy outcome studies with adolescents: 1978-1988. Adolescence, 26, 505-541.

MARKMAN, H. J., LEBER, B. D., CORDOA, A. D., & ST.PETERS, M. (1995). Behavioral observation and family psy-chology—Strange bedfellows or happy marriage?: Com-ment on Alexander et al. (1995). Journal of Family Psy-chology, 9, 371-379.

MARMAR, C. R. (1990). Psychotherapy process research:Progress, dilemmas, and future directions. Journal of Con-sulting and Clinical Psychology, 58, 265-272.

343


MESSER, S., TISHBY, O., & SRLLMAN, A. (1992). Takingcontext seriously in psychotherapy research: Relating thera-pist interventions to patient progress in brei psychodynamictherapy. Journal of Consulting and Clinical Psychology,60,678-688.

MINTZ, J. & LUBORSKY, L. (1971). Segments versus wholesessions: Which is the better unit for psychotherapy processresearch? Journal of Abnormal Psychology, 78, 180-191.

MONCHER, F. J. & PRINZ, R. J. (1991). Treatment fidelity inoutcome studies. Clinical Psychology Review, 11,247-266.

MONTALVO, B. (1973). Aspects of live supervision. FamilyProcess, 12, 343-359.

NEWMAN, F. & HOWARD, K. (1986). Therapeutic effort, treat-ment outcome, and national health policy. American Psy-chologist, 41, 181-187.

O'MALLEY, S. S., FOLEY, S. H., ROUNSAVILLE, B. J., WAT-KINS, J. T., SOTSKY, S. M., IMBER, S. D., & ELKJN, I.(1988). Therapist competence and patient outcome in inter-personal psychotherapy of depression. Journal of Con-sulting and Clinical Psychology, 56, 496-501.

ORUNSKY, D., GRAWE, K., & PARKS, B. (1994). Process andoutcome in psychotherapy-Noch einmal. In A. Bergin andS. Garfield (Eds.), Handbook of psychotherapy and behav-ior change (4th ed., pp. 270-376). New York: John Wiley.

PATVIO, S. & GREENBERG, L. (199S). Resolving "unfinishedbusiness": Efficacy of Experiential Therapy using empty-chair dialogue. Journal of Consulting and Clinical Psychol-ogy, 63, 419-425.

PINSOF, W. M. (1989). A conceptual fraemwork and methodo-logical criteria for family therapy process research. Journalof Consulting and Clinical Psychology, 57, 53-59.

PlNSOF, W. M. (1994). An integrative systems perspective onthe therapeutic alliance: Theoretical, clinical, and researchimplications. In A. Horvath and L. Greenberg (Eds.), Theworking alliance: Theory, research, and practice (pp. 173-195). New York: John Wiley.

PINSOF, W. & WYNNE, L. (1995a). The efficacy of maritaland family therapy: An empirical overview, conclusions,and recommendations. Journal of Marital and Family Ther-apy, 21, 585-613.

PINSOF, W. & WYNNE, L. (19956). The effectiveness andefficacy of marital and family therapy: Introduction to thespecial issue. Journal of Marital and Family Therapy, 21,341-343.

PRINCE, S. & JACOBSON, W. (1995). A review and evaluationof marital and family therapy: Introduction to the specialissue. Journal of Marital and Family Therapy, 21,341-343.

ROMUN, C , PLAIT, J., & SCUPPERS, G. (1990). Family ther-apy for Dutch drug abusers: Replication of an Americanstudy. The International Journal of the Addictions, 25,1127-1149.

RUSSELL, R. L. & STASZEWSKI, C. (1988). The unit problem:Some systematic distinctions and critical dilemmas for psy-chotherapy process research. Psychotherapy, 25,191-200.

SAFRAN, J. D., GREENBERG, L. S., & RICE, L. N. (1988).Integrating psychotherapy research and practice: Modelingthe change process. Psychotherapy, 25, 1-17.

SCHAFFER, N. D. (1982). Multidimensional measures of thera-pist behavior as predictors of outcome. Psychological Bul-letin, 92, 670-681.

SCHAFFER, N. D. (1983). Methodological issues of measuringthe skillfulness of therapeutic techniques. Psychotherapy:Theory, Research, and Practice, 20, 486-493.

SCHMIDT, S. E., LIDDLE, H. A., & DAKOF, G. A. (1996).Changes in parenting practices and adolescent drug abuse

during multidimensional family therapy. Journal of FamilyPsychology, 10, 12-27.

SECUREST, L. (1994). Recipes for psychotherapy. Journal ofConsulting and Clinical Psychology, 62, 952-954.

SHADISH, W. R., MONTGOMERY, L., WILSON, P., WILSON,M. R., BRIGHT, I., & OKWUMABUA, T. (1993). Effects offamily and marital psychotherapies: A meta-analysis. Journalof Consulting and Clinical Psychology, 61, 992-1002.

SHADISH, W. R., RAGSDALE, K., GLASER, R., & MONTGOM-ERY, L. (1995). The efficacy and effectiveness of maritaland family therapy: A perspective from meta-analysis.Journal of Marital and Family Therapy, 21, 345-360.

SHAPIRO, D. A. & SHAPIRO, D. (1982). Meta-analysis ofcomparative therapy outcome studies: A replication andrefinement. Psychological Bulletin, 92, 581-604.

SHAPIRO, D. A. & STARTUP, M. (1992). Measuring therapistadherence in exploratory psychotherapy. PsychotehrapyResearch, 2, 193-203.

SHAW, B. F. & DOBSON, K. S. (1988). Competency judgmentsin the training and evaluation of psychotherapists. Journal ofConsulting and Clinical Psychology, 56, 666-672.

SHEFLER, G., DASBERG, H., & BEN-SHAKHAR, G. (1995).A randomized controlled outcome and follow-up study ofMann's Time-Limited Psychotherapy. Journal of Con-sulting and Clinical Psychology, 63, 585-593.

SHIELDS, C , WYNNE, L., MCDANIEL, S., & GAWINSKI, B.(1994). The marginalization of family therapy: A historicaland continuing problem. Journal of Marital and FamilyTherapy, 20, 117-138.

SHOHAM-SALOMON, V. (1990). Interrelating research pro-cesses of process research. Journal of Consulting and Clini-cal Psychology, 58, 295-303.

SHOHAM-SALOMON, V. & HANNAH, M. (1991). Client-treatment interaction in the study of differential changeprocesses. Journal of Consulting and Clinical Psychology,59, 217-225.

SIMONS, A., GORDON, J., MONROE, S., &THASE, M. (1995).Toward an integration of psychologic, social, and biologicfactors in depression: Effects of outcome and course ofcognitive therapy. Journal of Consulting and Clinical Psy-chology, 63, 369-377.

SMITH, M. & GLASS, G. (1977). Meta-analysis of psychother-apy outcome studies. American Psychologist, 32,752-760.

SMITH, M., GLASS, G., & MILLER, T. (1980). The benefitsof psychotherapy. Baltimore: MD: Johns Hopkins Univer-sity Press.

STARTUP, M. & SHAPIRO, D. A. (1993). Therapist treatmentfidelity in prescriptive vs. exploratory psychotherapy. Brit-ish Journal of Clinical Psychology, 32, 443-456.

STILES, W. & SHAPIRO, D. (1989). Abuse of the drug metaphorin psychotherapy process-outcome research. Clinical Psy-chology Review, 9, 521-543.

STILES, W. & SHAPIRO, D. (1994). Disabuse of the drug meta-phor Psychotherapy process-outcome correlations. Journalof Consulting and Clinical Psychology, 62, 942-948.

STILES, W. & SHAPIRO, D. (1995). Verbal exchange structureof brief psychodynamic-interpersonal and cognitive-behav-ioral psychotherapy. Journal of Consulting and ClinicalPsychology, 63, 15-27.

STILES, W , SHAPIRO, D., & ELLIOTT, R. (1986). Are allpsychotherapies equivalent? American Psychologist, 41,165-180.

STINSON, C. H., MILBRATH, C , REIDBORD, S. P., & Bucci,W. (1994). Thematic sgementation of psychotherapy tran-scripts for convergent analyses. Psychotherapy, 31,36-48.

344


SZAFOCZNK, J . , KURTINES, W . , PEREZ-VlDAL, A . , HERV1S,O., & FOOTE, F. (1990). One-person family therapy. In R.A. Wells and V. A. Gianetti (Eds.), Handbook of briefpsychotherapies (pp. 493-510). New York: Plenum.

SZAPOCZNK, J., PEREZ-VIDAL, A., BRICKMAN, A., FOOTE,F., SANTISTEBAN, D., HERVIS, O., & KURTMES, W. (1988).Engaging adolescent drug abuseis and their families intreatment: A strategic structural systems approach. Journalof Consulting and Clinical Psychology, 56, 552-557.

SZAPOCZNK, J., Rio, A., MURRAY, E., COHEN, R., SCOPETTA,M., RIVAS-VASQUEZ, A., HERVIS, O., POSADA, V., & KUR-TMES, W. (1989). Structural family versus psychodynamic

child therapy for problematic Hispanic boys. Journal ofConsulting and Clinical Psychology, 57, 571-578.

WALTZ, J., ADDIS, M. E., KOERNER, K., & JACOBSON, N. S.(1993). Testing the integrity of a psychotherapy protocol:Assessment of adherence and competence. Journal of Con-sulting and Clinical Psychology, 61, 620-630.

WATZLAWICK, P. & WEAKLAND, J. (EDS.) (1977). The interac-

tional view. New York: W. W. Norton.YEATON, W. H. & SECUREST, L. (1981). Critical dimensions

in the choice and maintenance of successful treatments:Strength, integrity, and effectiveness. Journal of Con-sulting and Clinical Psychology, 49, 156-167.

345

Psychotherapy - MDFT · in psychotherapy outcome research as a means for testing whether...

Documents

Transcript of Psychotherapy - MDFT · in psychotherapy outcome research as a means for testing whether...