THE JOURNAL OF ALTERNATIVE AND COMPLEMENTARY MEDICINE. Volume 10, Number 3, 2004, pp. 580586 proach toward complementary and alternative medicine in …
THE JOURNAL OF ALTERNATIVE AND COMPLEMENTARY MEDICINE Volume 10, Number 3, 2004, pp 580586 Mary Ann Liebert, Inc
EDUCATION, INITIATIVES, AND INFORMATION RESOURCES
One Size Does Not Fit All: Aptitude x Treatment Interaction ATI as a Conceptual Framework for Complementary and Alternative Medicine Outcome Research Part 1–What Is ATI Research?
OPHER CASPI, MD, PhD, and IRIS R BELL, MD, PhD ABSTRACT When multiple treatment choices are available, the question is not just which treatment is the best? but more importantly best or better for whom, when, and why? Aptitude or attribute by treatment interaction ATI is a research paradigm that attempts to examine exactly that–how outcome depends on the match or mismatch between patients specific characteristics and the treatment they receive The purpose of this twopart paper is to introduce ATI methods as a conceptual framework into complementary and alternative medicine/integrative medicine CAM/IM outcome research Part 1 presents key concepts
in ATI research Part 2 will present ATI research designs and discusses their applications to the examination of the relationships between individuals and therapies, and the illumination of the mechanisms that make therapies differentially effective Based on this examination, we conclude that ATI research offers invaluable insights into the multifaceted package of care typically delivered in contemporary medicine and therefore should be included in the portfolio of all CAM/IM outcome research INTRODUCTION
acterized by more to know, more to do, more to manage, more to watch, and more people involved than ever before Institute of Medicine, 2001 However, offering complex multifaceted packages of care is challenging in at least one other way It results in more difficulty in making causal inferences regarding the relationship between interventions and outcomes This is because, depending on the situation, it is possible that some components of the package of care may augment each other,
others may be redundant of each other, and still others may cancel out each others effects Shoam and Rohrbaugh, 1995 Add to this the fact that human beings are uniquely complex and the result is a difficult puzzle to solve Therefore, outcome research that fails to account for this complexity systematically may inadvertently reach a misleading conclusion For example, null findings from comparative outcome studies may obscure systematic individual differ-
I
ncreasingly, medical treatments are administered as part of multifaceted packages of care that contain a variety of therapeutic components However, such an approach to patient care may inadvertently present some unique challenges First, from the patients cognitive standpoint, research suggests that medical care that involves multiple alternative therapeutic options may result in paradoxical paralysis Redelmeier and Shafir, 1995 Second, from a practical standpoint, it is unclear whether complex packages of care, compared to more
limited approaches, result in better outcomes National Center for Complementary and Alternative Medicine, 2000 Last, from a health policy standpoint, it is questionable whether this multifaceted approach to patient care is effective, efficient, and economically justifiable at a time of growing complexity of health care, which today is char-
Program in Integrative Medicine, University of Arizona, Tucson, AZ
580
APTITUDE-TREATMENT INTERACTION ences in response to specific treatments, and differentially effective treatment may be the result of systematic, predictable differences at the patient level Shoam and Rohrbaugh, 1995 Hence, for outcome research the key question is not just, which treatment is the best? but more importantly best or better for whom, when, and why? Snow, 1991 Aptitude or attribute by treatment interaction ATI is a research paradigm that attempts to examine exactly that– how outcome depends on the match or mismatch between patients specific characteristics and
the treatment they receive It promises to examine the relationship between individuals and therapies with respect to outcome and to illuminate the mechanisms and processes that make therapies differentially effective Cronbach, 1957; Shoam-Salomon, 1991 However, despite its pragmatic, heuristic, and methodological appeals to come in Part II ATI research has been used only sparsely in decades of research in diverse fields such as education, psychology, and medicine Beutler, 1991; Dance and Neufeld, 1988; Shoam-Salomon, 1991; Shoam-Salomon and Hannah, 1991; Smith and Sechrest, 1991; Snow, 1991 Thus, its promise remains largely unfulfilled, perhaps in part due to lack of awareness The purpose of this two-part paper, therefore, is to introduce this cluster of ATI research design and analysis methods as a conceptual framework into the growing portfolio of outcome research Because the principles and theory of ATI research are generic and de facto pertain to all avenues of human therapy, our
hope is that the papers will stimulate more hypothesis-driven outcome research across all medical disciplines and specialties
581 thereof, benefits what kind of patient? Shoam-Salomon and Hannah, 1991 In the ATI paradigm, the A aptitude or attribute stands for any individual difference variables or patient characteristics that may moderate the effects of a treatment T on an outcome O The T may represent any type of intervention, from a whole integrated therapy package to a specific, even single intervention As explained below, it refers not only to the intervention itself, but also to characteristics of the environment in which the therapy takes place, and the therapists characteristics The I interaction term is used in a statistical not social sense, referring to the moderating effect of A on the relationship between T and O Shoam and Rohrbaugh, 1995 To illustrate what ATI research is about better, consider the relationship between absorption and treatment as an example The
personality trait of absorption reflects an individuals capacity to immerse oneself in a self-absorbing experience Neff et al 1983 reported that how patients respond to treatment for headache could be predicted by their level of absorption That is, patients with vascular headaches high in absorption responded well to relaxation training but not to biofeedback while low absorbers improved significantly after biofeedback training On the other hand, patients with tension headaches high in absorption responded better to biofeedback while those low in absorption did not respond significantly to either biofeedback or relaxation Likewise, earlier research had shown that high absorbers did better with internal imagery methods while low absorbers did better with external biofeedback methods Qualls and Sheehan, 1979, 1981 Shea 1985 found that males low in absorption achieved better voluntary control of heart rate with hypnosis compared to imagery Gregerson et al 1996 observed that high absorbers
increased mucosal immunoglobulin A IgA in saliva using a combination of relaxation and imagery, whereas low absorbers decreased mucosal IgA under the same intervention Taken together, these studies suggest that prior knowledge of patient characteristics may be useful for tailoring the kind of intervention that would be most effective In what follows we describe ATI research in more details The rest of this paper explicates what exactly is meant by aptitude, treatment and interaction Part 2 will present ATI research designs and discuss their applications with respect to the examination of the relationships between individuals and therapies, and the illumination of the mechanisms and processes that make therapies differentially effective
WHAT IS ATI RESEARCH?
In a series of publications in the mid 1950s Edwards and Cronbach 1952 and later Cronbach 1957 proposed an innovative way to relate individuals to outcomes Rather than looking for main effects, they suggested looking at the
[o]therwise neglected interactions between organismic and treatment variables Cronback, 1957 so as to discover [s]pecific types of persons and situations for which [treatment] is superior Edwards and Cronbach, 1952 They termed this new approach to outcome research AptitudeTreatment Interaction ATI ATI research is designed to systematically take into account individual differences in the process of treatment evaluation It is meant to assess the degree to which different treatments have different effects as a function of persons characteristics so as to examine whether particular treatments can be chosen or adapted to optimally fit a particular person Snow, 1991 As such, rather than asking, who benefits most from undifferentiated therapy? or what therapy is best for undifferentiated patients?, ATI research focuses on the question, what kind of therapy, or elements
APTITUDE
The term aptitude as used within the ATI framework refers to a complex set of personal characteristics identified
before and during treatment that are hypothesized to be causally related to the persons outcome at the end of the
582 treatment According to Snow 1991, aptitudes can be related to all relevant individual differences, whether cultural, psychologic, or biologic It is not limited to intelligence or some fixed list of differential abilities Nor is any particular personality theory implied Rather, aptitude refers to a heterogeneous set of personality and motivational variables along with styles, attitudes, and beliefs, to name only a few of the constructs and variables that are of interest in this research paradigm Some researchers Shoam-Salomon and Hannah, 1991 have even advocated that aptitude be expanded to include relational variables eg, the quality of the patients marriage; Barlow et al, 1984; Rohrbaugh et al, 2002 An aptitude is thus a dynamic, relational, and contextual construct Increasingly, aptitudes are being treated not just as dynamic constructs, but more as processes that,
once activated, can affect the treatment that has activated them Snow, 1980 This reciprocal relationship may alter the course of treatment, the meaning attributed to it, and the aptitude processes involved reflecting aptitude as a state not just a trait Shoam-Salomon and Hannah, 1991; Snow, 1991 Review of the ATI literature reveals, however, that most ATI research focuses on presumably stable patient characteristics such as personality traits and coping styles see Dance and Neufeld, 1988, for a review Again, this seems to be an unwarranted simplification because it assumes that the stable patient characteristic, measured before therapy, interacts with the treatment to influence treatment outcome Shoam and Rhorbaugh, 1995 Looking at aptitude as a process rather than as a fixed entity implies that the treatment may produce a change in the moderator variable itself Thus, moderators, when shaped by the treatment, may turn into mediators and further interact with the treatment Shoam and
Rohrbaugh, 1995 For example, patients expectations were shown to moderate and mediate various medical outcomes such as quality of life Carr et al, 2001; Crow et al, 1999 Many researchers have advocated examining process variables that may as well moderate the effects of a treatment Mitchell and Cormack, 1998; Shoam and Rohrbaugh, 1995; Snow, 1991 Di Blasi et al 2001, for example, examined whether doctorpatient relationships have an important therapeutic effect, irrespective of any prescribed drug or treatment Their findings that some physicians styles are more effective than others for example, adopting a warm, friendly, and reassuring manner suggest that if we really are committed to understanding outcomes research, we need to consider provider aptitudes in addition to patient aptitudes Indeed, the recent JNC 7 report on prevention, detection, evaluation, and treatment of hypertension suggests that empathy and trust may mediate patients outcome through their effect on adherence to
treatment Chobanian et al, 2003
CASPI AND BELL one way or another is not enough Consider individualized psychotherapy as an example No two sessions are alike So just labeling treatment as psychotherapy may represent only a remote approximation for the infinite richness of the therapeutic experience Labels, in short, do not make treatments what they are That is, patients interact with the treatment as it is administered and not with its label or theoretical orientation Shoam-Salomon and Hannah, 1991 Rohrbaugh et al 2002, advocated a dismantling approach to packages of care According to their method, studying isolated treatments, or treatments within treatments little ts, in their language instead of integrative, broadband, multi-component treatment packages big Ts, in their language will allow us to understand better which aspects of the integrative package make a difference for whom, or how the various components alone or in combination actually work However, this approach does not
take into consideration the possibility that the whole may exhibit properties that its separate parts do not possess Thus, unlike the reductionistic approach to research that typically examines singular parts of health care and parts of the individual, one at a time, but not the complete system, we advocate for a more integrative approach that studies the whole Bell et al, 2002 And yet, adopting the latter approach poses some real challenges For example, if ATIs are real, should we expect certain patient characteristics to interact with the whole treatment package itself as if it were a single independent variable or with a specific element of that intervention? If so, which one: the specific, the nonspecific, or perhaps both? Caspi and Bootzin, 2002; Grünbaum, 1985 ATI research seems particularly useful within the framework of this debate It can help determine whether interaction exists between little ts and big Ts, such that the effectiveness of a specific intervention depends on the
larger model in which it is applied a phenomenon known as contextual moderation; Shoam and Rhorbaugh, 1996 Evidence of such interaction casts doubt on the empirical legitimacy of decontextualizing specific intervention techniques little ts by using them in other big Ts context Elsewhere, we have argued that the cut-and-paste approach toward complementary and alternative medicine integration, in which empirically supported therapies are integrated into the health care delivery system without considering their philosophical underpinnings, is too reductionistic Bell et al, 2002; Caspi, 2001 ATI research allows us to test this proposition
OUTCOMES
Choosing the right outcomes seems to be an especially challenging task in the ATI research paradigm This is because different individuals may experience qualitatively different but equally worthwhile outcomes as a result of receiving the same therapy Shoam-Salomon and Hannah, 1991 For example, Shoam-Salomon et al 1989 hypoth-
TREATMENT
Within
the ATI framework, it is imperative to define what exactly is meant by treatment Only labeling a treatment
APTITUDE-TREATMENT INTERACTION esized and found that given paradoxical interventions ie, trying to induce change by discouraging it, high-reactance clients manifest a behavioral change, whereas low-reactance clients manifest a cognitive change The former have become more effective in their daily behaviors, whereas the latter have derived improved self-efficacy for dealing with their problems The realization that different individuals may experience qualitatively different but equally worthwhile outcomes led Shoam-Salomon and Hannah 1991 to propose aptitudeoutcome interaction AOI as another type of interaction worth studying However, despite the fact that both ATI and AOI can shed light on the processes of therapeutic change, they are distinctly different In ATI research one studies the differential effectiveness of at least two therapies for individuals who vary on at least one
relevant aptitude The outcome criterion is common to both therapies In AOI, on the other hand, one focuses on one therapy and at least one aptitude but seeks an interaction that is manifested in different outcome criteria Shoam-Salomon Hannah, 1991 Smith and Sechrest 1991 proposed yet another possibility of aptitude x treatment x outcome Although AOI research was originally developed for psychotherapy outcome research, it seems relevant to all fields of medical outcome research because therapy-related effects may be mapped onto any number of outcomes along the continuum of curecare In that respect then, outcomes are never ultimate Every outcome is an intermediate step toward still another outcome Shoam-Salomon and Hannah, 1991 For that reason, the recommendation to use multiple outcome variables in order to obtain a richer understanding of potential ATIs Dance and Neufeld, 1988; Shoam-Salomon and Hannah, 1991; Snow, 1991 fits well within the complex systems theory approach to outcome
research we advocated Bell et al, 2002
583 regression lines are nonparallel, various methods exist for determining range of the aptitude over which the regression lines differ significantly, specifying the exact value of the patients characteristics above which one treatment or a package of treatments is superior to another For a summary review of those techniques see Cronbach and Snow, 1977; Karpman, 1983; Rogosa, 1980 Two types of interactions are possible: disordinal, meaning that treatment alternatives are differentially effective across different values of a given aptitude and ordinal, meaning that a given treatment is always superior to an alternative, regardless of the level of the aptitude A disordinal interaction is apparent if the two slopes intersect somewhere within the range of the aptitude measured Fig 1 The example displayed in Figure 1 implies that individuals scoring high on aptitude X have a better outcome on outcome Y when given treatment A In contrast, individuals
low on aptitude X obtain better results on outcome Y with treatment B In an ordinal interaction, one regression line remains above the others across the sampled aptitude range, but the lines are nevertheless nonparallel Fig 2 This result would appear to indicate that treatment A is superior for all subjects measured on the aptitude Although ATI is taken to refer to the greater effect and/or efficiency of a treatment in the presence of some characteristics than others, Smith and Sechrest 1991, have argued that for ATI to have both pragmatic and heuristic value see below the demonstration of interaction is meaningful only if it shows specificity That is, in order to infer an interaction, it is not enough to show that a specific therapy is superior to another therapy as depicted for example in Fig 2 One must also demonstrate that the difference between them is smaller or larger than for some other condition Thus, ATI research bears the minimum requirement that two treatments and two
aptitudes be compared To better illustrate this point, consider the following example Kalauokalani et al 2001, hypothesized an associ-
INTERACTION
An effect of statistical interaction occurs when the relation between two or more variables is modified by at least one other variable Two variables interact if the effect of one of the variables differs depending on the level of the other variable That is, the magnitude of an effect is greater at one level of a variable than at another Statistical interaction is thus different from main effect, which is the effect of an independent variable averaging over all of the other variables To find out whether interaction within the experimental framework of ATI research exists, subjects are measured on some aptitude and then randomly assigned to one of two or more treatments Outcome is measured on a separate variable Regression equations are then calculated within each treatment group, using the individual aptitude measures as a predictor of the
outcome An interaction of aptitude with treatment is present when the obtained regression slopes differ Dance and Neufeld, 1988 After determining that the
FIG 1 Disordinal interaction, in which two slopes intersect somewhere within the range of the aptitude measured Adapted from Dance and Neufeld, 1998
584
CASPI AND BELL stantial within group variation across all types of human therapy serves as ample testimony that interaction effects are an integral part of the ecology of health care Beutler, 1991 Findings of patienttherapy interaction provide further support for the critique of the patient uniformity proposition, a myth that is based on the tacit assumption that patients with a given disorder constitute a homogenous group Dance and Neufeld, 1988; Kiesler, 1966 This myth stands in sharp contrast to the wisdom of day-to-day medicine and ATI research, both proving that therapies, patients, and providers are not interchangeable Good medicine means individualization of treatments so
that treatments are tailored and individually customized to fit the dynamic nature of the patient bio-psycho-social-spiritual makeup or profile Aptitudes, such as personal readiness to profit from particular treatment situations Prochaska, 1999 or perception of locus of control Wallston et al, 1994 may render patients with similar diagnoses dissimilar with respect to both treatment and outcome If outcome research fails to account for such differences, a potentially wrong conclusion may be made, such as in the famous case of comparative psychotherapy, where much to the disappointment of many researchers the conclusion was everyone has won and all must have prizes Luborsky et al, 1975 We mentioned previously that aptitude, as used within the ATI framework, could be related to all relevant individual differences, whether cultural, psychologic, or biologic To illustrate how biologic differences support the face validity of ATI research, consider drug therapy For many years medicine had no
reliable ways to account for individual differences among patients As a result, physicians prescribed approximately the same dosage of medications to everyone, young and old, men and women There are only a few exceptions to this practice For example, pediatric dosage is routinely adjusted for weight Likewise, the dosage of cytotoxic medications is often adjusted to body area or height However, it has now become evident that these simple adjustments to medication prescription are probably too simplistic With the advent of pharmacogenomics and pharmacogenetics individual differences can finally be modeled and taken into account Suddenly, measures such as height and weight appear to be crude, maybe even archaic Isoenzyme phenotyping emerges as a more precise method to understand and predict drug responses Differential response to alcohol Room, 2001 and the case of multidrug resistance are just two examples for how individual differences moderate treatment outcomes Goldie, 2001 It appears
then that the principles of ATI are not foreign to medicine They simply are not addressed as such
FIG 2 Ordinal interaction, in which one regression line remains above the other across the sampled aptitude range, but the lines are nevertheless nonparallel Adapted from Dance and Neufeld, 1998
ation between patient expectation for benefit from a specific treatment and improved functional outcomes To test this hypothesis they conducted a subanalysis of data derived from a randomized clinical trial of 135 patients with chronic low-back pain who received acupuncture or massage Study participants were asked before randomization to describe their expectations regarding the helpfulness of each treatment on a scale of 0 to 10 The primary outcome was level of function at 10 weeks They found that after adjustment for baseline characteristics, improved function was observed for 86 of the participants with higher expectations for the treatment they received, as compared with 68 of those with
lower expectations p 001 Furthermore, patients who expected greater benefit from massage than from acupuncture were more likely to experience better outcomes with massage than with acupuncture, and vice versa p 003 The results of this study suggest that patient expectations may specifically influence clinical outcome through the interaction with the treatment itself In contrast, general optimism about treatment, divorced from any specific treatment, was not strongly associated with outcome
FACE VALIDITY OF ATI RESEARCH
As Shoam and Rohrbaugh 1995 suggested, [B]oth clinicians and clients seem to know that different folks benefit from different strokes: the idea feels right Smith and Sechrest 1991 also agreed that, [C]ommon sense suggests that there should be at least some ATI in therapy But even if different folks benefit in much the same way regardless of treatment ie, their outcomes would be the same, they may still exhibit different change mechanisms when exposed to certain types of
interventions to be covered in Part II– Change Curves; Shoam-Salomon, 1991 Indeed, the sub-
HOW AND WHY INTERVENTIONS WORK?
As discussed above, the cluster of ATI research methods offers a way to study the effects of variables that mediate
APTITUDE-TREATMENT INTERACTION outcomes such as patient aptitudes Accordingly, ATI research seeks to understand questions such as, who does and does not benefit from a specific treatment? Would the patients who do not benefit from one treatment benefit from another treatment? Would patients who do well under one treatment benefit less if assigned to another treatment? Are treatments for well-matched patients more effective than treatment for poorly matched or randomly assigned patients? Shoam and Rohrbaugh, 1995 Symmetrically, ATI research may also be useful for detecting interventions that are harmful Ridenour et al, 1999 Thus, ATI research may have both heuristic and pragmatic values Heuristically, ATI research promises to provide information
that would result in a better understanding of the mechanisms that underlie the outcome achieved by the administration of different treatments to different individuals That is, ATI research may help advance knowledge about the nature and course of clinical disorders and hence impact theory development Pragmatically, ATI research aims at providing information that would result in optimal matching between different patients based on their aptitudes and different treatments in order to maximize treatment efficiency and effectiveness Thus, by identifying components of an intervention that are more or less effective under different conditions ATI research may help determine the type of interventions that need to be adapted in order to fit local needs and to increase the potency and cost-effectiveness of care National Institute of Mental Health, 2002 In Part 2 of this paper we will discuss how exactly this can be accomplished
585
ordinating Committee The seventh report of the Joint National
Committee on Prevention, Detection, Evaluation, and Treatment of high blood pressure The JNC 7 report JAMA 2003;289: 25602572 Cronbach, LJ The two disciplines of scientific psychology Am Psychol 1957;12:671684 Cronbach, LJ, Snow RE Aptitude and Instructional Methods: A handbook for Research on Interactions New York, NY: Irvington, 1977 Crow R, Gage H, Hampson S, Hart J, Kimber A, Thomas H The role of expectancies in the placebo effect and their use in the delivery of health care: A systematic review Health Technol Assess 1999;3:3 Dance KA, Neufeld RWJ Aptitude-treatment interaction research in the clinical setting: A review of attempts to dispel the patient uniformity myth Psychol Bull 1988;104:192213 Di Blasi Z, Harkness E, Ernst E, Georgiou A, Kleijnen J Influence of context effects on health outcomes: A systematic review Lancet 2001;357:757762 Edwards AL, Cronbach LJ Experimental design for research in psychotherapy J Clin Psychol 1952:8;5159 Goldie JH Drug resistance in cancer: A
perspective Cancer Metastasis Rev 2001;20:6368 Gregerson MB, Roberts IM, Amiri MM Absorption and imagery locate immune responses in the body Biofeedback Self Regul 1996;21:149165 Grünbaum A Explication and implications of the placebo concept In: White L, Tursky B, Schwartz GE eds Placebo: Theory, Research, and Mechanisms New York: Guilford Press, 1985:936 Institute of Medicine Crossing the Quality Chasm: A New Health System for the 21st Century Washington, DC: National Academy Press, 2001 Kalauokalani D, Cherkin DC, Sherman KJ, Koepsell TD, Deyo RA Lessons from a trial of acupuncture and massage for low back pain: Patient expectations and treatment effects Spine 2001;26:14181424 Karpman MB The Johnson-Neyman technique using SPSS or BMDP Educ Psychol Meas 1983;43:137147 Kiesler DJ Some myths of psychotherapy research and the search for a paradigm Psychol Bull 1966;65:110136 Luborsky L, Singer B, Luborsky L Comparative studies of psychotherapies: Is it true that everybody has won and all
must have prizes? Arch Gen Psychiatry 1975;32:9951008 Mitchell A, Cormack M Towards a model of treatment in health care In: The Therapeutic Relationship in Complementary Health Care Edinburgh: Churchill Livingstone, 1998:2942 National Center for Complementary and Alternative Medicine Integration of complementary and alternative medicine: A health services research perspective RFA: AT-01-001, 2000 Online document at: grants2nihgov/grants/guide/rfa-files/RFA-AT01-001html National Institute of Mental Health Psychotherapeutic interventions: How and why they work A workshop by the National Institute of Mental Health Rockville, Maryland, 2002 Online document at wwwnimhnihgov/research/interventionscfm Neff DF, Blanchard EB, Andrasik F The relationship between capacity for absorption and chronic headache patients response to relaxation and biofeedback treatment Biofeedback Self Regul 1983;8:177183 Prochaska JO How do people change, and how can we change to
REFERENCES
Barlow DH, OBrien FT,
Last CG Couples treatment of agoraphobia: initial outcome In: Craig KD, McMahon RJ, eds Advances in Clinical Behavior Therapy New York: Brunner/ Mazel, 1984 Bell IR, Caspi O, Schwartz GRE, Grant KL, Rychener D, Gaudet TW, Maizes V, Weil AT Integrative medicine and systematic outcomes research: Issues in the emergence of a new model for primary health care Arch Intern Med 2002;162:133140 Beutler LE Have all won and must all have prizes? Revisiting Luborsky et als verdict J Consult Clin Psychol 1991;59: 226232 Carr AJ, Gibson B, Robinson PG Measuring quality of life: Is quality of life determined by expectations or experience? BMJ 2001;322:12401243 Caspi O Integrated medicine: Orthodox meets alternative Bringing complementary and alternative medicine CAM into mainstream is not integration BMJ 2001;322:168 Caspi O, Bootzin RR Evaluating how placebos produce change: Logical and causal traps and understanding cognitive explanatory mechanisms Eval Health Prof 2002;25:436464 Chobanian AV,
Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL, Jones DW, Materson BJ, Oparil S, Wright JT, Roccella EJ, The National High Blood Pressure Euducation Program Co-
586
help many more people? In: Hubble DB, Miller, SD, ed The Heart Soul of Change: What Works in Therapy Washington, DC: American Psychological Association, 1999:227255 Qualls PJ, Sheehan PW Capacity for absorption and relaxation during electromyograph biofeedback and no-feedback conditions J Abnorm Psychol 1979;88:652662 Qualls PJ, Sheehan PW Imagery engagement, absorption capacity, and relaxation during electromyograph biofeedback J Pers Soc Psychol 1981;41:370379 Redelmeier DA, Shafir E Medical decision making in situations that offer multiple alternatives JAMA 1995;273:302305 Ridenour TA, Treloar JH, Biner PM, Henriksen LW, Dean RS Attribute-treatment interaction research using nonexperimental data and the allocation average Int J Neurosci 1999;99:1947 Rogosa D Comparing nonparallel regression lines Psychol Bull
1980;88:307321 Rohrbaugh M, Shoam V, Racioppo MW Toward family-level attribute x treatment interaction ATI research In: Lidle HA, Santisteban D, Levant RF, Bray JH, eds Family Psychology: Science-Based Interventions Washington, DC: American Psychological Association, 2002 Room R Intoxication and bad behavior: Understanding cultural differences in the link Soc Sci Med 2001;53:189198 Shea JD Effects of absorption and instructions on heart rate control J Mental Imagery 1985;9:87 Shoam V, Rohrbaugh M Aptitude x treatment interaction ATI research: Sharpening the focus, widening the lens In: Aveline M, Shapiro DA, eds Research Foundation for Psychotherapy Practice Sussex, England: John Wiley Sons, 1975:7395 Shoam V, Rohrbaugh M Promises and perils of empirically supported psychotherapy integration J Psychother Integr 1996;6: 191206
CASPI AND BELL
Shoam-Salomon V Introduction to special section on client-therapy interaction research J Consult Clin Psychol 1991;59: 203204 Shoam-Salomon V,
Avner R, Neeman R Youre changed if you do and changed if you dont: Mechanisms underlying paradoxical interventions J Consult Clin Psychol 1989;57:590598 Shoam-Salomon V, Hannah MT Client-treatment interaction in the study of differential change processes J Consult Clin Psychol 1991;59:217225 Smith B, Sechrest L Treatment of aptitude x treatment interactions J Consult Clin Psychol 1991;59:233244 Snow RE Aptitude processes In: Snow RE, Federico P, Montague WE, eds Aptitude, Learning, and Instruction: Vol 1: Cognitive Processes Analyses of Aptitude Hillsdale, NJ: Erlbaum, 1980 Snow RE Aptitude-treatment interaction as a framework for research on individual differences in psychotherapy J Consult Clin Psychol 1991;59:205216 Wallston KA, Stein MJ, Smith CA Form C of the MHLC scales: A condition-specific measure of locus of control J Pers Assess 1994;63:534553
Address reprint requests to: Opher Caspi, MD, PhD Program in Integrative Medicine Department of Medicine College of Medicine The
University of Arizona–Health Sciences Center PO Box 245153 Tucson, AZ 85724-5153 E-mail: ocaspi@ahscarizonaedu
Source:hpna.org
