CHAPTER 1 3

CONTEMPORARY CLINICAL INTERVIEWING: INTEGRATION OF THE DSM-IV, MANAGED CARE CONCERNS, MENTAL STATUS, AND RESEARCH

Shawn Christopher Shea

INTRODUCTION DSM-IV, to more classic psychodynamic approaches and engagement skills. This clinical Interviewing is the backbone of all mental health challenge has been made even more difficult by yet professions. It is a dynamic and creative process, another new influence, the powerful presence of which represents a somewhat elusive set of skills. managed care and the constant ticking of "the The importance of this set of skills has been high- clock" concerning the number of sessions avail- lighted by Langsley and Hollender (1982). Their able to the client. In the past a skilled clinican survey of 482 psychiatric teachers and practitio- could perform a sound diagnostic assessment ners revealed that 99.4 percent ranked conducting within an hour, although many chose to take a comprehensive interview as an important longer. The difference is that today the clinician requirement for a psychiatrist. This represented the does not have a choice; managed-care principles highest ranking of 32 skills listed in the survey. dictate that he or she must complete the assessment Seven of the top 10 skills were directly related to within an hour and subsequently rapidly write up interviewing technique, including skills such as the the document as well. assessment of suicide and homicide potential, the Such a daunting integrative task, performed ability to make accurate diagnoses, and the ability under tight time constraints, can represent a major to recognize countertransference problems and hurdle for the developing clinician. This educa- other personal idiosyncrasies as they influence tional expectation was somewhat wryly stated by interactions with patients. These results were repli- Sullivan (1970) decades ago when he wrote: "The cated in a follow-up survey (Langsley & Yager, psychiatric expert is presumed, from the cultural 1988). definition of an expert, and from the general It can be seen from this list that the contempo- rumors and beliefs about psychiatry, to be quite rary clinician is being asked to combine an impres- able to handle a psychiatric interview." But the sive list of complex skills, ranging from structuring ability to handle the initial assessment interview techniques and diagnostic explorations using the has become a considerably more complicated task

339 340 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT since the time of Sullivan's quote, for there has By utilizing these two concepts, several inter- been an evolution in psychiatry and mental-health view types can be defined. In the free-format inter- care of immense proportions in the past 40 years. view, the interviewer has little standardization of This chapter is about this ongoing evolution and database and is highly interested in the spontane- its impact on assessment interviewing. Perhaps the ous content produced by the patient. Such free-for- single most striking legacy of the evolution is the mat interviews place little emphasis on scheduling disappearance of the psychiatric interview. Instead and tend to follow the natural wanderings of the of a single style of interviewing, the contemporary patient. These interviews are valuable for uncover- clinician must learn to perform an impressive array ing patient psychodynamics and revealing patient of interviews suited to the specific clinical task at feelings, opinions, and defenses. hand, including assessments as diverse as those At the opposite end of the spectrum is the fully required in an emergency room; an inpatient unit; a structured interview that is highly standardized and psychotherapy practice, consultation and liaison strictly scheduled. In fully structured interviews setting; and a managed-care clinic. the required informational areas are specified in This chapter is designed for both academicians detail and the ways of exploring them are also pre- interested in the theoretical and research underpin- scribed. An example of this type of interview is the nings of the interview process and clinical students Diagnostic Interview Schedule (Robins, Helzer, concerned with practical interviewing techniques. Croughan, & Ratcliff, 1981), developed for com- It makes no attempt to be an exhaustive overview; munity surveys by lay interviewers. instead, the reader is provided with a conceptual Semistructured interviews represent procedures guide that provides a wealth of references for more in which the informational areas to be explored are in-depth study. specified, but the sequence and wording to be used The following areas are discussed: (a) an histor- in data gathering are only moderately predeter- ical overview and description of the influences that mined. In these interviews, general guidelines have shaped the evolution of clinical interviewing about the interview sequence, such as beginning mentioned earlier; (b) a practical introduction to with the chief complaint and following with epi- two of the major clinical cornerstones of current sodes of the present illness, may be provided, but assessment interviewing: the mental status exami- the clinician is given some latitude to move within nation and the DSM-IV; and (c) a review of some this framework. Semistructured interviews are of of the major research efforts with regard to inter- value in both research and clinical settings. They viewing, including clinician phrasing of responses, frequently can provide standardized databases as nonverbal concerns, alliance issues and empathy, pioneered by Mezzich in the Initial Evaluation structured interviews, and educational research. Form (Mezzich, Dow, Rich, Costello, & Himmel- Before proceeding it will be of use to define a hoch, 1981; Mezzich, Dow, Ganguli, Munetz, & few terms that clarify many of the complicated Zettler-Segal, 1986). issues regarding interviewing style. The style of The last major type of interview is the flexibly any specific clinical or research interview is structured interview. The flexibly structured inter- greatly determined by the following structural fac- view represents the most popular clinical inter- tors: (a) specific content areas required to make a view, and when performed by an experienced clinical decision or to satisfy a research data base, clinician, holds promise as a research tool. With (b) quantity of data required, (c) importance placed the flexibly structured interview, the clinician has a on acquiring valid historical and symptomatic data standardized database (pre-determined by the clin- as opposed to patient opinion and psychodynamic ical or research task at hand) but is given total free- understanding, and (d) time constraints placed dom in scheduling. The interview begins with a upon the interviewer. free-format style in which the clinician moves with With regard to these structural concerns of the whatever topics appear to be most pressing for the interview, two concepts outlined by Richardson, patient. Once the engagement is secured the clini- Dohrenwend, and Klein (1965) are useful: stan- cian begins to structure the interview sensitively. dardization and scheduling. Standardization refers With flexibly structured interviews the actual to the extent to which informational areas or items scheduling will be relatively unique to each clini- to be explored are specified in the interview proce- cian-patient dyad, for the interviewer fluidly alters dure. Scheduling refers to the prespecification of the style of scheduling to gather the standardized the wording and sequence of the interview process. database most effectively while working with the CONTEMPORARY CLINICAL INTERVIEWING 341 specific needs and defenses of the patient. These his interests would move forward both the flee-for- interviews require a high degree of sophistication mat style and a more semistructured approach. from the clinician and allow him or her to insert Meyer professed a psychobiological approach to areas of free format and dynamic questioning the patient, in which it was deemed important to whenever expedient. Most experienced clinicians, determine a "biography" of the patient that whether consciously or by habit, utilize a flexibly included biological, historical, psychological, and structured format. The complexities and nuances social influences on the patient's current behavior of the flexibly structured clinical interview have (Kaplan, Freedman, & Sadock, 1980). His interest been most recently explored in detail by Shea in psychological and social influences further (1998) and Othmer and Othmer (1994). A compre- advanced a style of interviewing in which there hensive annotated bibliography on the literature was an appreciation for the value of the flee-format concerning clinical interviewing and training style (Siassi, 1984). appears in Core Readings of Psychiatry (Shea, On the other hand, Meyer' s interest in determin- 1995). ing a sharp conceptualization of biological influ- Historically, clinical interview styles have var- ences as well as a clear presentation of the patient' s ied in popularity; they have ranged from semis- immediate symptomatology moved him toward an tructured interviews that were partially based on appreciation of semistructured or flexibly struc- the medical model to more free-form analytic tured formats. For instance, Meyer believed that interviews and flexibly structured styles. It is to the clinician should begin the interview with a this evolution that attention is now turned. careful exploration of the patient' s chief complaint (Kaplan, Freedman, & Sadock, 1980). In his work "Outlines of Examinations" (Meyer, 1951), which HISTORICAL FOUNDATIONS was printed privately in 1918, Meyer was the first to define the term "mental status" (Donnelly, When studying the historical evolution of the Rosenberg, & Fleeson, 1970). interview, it is helpful to look for underlying prin- By the end of the first quarter of the century ciples of development. Perhaps the most useful many of the major components of the psychiatric principle is that interview styles tend to evolve out interview had been established. These key content of whatever theoretical knowledge base is most regions included chief complaint, history of the popular in a given age. In particular, the more present illness, social history, family history, med- numerous and syndrome-specific the available ical history, and mental status. All of these were treatment modalities are, the more likely it is that a related to an underlying attempt to arrive at a diag- standardized database will be sought. If the stan- nostic overview. But a diagnostic system based on dardized database requirements become large, mutually agreed-upon criteria was not well estab- there is a gradual shifting toward methods of struc- lished, and consequently, most of the interview turing, whether done by rigid schedule or by flexi- was not directed primarily toward establishing a ble maneuvering. This relationship between the specific diagnosis. availability of treatment modality and interview Such lack of diagnostic specificity, coupled with style is seldom noted but represents a powerful and a relative paucity of treatment interventions, unifying historical principle. resulted in a database that did not require a high Early in the century the approach to clinical degree of scheduling. In the first place, because assessment was rooted in the medical model. Krae- there were few diagnostic-related interventions, pelin had attempted to classify mental illnesses and there was not a pressing need to complete the ini- indeed had been able to differentiate manic depres- tial assessment quickly. The clinician could spend sion from dementia praecox (Kaplan, Freedman, & many hours over many days eliciting data for the Sadock, 1980). Although there was not an abun- initial interview. In the second place, the diagnos- dance of treatment modalities present, the gestalt tic schema were so limited that there was not a sig- of the moment was toward a careful detailing of nificant need to cover large areas of symptomatol- behaviors and symptoms in an effort to determine ogy quickly. The resulting relative lack of specific syndromes and diseases. scheduling and structure was to have a major thrust At this time the gifted psychiatrist and educator toward even more emphasis on free format. Adolf Meyer proved to be a catalyst in the devel- Psychoanalysis arrived on the shores of America opment of the psychiatric interview. Paradoxically, like a native-born son. By the 1940s it had become 342 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT well established. Freud's pioneering work had an as they revealed themselves in the initial interview enormous impact on interviewing technique. His and subsequent therapy. Few, if any, books basic theories seemed to move away from empha- describe more lucidly and insightfully the subtle sis on diagnosis in a medical sense toward a more relationships between patient defense mechanisms probing investigation of actual psychological pro- and clinician style. cesses. With the development of ego psychology In the early 1950s another major force was to and a further investigation of defense mechanisms have an impact on the psychiatric interview. That by theorists such as Heinz Hartman and Anna force would be a single man: Harry Stack Sullivan. Freud, the emphasis further shifted toward an During his life, Sullivan proved to be one of the understanding of how the patient's defenses were most gifted interviewers of all time. His book The manifested in the context of the interview itself. Psychiatric Interview was published posthumously Interviewing and therapy seemed to become less in 1954 (Sullivan, 1970). The book would estab- distinct. lish forever the importance of the interpersonal A free-format style of interviewing became matrix as one of the major areas through which to more common. Clinicians became increasingly understand the interview process. Sullivan stressed aware of the value of spontaneous speech as a fer- the importance of viewing the interview as a socio- tile ground for uncovering patient defenses and logical phenomenon in which the patient and the conflicts. The elicitation and description of these clinician form a unique and dynamic dyad, with defenses and a basic description of the patient's the behavior of each affecting the other. ego structure became goals of the interview. One of Sullivan's key terms was participant Important advances in interviewing technique observation. This concept emphasized the need of evolved during this time. Emphasis was placed not the interviewer to "step aside" during the interview only on what the patient said but what the patient itself in the sense of viewing his or her own behav- either consciously or unconsciously did not say. ior and the impact of that behavior upon the Resistance came to be seen as a golden door for patient. Sullivan saw that the measuring instrument entering the dynamics and conflicts of the patient. itself, in this case the interviewer, could actually A free-format style of interview provided a rich change the database, that is, the patient' s behaviors psychological milieu in which to observe directly and degree of distortion in relating symptomatol- the maneuverings of the patient's unconscious ogy. defenses. Sullivan was also one of the first interviewers to Several books helped clinicians to adapt to this emphasize the importance of structuring, and he new emphasis in interviewing style. One was Lis- discussed specific methods of making transitions tening With the Third Ear by Theodore Reik during the interview from one topic to another. In (1952). In the section entitled "The Workshop," this sense Sullivan recognized the importance of Reik provides a variety of insights concerning free-format style as well as scheduling issues, and issues such as freefloating attention, conscious and essentially developed a flexibly structured style of unconscious observation, and the therapist-patient interviewing in which these various techniques alliance. could be intermixed at the will of the clinician. Another important analytic contribution was The Near the time of Sullivan's book, another work Clinical Interview (Vols. I & 2) written by Felix appeared entitled The Initial Interview in Psychiat- Deutsch and William Murphy (1955a, 1955b). ric Practice by Gill, Newman, and Redlich (1954). Working out of Boston, Deutsch and Murphy This work was strongly influenced by Sullivan's described the technique of associative anamnesis. interpersonal perspective. Innovatively, the book This technique emphasizes a free-format style in includes three fully annotated transcripts of inter- which free association and gentle probing by the views which were accompanied by phonographic clinician open a window into the symbolic world records of the actual patient/physician dialogue. It that lies "between the lines" of the patient's report. also contains an excellent history of interviewing But perhaps the most influential book dedicated technique. to interviewing from an analytic point is the classic With regard to the interpersonal perspective, text, The Psychiatric Interview in Clinical Prac- Sullivan's classic text had been predated by J. C. tice, by MacKinnon and Michels (1971). This Whitehorn. In 1944 Whitehorn published an influ- book provided an easily read yet highly rewarding ential article in the Archives of Neurology and Psy- introduction to understanding dynamic principles chiatry entitled, "Guide to Interviewing and CONTEMPORARY CLINICAL INTERVIEWING 343

Clinical Personality Study." One of Whitehorn's verbal and nonverbal, which allow the clinician to contributions lay in his emphasis on eliciting relate favorably to the interviewee. patient opinion as both a powerful engagement Carl Rogers represents one of the most powerful technique, and a method of looking at unconscious influences in this regard. His "client-centered dynamics. In particular, the patient' s opinions con- approach" emphasized empathic techniques. He cerning interpersonal relations and reasons for car- described empathy as the clinician's ability "to ing for others represented major avenues for perceive the intemal frame of reference of another exploration. with accuracy, and with the emotional components Closely related to the analytic and interpersonal and meanings which pertain thereto, as if one were schools was the European-based school of phe- the other person, but without ever losing the 'as if' nomenological psychiatry and psychology. Giants condition (Rogers, 1951, 1959)." in the field during the first half of this century, such Rogers is also well known for his concept of as Karl Jaspers and Medard Boss, emphasized an "unconditional positive regard." A clinician con- approach to the patient in which the focus was on veys this value when he or she listens without pass- developing an understanding of the exact ways in ing judgment on the patient's behaviors, thoughts, which the patient experienced "being in the world" or feelings. These ideas were pivotal in conveying (Hall & Lindzey, 1978). In this approach, while the idea that the clinician should not appear remote utilizing a phenomenological style of interview, or distant during the interviewing process, for such the clinician delicately probes the patient for care- artificial remoteness could seriously disengage ful descriptions of the patient's symptoms, feel- patients. Interviewers were allowed to utilize in a ings, perceptions, and opinions. Through a shared naturalistic sense their social skills and personality. process of precise questioning and at times, self- Other counselors, such as Alfred Benjamin transparency, the clinician arrives at a vivid picture (1969) in The Helping Interview and Gerard Egan of the patient's universe, a picture which some- (1975) in The Skilled Helper, carried on this tradi- times even surprises the patient as defenses and tion of emphasizing genuineness and common distortions are worked through by the clinician's sense in the therapeutic relationship. Benjamin style of questioning. emphasized the need to develop a trusting relation- In more recent years, authors such as Alfred ship with the patient, avoiding the tendency to hide Margulies and Leston Havens have reemphasized behind rules, position, or sense of authority. Egan the importance of a phenomenological approach attempted to help clinicians develop these abilities (Havens, 1978, 1979; Margulies, 1984; Margulies by describing a concrete language with which to & Havens, 1981). A particularly fascinating tech- help convey these ideas in an educational sense, nique, known as counterprojection, has been highlighted by a self-programmed manual to described by Havens. The counterprojective tech- accompany his text. Most of the interviewing tech- nique deflects paranoid projections before they niques developed by these authors and other coun- manifest onto the interviewer. Such techniques are selors are distinctly non-diagnostically focused. valuable in consolidating an alliance with fright- Consequently, as one would expect, they tend to be ened, hostile, angry, or actively paranoid patients neither highly standardized nor scheduled. (Havens, 1980). But the fields of counseling and psychology did In summary, it can be seen that during the mid- not ignore the importance of the database. To the dle years of this century and later, psychiatrists contrary, an emphasis on developing an increas- from the analytic, interpersonal, and phenomeno- ingly sophisticated understanding of the impact of logical schools exerted a strong influence on inter- interviewing technique on the validity of data, viewing technique. The next impact would come grew out of the empirical studies and behavioral from a nonmedical tradition. approaches pioneered by nonmedical researchers. In the 1950s, 1960s, 1970s, and 1980s the fields For example, Richardson, Dohrenwend, and of psychology and counseling had an enormous Klein (1965) whose schema of standardization and impact on clinical interviewing. More than psychi- scheduling was mentioned earlier, produced a par- atry, these fields emphasized the need for empiri- ticularly incisive work entitled Interviewing: Its cal research concerning the flexibly structured Forms and Functions. The power of the text lies in clinical interview, which will be discussed in more the authors' attempts systematically to define and detail later. These research approaches opened up study various characteristics of the interview pro- an increased awareness of the specific factors, both cess ranging from the style of questioning (such as 344 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT open-ended versus closed-ended) to the impact of Ironically, the most powerful forces to operate patient and clinician characteristics on the inter- on interviewing style in the last several decades viewing process. Their work transformed a process were not directly related to attempts to advance that was heretofore somewhat nebulous in nature interviewing per se but were related to the diagnos- into a process that could be studied behaviorally. tic and therapeutic advances occurring within psy- Another psychologist, Gerald Pascal (1983), chiatry proper and the other disciplines of mental- described a technique known as the behavioral health care. The evolution of the psychiatric inter- incident. Although simple in nature, this technique view was the direct result of a revolution in three represents one of the most significant and easily areas: (1) treatment modalities, (2) diagnostic sys- taught of all interviewing techniques in the last tems, and (3) managed-care principles. several decades. The technique is based on the Concerning the first factor, in the past 30 years premise that questions can range on a continuum an impressive array of new therapeutic interven- from those that request patient opinion to those that tions has emerged. These revolutionary advances ask for historical or behavioral description. The include modalities such as tricyclic antidepres- latter style of questioning is more apt to yield valid sants, serotonin-selective reuptake inhibitors, information, whereas questions which request mood stabilizers, antipsychotic medications, cog- patient opinion are dangerously prone to patient nitive behavioral therapy, family therapy, group distortion. therapy, and more sophisticated forms of dynamic and behavioral therapies such as interpersonal- The behavioral incident provides a more reliable dynamic psychotherapy and dialectical behavioral tool for exploring areas of particular sensitivity therapy, to name only a few. In the same fashion where patient distortion may be high, as in the that the rapid acceptance of analytic thinking assessment of suicide potential, child abuse, sub- resulted in further development of the free-format stance abuse, and antisocial behavior. For exam- style of interviewing, these new interventions, ple, a clinician may phrase a lethality probe in this which frequently are chosen in relation to a DSM- fashion, "Have you ever had any serious suicide IV diagnosis, have led interviewers to reexamine attempts?" It is then up to the patient to interpret the importance of developing both a thorough and the notion of what constitutes a "serious" attempt. valid data base. To the patient, an overdose of 20 pills may not The development of new treatment interventions seem serious and consequently may not be directly spawned the second major force molding reported. Using the behavioral-incident technique the contemporary interview. Researchers and clini- the clinician asks a series of questions focused cians quickly realized that better diagnostic sys- directly on patient behaviors: "Tell me exactly tems, which would decrease variability and what methods of killing yourself you have ever unreliability, needed to be developed, for treatment tried, even if only in a small way," and "When you modalities were being increasingly determined by took the pills how many did you take?" With these diagnosis. types of questions the patient is asked to provide One of the most influential of the modern diag- concrete information. It is up to the clinician to nostic systems that resulted was the Feighner crite- arrive at an opinion as to what is "serious". ria (Feighner et al., 1972). These criteria were Another significant book concerning the specific developed in the Department of Psychiatry at phrasing of questions with regard to their impact Washington University in St. Louis. This system on data gathering was The Structure ofMagic I by delineated 15 diagnostic categories by using both Grinder and Bandler (1975). Although some of exclusion and inclusion criteria. Building on this their latter work has been controversial in nature, base, Spitzer, Endicott, and Robins (1978) devel- this early volume was sound, penetrating, and to oped the Research Diagnostic Criteria (RDC). the point. They described a variety of techniques With the RDC system the psychopathological for phrasing questions in such a manner that the range was increased to include 23 disorders. patient's hidden thoughts would be gradually Of particular note to the history of interviewing pulled to the surface. The work is based on an was the subsequent development of a semistruc- understanding of transformational grammar and is tured interview designed to delineate the diagnoses enhanced by the self-programmed layout of the described by the Research Diagnostic Criteria. book, which literally forces the reader to make This interview, the Schedule for Affective Disor- actual changes in style of questioning. ders and Schizophrenia (SADS), was developed by CONTEMPORARY CLINICAL INTERVIEWING 345

Endicott and Spitzer (1978). It was a powerful tool DSM-IV differential into an equally sound biopsy- with good reliability and it became popular as a chosocial evaluation remained daunting. research instrument. A second interview that was The task would prove to be further complicated both highly standardized and rigidly scheduled by the advent of a philosophical/economic para- was the Diagnostic Interview Schedule (DIS) digm shift, representing the third significant factor developed by Robins (Robins, Helzer, Croughan, molding the contemporary psychiatric interview. & Ratcliff, 1981). This interview was designed to Managed care, an approach that gained enormous be used by lay interviewers and hence was highly power in the early 1990s, placed a heavy emphasis scheduled to ensure interrater reliability. upon efficient use of resources. In its healthy func- The semistructured and structured formats dis- tioning it pushes clinicians to always work with a played by the SADS and the DIS, respectively, sound cost-mindfulness. In its unhealthy function- were not overly popular with clinicians. Such lack ing it results in inadequate treatment, sometimes of enthusiasm demonstrated that even though clini- caused by clinicians missing critical treatable diag- cians were progressively required to obtain a noses in the initial session, in an effort to prema- highly standardized database, the method to turely begin therapy and cut corners. For the most achieve this goal while flexibly engaging the part, clinicians no longer have an option; the initial patient and handling resistance was not clear. diagnostic session must be completed in sixty min- The movement toward the need for a highly utes with roughly another 30 minutes allotted, if standardized database with regard to diagnostic the clinician is lucky, for the written document. information was given further momentum in the In order to determine a correct DSM-IV diagno- United States by the publication of the third edition sis, perform a sound biopsychosocial assessment, of the Diagnostic and Statistical Manual of Mental and spot the client' s strengths to capitalize upon in Disorders (DSM-III) by the American Psychiatric brief therapy formats, the contemporary inter- Association (1980). This manual emphasized a viewer must gather an amount of concrete informa- multi-axial approach which will be described in tion in 60 minutes which might have seemed quite more detail later in this chapter. Seven years later unmanageable to an interviewer of 40 years ago. the revised edition, the DSM-III-R, appeared Consequently, interviewers have reexamined their (APA, 1987) and was followed by the DSM-IV approaches to scheduling, moving toward partially (APA,1994). With the advent of these widely scheduled interviews, as seen in the semistructured accepted diagnostic systems, interviewers were format, or toward a method of tracking the data- faced with the necessity of gathering sensitively base while maximizing interviewer spontaneity, as the data that would be required for a sophisticated seen in the flexibly structured format. differential diagnosis. This would prove to be no The lead toward resolving some of the complex easy task. integrative tasks facing the contemporary psychiat- From the arena of semi-structured interviews, ric clinician came from the Western Psychiatric this task was approached through the development Institute and Clinic at the University of Pittsburgh. of ever improving and more "user friendly" for- In 1985, Hersen and Turner edited an innovative mats such as the Structured Clinical Interview for book entitled Diagnostic Interviewing. This book the DSM-III (the SCID-III, SCID-III-R, and the represented one of the first attempts to acknowl- SCID-IV), developed by Spitzer and Williams edge fully that interviewers should become famil- (1983). The SCID-IV has seen wide use, but has iar with specific techniques for sensitively some limitations due to its length and its restriction exploring the diagnostic criteria from the various to the DSM-IV system. The Mini-International diagnoses in the DSM-III. To accomplish this edu- Neuropsychiatric Interview (M.I.N.I.) pioneered in cational task, various experts contributed chapters the United States by Sheehan and in France by on a wide range of DSM-III categories from Lecrubier (1999) has seen wider international schizophrenia to sexual disorders. acceptance. The M.I.N.I. is elegant, practical and At the same time, also at the Western Psychiatric remarkably brief to administer with a median dura- Institute and Clinic, a variety of innovations, both tion of 15 minutesmall qualities that have with regard to interviewing technique and inter- enhanced its acceptance by clinicians. But from a viewing training, were presented in Psychiatric practical front-line clinician's standpoint, who Interviewing: The Art of Understanding (Shea, must arrive at much more than merely a diagnosis 1988). This book represented the first attempt to in 60 minutes, the task of integrating a sound synthesize smoothly the divergent streams of inter- 346 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT viewing knowledge developed in the various men- skill level of the clinician easily documented for tal-health fields over the previous 50 years. In quality assurance purposes. particular, it acknowledged the confusing task fac- While these advances were being made, an out- ing contemporary clinicians who had to synthesize standing and highly influential textbook, The Clin- a wide range of important information (including ical Interview Using DSM-III-R was written by regions such as the chief complaint, history of the Othmer and Othmer (1989) which, more than any present illness, the social history, the family his- other textbook on interviewing, firmly secured the tory, the medical history, the mental status, and the DSM-III-R system as an accepted clinical tool by DSM-III-R diagnostic regions) into an interview front-line clinicians. This text, filled with practical that was naturally flowing and to which appropri- tips and model questions, has secured itself as a ate energy could be given to dynamic issues and classic. It was later expanded to two volumes to resistance concerns. further address interview situations involving com- To accomplish this task, Shea developed a plicated clinical interactions with borderline and supervision language that would provide a readily resistant patients (Othmer and Othmer, 1994). understandable system with which to study the Other outstanding inteviewing texts soon followed structuring, flow, and time management of the including Morrison' s The First Interview: A Guide interview. This widely accepted supervision sys- for Clinicians (1993), Shea's Psychiatric Inter- tem would later prove to be of immediate value in viewing: The Art of Understanding, 2nd edition graduate training across disciplines and in the (1998), Sommers-Flanagan's Clinical Interview- development of quality assurance programs in the ing, 2nd edition (1999), and Carlat's The Psychiat- fast-paced world of the managed-care asssessment. ric Interview: A Practical Guide (1999). This study of the structure and flow of the inter- From a historical perspective, clinical interview- view process was named facilics (Shea, 1988) ing has continually evolved and undoubtedly will derived from the Latin root facilis (ease of move- continue to change as clarifying theories and treat- ment). The study of facilics emphasized a rigorous ment modalities grow in number and depth. Clini- examination of the overall structuring of the inter- cians have been forced to cope with the realization view as it related to time constraints and clinical that the contemporary clinical interview frequently tasks. To enhance learning further, a schematic requires a high degree of standardization, as exem- supervision system was designed, made up of sym- bols which depict the various transitions and types plified by the demand for larger amounts of data, of topical expansions utilized by the trainee. This of both diagnostic and psychosocial importance, to shorthand system clarifies educational concepts be gathered in relatively short periods of time. At and highlights structural elements, while present- first it appeared that these requirements would ing an immediately understandable and permanent necessitate clinical interviews to be tightly sched- record of what took place in the interview. uled or semistructured in nature. But with the In subsequent years facilic principles were advent of sensitive structuring approaches, such as applied to specific interviewing tasks such as the facilics, clinicians remain free to utilize flexibly elicitation of suicidal ideation (Shea, 1998). This structured styles of interviewing, as exemplified by resulted in the development of innovative inter- the CASE strategy and the work of Othmer and viewing strategies such as the Chronological Othmer. Such styles provide clinicians with meth- Assessment of Suicide Events (the CASE ods of gathering thorough data-bases in relatively Approach) and its counterpart for uncovering vio- short periods of time, scheduling the interview as lent ideation the Chronological Assessment of they go along, each interview representing a Dangerous Events (the CADE Approach). The unique creative venture. CASE Approach, delineated in The Practical Art With the historical review completed it is valu- of Suicidal Assessment (Shea, 1999) was a flexible, able to provide a practical introduction to two of practical, and easily learned inteview strategy for the most powerful influences mentioned earlier. eliciting suicidal ideation, planning, and intent. It The first influence, which dates back to the pio- was designed to increase validity and decrease neering work of Adolf Meyer, is the mental status. potentially dangerous errors of omission. Because The second, and much more recent influence, is the techniques of the CASE Approach were behav- development of the DSM-III, the DSM-III-R, and iorally concrete it could be readily taught and the the DSM-IV diagnostic systems. CONTEMPORARY CLINICAL INTERVIEWING 347

THE MENTAL STATUS EXAMINATION sentation. One place to start is with a description of the patient's clothes and self-care. Striking charac- The mental status represents an attempt to teristics such as scars and deformities should be describe objectively the behaviors, thoughts, feel- noted, as well as any tendencies for the patient to ings, and perceptions of the patient during the look older or younger than his or her chronological course of the interview itself. These observations age. Eye contact is usually mentioned. Any pecu- are usually written as a separate section of the patient' s evaluation. The general topics covered by liar mannerisms are noted, such as twitches or the the mental status are categorized as follows: patient's apparent responses to hallucinations, appearance and behavior, speech characteristics which may be evident through tracking move- and thought process, thought content, perception, ments of the eyes or a shaking of the head as if mood and affect, sensorium, cognitive ability, and shutting out an unwanted voice. The clinician insight. Clinicians may vary on the exact catego- should note the patient' s motor behavior; common ries that are used, and some clinicians collect all of descriptive terms include restless, agitated, sub- these observations into a single narrative para- dued, shaking, tremulous, rigid, pacing, and with- graph. In any case, the clinician attempts to convey drawn. Displacement activities such as picking at a the state of the patient during the interview itself, cup or chain smoking are frequently mentioned. as if a cross-section were being taken of the An important, and often forgotten characteristic, is patient' s behavior for 60 minutes. the patient's apparent attitude toward the inter- In a sense, the mental status consists of a variety viewer. With these ideas as a guide, the following of different clinician activities ranging from obser- excerpt represents a relatively poor description. vation to the written record. Part of the mental sta- tus occurs informally as the clinician observes Clinician A: The patient appeared disheveled. Her various characteristics of the patient while the behavior was somewhat odd and her eye contact did patient spontaneously describes symptoms or his- not seem right. She appeared restless and her cloth- tory. The clinician may note whether the patient ing seemed inappropriate. appears to be shabbily dressed or able to concen- trate. Other aspects of the mental status are more Although this selection gives some idea of the formal in nature as the clinician asks direct ques- patient' s appearance, one does not come away with tions concerning areas of psychopathology, such as a sense of what it would be like to meet this patient. inquiries regarding hallucinations or delusions. Generalities are used instead of specifics. The fol- Finally, certain aspects of the mental status may be lowing description of the same patient captures her quite formalized as is seen during the formal cog- presence more aptly. nitive examination. During this part the patient is asked to perform tasks, such as calculations or Clinician B: The patient presents in tattered clothes, digit spans. all of which appear filthy. Her nails are laden with All of these clinician activities are synthesized dirt, and she literally has her soiled wig on back- into the written mental status. Indeed, it is by wards. She is wearing two wrist watches on her left examining the thought processes required to pro- wrist and tightly grasps a third watch in her right duce a sound written mental status that one can hand, which she will not open to shake hands. Her best discuss the more intangible processes at work arms and knees moved restlessly throughout the interview, and she stood up to pace on a few occa- during the "gathering of the mental status informa- sions. She did not give any evidence of active tion." Consequently, in this chapter each segment response to hallucinations. She smelled badly but did of the mental status is examined as it might appear not smell of alcohol. At times she seemed mildly in a standard written evaluation. An effort is made uncooperative. to summarize commonly utilized descriptive terms, to clarify confusing terms, to point out com- This passage presents a more vivid picture of her mon mistakes, and to provide an example of a behavior, a pattern that may be consistent with a well-written mental status. manic or psychotic process. Her "odd" behaviors have been concretely described. The clinician has Appearance and Behavior included pertinent negatives, indicating that she In this section the clinician attempts to describe shows no immediate evidence of hallucinating, as accurately the patient' s outward behavior and pre- might be seen in a delirium. 348 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT

Speech Characteristics and better, cherry tarts and Mom's hot breath keeps you Thought Process going and rolling along life's highways." If loosen- ing becomes extremely severe it is sometimes referred to as a word salad. The clinician can address various aspects of the patient's speech, including the speech rate, vol- Flight of ideas: In my opinion this is a relatively ume, and tone of voice. At the same time, the clini- weak term, for it essentially represents combinations cian attempts to describe the thought process of the of the above terms. This is why most trainees find it patient, as it is reflected in the manner with which confusing. For flight of ideas to occur, the patient must demonstrate tangential thought or a loosening the patient's words are organized. The term formal of associations in conjunction with a significantly thought disorder is utilized to suggest the presence pressured speech. Usually there are connections of abnormalities in the form and organization of between the thoughts, but, at times, a true loosening the patient's thought. The less commonly used of associations is seen. A frequently, but not always, term, content thought disorder, refers specifically seen characteristic of flight of ideas is the tendency for the patient's speech to be triggered by distracting to the presence of delusions, and is addressed in a stimuli or to demonstrate plays on words. When subsequent section of the mental status. The more present, these features represent more distinguishing generic term thought disorder includes both the hallmarks of a flight of ideas. Flight of ideas is com- concept of a formal thought disorder and of a con- monly seen in mania, but can appear in any severely agitated or psychotic state. tent thought disorder. In this section of the mental status the emphasis is on the process of the thought Thought blocking: The patients stop in mid-sentence (presence of a formal thought disorder), not the and never return to the original idea. These patients content of the speech. Terms frequently used by appear as if something has abruptly interrupted their clinicians include the following: train of thought and, indeed, something usually has, such as an hallucination or an influx of confusing ideation. Thought blocking is very frequently a sign Pressured speech: This term refers to an increased of psychosis. It is not the same as exhibiting long rate of speech, which may possibly best be described periods of silence before answering questions. Some as a "speech sans punctuation." Sometimes it is only dynamic theorists believe it can also be seen in neu- mildly pressured, whereas at other times, the rotic conditions, when a repressed impulse is threat- patient's speech may virtually gush forth in an end- ening to break into consciousness. less stream. It is commonly seen in mania, agitated psychotic states, or during extreme anxiety or anger. Illogical thought: The patient displays illogical con- clusions. This is different from a delusion, which Tangential thought: The patient's thoughts tend to represents a false belief but generally has logical rea- wander off the subject as he or she proceeds to take soning behind it. An example of a mildly illogical tangents off his or her own statements. There tends to thought follows: "My brother has spent a lot of time be some connection between the preceding thought with his income taxes so he must be extremely and the subsequent statement. An example of fairly wealthy. And everyone knows this as a fact because striking tangential thought would be as follows: "I I see a lot of people deferring to him." These conclu- really have not felt very good recently. My mood is sions may be true, but they do not necessarily logi- shot, sort of like it was back in Kansas. Oh boy, those cally follow. Of course, in a more severe form, the were bad days back in Kansas. I'd just come up from illogical pattern may be quite striking: "I went to the Army and I was really homesick. Nothing can Mass every Sunday, so my boss should have given really beat home if you know what I mean. I vividly me a raise. That bum didn't even recognize my reli- remember my mother's hot cherry tarts. Boy, they gious commitment." were good. Home cooking just can't be beat." Cir- cumstantial thought is identical in nature to tangen- In the following excerpt, the speech and thought tial thought but differs in that the patient returns to process of the woman with two watches on her the original topic. wrist is depicted. Once again the description dem- Loosening of associations: The patient's thoughts at onstrates some areas in need of improvement. times appear unconnected. Of course, to the patient, there may be obvious connections, but a normal lis- Clinician A: Patient positive for loosening of associ- tener would have trouble making them. In mild ations and tangential thought. Otherwise grossly forms, loosening of associations may represent within normal limits. severe anxiety or evidence of a schizotypal character structure. In moderate or severe degrees, it is gener- This clinician has made no reference to the ally an indicator of psychosis. An example of a mod- degree of severity of the formal thought disorder. erate degree of loosening would look like this: "I haven't felt good recently. My mood is shot, fluid Specifically, does this patient have a mild loosen- like a waterfall that's black, back home I felt much ing of associations or does she verge upon a word CONTEMPORARY CLINICAL INTERVIEWING 349 salad? Moreover, the clinician makes no reference find these obsessive thought processes to be both to her speech rate and volume, characteristics that odd and painful. They frequently have tried vari- are frequently abnormal in manic patients. The fol- ous techniques to interrupt the process. Common lowing brief description supplies a significantly themes for obsessions include thoughts of com- richer data base: mitting violence, homosexual fears, issues of right and wrong, and worries concerning dirt or Clinician B: The patient demonstrates a moderate filth. Obsessions may consist of recurrent ideas, pressure to her speech accompanied at times by loud thoughts, fantasies, images, or impulses. If the outbursts. Even her baseline speech is slightly louder clinician takes the time to listen carefully to the than normal. Her speech is moderately tangential, with rare instances of a mild loosening of associa- patient, bearing the above phenomenological tions. Without thought blocking or illogical thought. issues in mind, he or she can usually differenti- ate between ruminations and obsessions. Slowly one is beginning to develop a clearer pic- Delusions represent strongly held beliefs, that ture of the degree of this patient's psychopathol- are not correct or held to be true by the vast major- ogy. More evidence is mounting that there may be ity of the patient's culture. They may range from both a manic-like appearance and a psychotic pro- bizarre thoughts, such as invasion of the world by cess. In any case, the patient' s speech coupled with aliens, to delusions of an intense feeling of worth- her strikingly disheveled appearance, may lead the lessness and hopelessness. clinician to suspect that the patient is having trou- The fourth issue consists of statements con- ble managing herself. cerning lethality. Because all patients should be asked about current lethality issues, these issues should always be addressed in the written men- Thought Content tal status. In general, the clinician should make some statement regarding the presence of sui- This section refers primarily to four broad cidal wishes, plans, and degree of intent to fol- issues: ruminations, obsessions, delusions, and low the plans in an immediate sense. If a plan is the presence of suicidal or homicidal ideation. mentioned, the clinician should state the degree Ruminations are frequently seen in a variety of to which any action has been taken on it. He or anxiety states and are particularly common in she should also note whether any homicidal ide- depressed patients. Significantly depressed ation is present and to what degree, as with sui- patients will tend to be preoccupied with worries cidal ideation. and feelings of guilt, constantly turning the thoughts over in their minds. The thinking pro- Clinician A: The patient is psychotic and can't take cess itself does not appear strange to these care of herself. She seems delirious. patients, and they do not generally try to stop it. Instead, they are too caught up in the process to This excerpt is just simply sloppy. The first do much other than talk about their problems. In statement has no place in a mental status, for it is contrast, obsessions have a different flavor to the beginning of the clinician's clinical assess- them, although they may overlap with rumina- ment. The description of the delusion is threadbare tions at times. and unrevealing. The clinician has also omitted the An obsession is a specific thought that is questioning concerning lethality. Assuming the repeated over and over by the patient as if he or clinician asked but forgot to record this informa- she is seeking an answer to some question. tion, he or she may sorely regret this omission if Indeed, the patient frequently demonstrates this patient were to kill herself and the clinician obsessions over a question and its answer. As was taken to court to face a malpractice suit for soon as the question is answered, the patient possible negligence. A more useful description is feels an intense need to ask it again, as if some given below. process had been left undone. The patient may repeat this process hundreds of times in a row Clinician B: The patient appears convinced that if until it "feels right." If one interrupts the patient the watch is removed from her right hand, the world will come to an end. She relates that, consequently, while this process is occurring, the patient will she has not bathed for three weeks. She also feels frequently feel a need to start the whole process that an army of rats is following her and is intending again. Unlike the case with ruminations, patients to enter her intestines to destroy "my vital essence." 350 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT

She denies current suicidal ideation or plans. She ing information out of fear that the voices represent denies homicidal ideation. Without ruminations or a sickness. Numerous reasons exist for a patient to obsessions. avoid sharing the presence of hallucinations with a clinician, including instructions to the patient from With this description it has become clear that the the voices not to speak about them to the clinician. patient is psychotic, as evidenced by her delusions. It may be more accurate to state that the patient The next question is whether hallucinations play a denied having hallucinations rather than to report role in her psychotic process. categorically that the patient is without them. A more sophisticated report would be as follows:

Perception Clinician B: The patient denied both visual and auditory hallucinations and any other perceptual This section refers to the presence or absence of abnormality. hallucinations or illusions. It is of value to note that there is sometimes a close relationship between delusions and hallucinations. It is not uncommon Mood and Affect for the presence of hallucinations eventually to trigger the development of delusional thinking, but Mood is a symptom, reported by the patient, the two should not be confused. Let us assume that concerning how he or she has generally been feel- a patient is being hounded by a voice screaming, ing recently, and it tends to be relatively persistent. "You are possessed. You are a worthless demon." Affect is a physical indicator noted by the clinician If the patient refuses to believe in the reality of the as to the immediately demonstrated feelings of the voice, then one would say that the patient is hear- patient. Affect is demonstrated by the patient's ing voices but is not delusional. If, on the other facial expressions and other nonverbal clues dur- hand, the patient eventually begins to believe in the ing the interview itself; it is frequently of a tran- existence of the voice and feels that the devil is sient nature. Mood is a self-reported symptom; planning her death, then the patient is said to have affect is a physical sign. If a patient refuses to talk, developed a delusion as well. The following the clinician can say essentially nothing about description of perceptual phenomena is obviously mood in the mental status itself, except that the threadbare. patient refused to comment on mood. Later in the narrative assessment, the clinician will have ample Clinician A: Without abnormal perceptions. space to describe his or her impressions of what the patient's actual mood has been. In contrast to There is a question concerning the appropriate- mood, in which the clinician is dependent upon the ness in the mental status of using phrases such as, patient's self-report, the clinician can always say "grossly within normal limits" or "without abnor- something about the patient' s affect. mality." Generally, the mental status is improved by the use of more precise and specific descrip- Clinician A: The patient's mood is fine and her affect tions, but sometimes clinical situations require is appropriate but angry at times. flexibility. For example, if the clinician is working under extreme time constraints, such global state- This statement is somewhat confusing. In which ments may be appropriate; in most situations, how- sense is her affect appropriate? Is it appropriately ever, it is preferable to state specifically the main fearful for a person who believes that rats are entities that were ruled out, for this tells the reader invading her intestines or does the clinician mean that the clinician actually looked for these specific that her affect is appropriate for a person without a processes. Stated differently, with these global delusional system? The clinician should always phrases, the reader does not know whether they are first state what the patient' s affect is and then com- accurate or the end result of a sloppy examination. ment upon its appropriateness. Typical terms used If one has performed a careful examination, it to describe affect include normal (broad) affect seems best to let the reader know this. with full range of expression, restricted affect There is another problem with the phrasing used (some decrease in facial animation), blunted affect by Clinician A: he or she has stated that the patient (a fairly striking decrease in facial expression), a does not, in actuality, have hallucinations. It is pos- fiat affect (essentially no sign of spontaneous facial sible, however, that this patient is simply withhold- activity), buoyant affect, angry affect, suspicious CONTEMPORARY CLINICAL INTERVIEWING 3 51 affect, frightened affect, flirtatious affect, silly Arguably the single best introduction to the mental affect, threatening affect, labile affect, and edgy status is Robinson's and Chapman's Brian Cali- affect. The following description gives a much pers: A Guide to a Successful Mental Status Exam clearer feeling for this patient' s presentation: (1997). This very readable primer is written with both wit and a keen eye for practicality. On a more Clinician B: When asked about her mood, the patient comprehensive level, Trzepacz and Baker's book, abruptly retorted, "My mood is just fine, thank you !" The Psychiatric Mental Status (1993) is, in my Throughout much of the interview she presented a opinion, the single best reference book on the men- guarded and mildly hostile affect, frequently clip- ping off her answers tersely. When talking about the tal status currently available, filled with concise nurse in the waiting area she became particularly definitions and clinical applications. suspicious and seemed genuinely frightened. With- With regard to the patient in question, the fol- out tearfulness or a lability of affect. lowing description is a weak one and could use some polishing:

Sensorium, Cognitive Clinician A: The patient seemed alert. She was ori- Functioning, and Insight ented. Memory seemed fine and cognitive function- ing was grossly within normal limits. In this section the clinician attempts to convey a sense of the patient's basic level of functioning Once again this clinician's report is vague. Most with regard to the level of consciousness, intellec- importantly, the reader has no idea how much cog- tual functioning, insight, and motivation. It is nitive testing was performed. No mention has been always important to note whether a patient pre- made regarding the patient' s insight or motivation. sents with a normal level of consciousness, using The following excerpt provides a more clarifying phrases such as "The patient appeared alert with a picture: stable level of consciousness," or "The patient's consciousness fluctuated rapidly from somnolence Clinician B: The patient appeared alert with a stable to agitation." level of consciousness throughout the interview. Indeed, at times, she seemed hyperalert and overly It should be noted that this section of the mental aware of her environment. She was oriented to per- status examination may have evolved from two son, place, and time. She could repeat six digits for- processes: the informal cognitive examination and ward and four backward. She accurately recalled the formal cognitive examination. The informal three objects after five minutes. Other formal testing was not performed. Her insight was very poor as was cognitive examination is artfully performed her judgment. She does not want help at this time and throughout the interview in a noninvasive fashion. flatly refuses the use of any medication. The clinician essentially "eyeballs" the patient's concentration and memory by noting the method When done well, as described above, the mental by which he or she responds to questions. If the cli- status can provide a fellow clinician with a reliable nician chooses to perform a more formal cognitive image of the patient's actual presentation over the examination, it can range from a brief survey of course of the interview. It should be openly orientation, digit spans, and short term memory, to acknowledged that, in actual practice, the written a much more comprehensive examination, perhaps mental status may need to be significantly briefer, lasting 20 minutes or so. Clinical considerations but the principles outlined above remain important will determine which approach is most appropri- and can help prevent the briefer mental status from ate. For a fast reading and penetrating discussion being transformed into an inept mental status. of the use of the formal cognitive exam, the reader is referred to The Mental Status Examination in Neurology, (Strub & Black, 1979). The reader may THE DSM-IV also be interested in becoming familiar with the 30-point Folstein Mini-Mental State Exam. This The importance, with respect to interviewing, of exam can be given in about 10 minutes, provides a the DSM-IV system does not pertain to any spe- standardized set of scores for comparison, and is cific interviewing technique or mode of question- extremely popular (Folstein, Folstein, & McHugh, ing. The DSM-IV is not a style of interview; it is a 1975). There are two further outstanding resources diagnostic system. Its impact on interviewing on the mental status for the interested reader. derives from its having established an important 352 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT

set of symptoms that must be covered in order for conservative stance towards change, essentially a thorough assessment of the stated criteria to take making changes only when such changes would place. In this sense, the DSM-IV has become an simplify the system or bring it into agreement with important factor in determining the type and available empirical data. amount of data that contemporary clinicians must Perhaps the most important change was a philo- address. With the advent of the DSM-III, the sophical clarification regarding the etiology of DSM-III-R, and the DSM-IV, the degree of stan- some major mental disorders. The term, "Organic dardization required in a typical "intake interview" Disorders", which was used for diseases such as has increased significantly, for the required data- dementia and delirium in the DSM-III-R was base has grown significantly. removed. This term gave the misleading impres- In this section a brief outline of the DSM-IV sys- sion that other mental disorders such as schizo- tem is provided as an introduction to utilizing the phrenia and bipolar disorder were not caused by system in practice. An attempt is also made to high- organic dysfunction, despite the fact that there is light some of the more important conceptual substantial data suggesting that biochemical dys- advances of the DSM-III system as it was revised function plays an etiologic role in such disorders. and ultimately developed into the DSM-IV itself. A second set of changes was the attempt to sim- For the more interested reader, Frances, First, and plify diagnoses if possible. Perhaps the best exam- Ross (1995) have written an excellent brief review ple of this is the criteria list for Somatization of the changes in the DSM-IV from the DSM-III-R. Disorder, that decreased from an intimidating and When the DSM-III appeared in 1980, it repre- hard-to-remember list of 35 symptoms to a list that sented several major advances. First, as compared only must include four pain symptoms, two gas- to the Feighner criteria or the SADS, it was a sys- trointestinal symptoms, one sexual symptom, and tem designed primarily for clinical practice rather one pseudoneurological symptom. than for application to a research setting. This clin- A third set of changes was the attempt to clarify ical orientation mandated that all areas of psycho- certain variable characteristics of disorders that pathology be delineated. The actual diagnoses may have a direct impact upon treatment decisions were intended to be distinct from one another. or that might push a clinician to hunt more aggres- Consequently, a second major advance, in compar- ison to the DSM-I and the DSM-II systems, was an sively for a specific disorder. For example, in emphasis on well-defined criteria for almost all the Mood Disorders, Bipolar II Disorder (at least one diagnostic categories. major depressive disorder plus at least one The third major advance, and perhaps the most hypomanic episode without overt mania) was important, was the utilization of a multi-axial sys- added. Specifiers such as "with rapid cycling" and tem, in which the patient's presentation was not "with seasonal pattern" were added as qualifiers limited to a single diagnosis. The clinician was that could indicate treatment interventions and pushed to look at the patient' s primary psychiatric have prognostic significance. Another nice exam- diagnosis within the context of a variety of inter- ple was the splitting of Attention-Deficit/Hyperac- acting systems, such as the patient's physical tivity Disorder into three subtypes: Predominantly health, level of stress, and level of functioning. As Hyperactive-Impulsive Type, Predominantly Inat- Mezzich (1985) has pointed out, the DSM-III sys- tentive Type, and Combined Type. This categori- tem evolved from pioneering work with multi- zation will push clinicians to look for the axial systems across the world including England inattentive subtype that was probably underdiag- (Rutter, Shaffer, & Shepherd, 1975; Wing, 1970), nosed in the past. Germany (Helmchen, 1975; von Cranach, 1977), The fourth set of changes was the effort that was Japan (Kato, 1977), and Sweden (Ottosson & Per- put into expanding and enriching the narrative text ris, 1973). sections. This enrichment helps the clinician to There are very few major changes in the DSM- better understand the phenomenology of these dis- IV over its immediate predecessor the DSM-III-R. orders and to make a better differential diagnosis. This is because of the extensive work that went It will also serve as an excellent introduction to into the preparation of the DSM-II-R itself, with a these disorders for the more novice clinician. An heavy emphasis on empirical trials and data as effort has also been made to help the clinician opposed to expert opinion. This tradition was car- understand the cross-cultural variations of these ried on with the DSM-IV Task Force, which took a disorders. CONTEMPORARY CLINICAL INTERVIEWING 3 5 3

In the DSM-IV the clinical formulation is sum- 16. Other conditions that may be a focus of clinical marized on the following five axes: attention (includes V codes and entities such as psychological symptoms affecting a medical Axis I: All clinical disorders and other condi- condition or medication-induced movement tions that may be a focus of clinicial disorders such as tardive dyskinesia) attention (except for personality disor- ders and mental retardation) Clues to which general categories of disorders Axis II: Personality disorders and mental retar- are most relevant to the patient in question will dation arise as the clinician explores the patient's history Axis III: General medical conditions of the present illness, both spontaneously and with Axis IV: Psychosocial and the use of probe questions. The clinician should environmental problems keep in mind that there are childhood diagnoses, Axis V: Global assessment of functioning that may first come to clinical attention in adult- hood, such as attention-deficit disorder. Each of these axes is examined in more detail It should also be kept in mind that developmental below. disorders are coded on Axis I and may reflect lim- ited cognitive delays or pervasive developmental disorders involving serious cognitive, social, Axis ! motor, and language disturbances. Examples At first glance, Axis I may appear somewhat include mathematics disorder, developmental coor- intimidating because of the large number of diag- dination disorder, expressive language disorder, nostic entities it contains. But the clinician can autistic disorder, and Rett's disorder. approach the system in a two-step manner which The second step or secondary delineation con- greatly simplifies the task. In the first step or pri- sists of delineating the specific diagnoses under mary delineation, the clinician determines whether each broad category. In the secondary delineation the patient's symptoms suggest one or more of the the clinician clarifies the database so that an exact major diagnostic regions of Axis I which are con- DSM-IV diagnosis can be determined. Thus, if the fined to the following 16 relatively easily remem- clinician suspects a mood disorder, he or she will bered categories: eventually search for criteria substantiating spe- cific mood diagnoses, such as major depression, 1. Disorders usually first diagnosed in infancy, bipolar disorder, dysthymic disorder, cyclothymic childhood, or adolescence disorder, mood disorder due to a medical condi- 2. Delirium, dementia, amnestic and other cogni- tion, mood disorder due to substance abuse, bipo- tive disorders lar disorder not otherwise specified, depressive 3. Mental disorders due to a general medical con- disorder not otherwise specified, and mood disor- dition (e.g., personality change secondary to a der not otherwise specified. This secondary delin- frontal lobe tumor) eation is performed in each broad diagnostic area deemed pertinent. 4. Substance-related disorders With regard to the interview process itself, the 5. Schizophrenia and other psychotic disorders trained clinician performs these delineations in a 6. Mood disorders highly flexible manner, always patterning the 7. Anxiety disorders questioning in the fashion most compatible with 8. Somatoform disorders the needs of the patient. Utilizing a flexibly struc- 9. Factitious disorders tured format, the clinician can weave in and out of 10. Dissociative disorders these diagnostic regions, as well as any other areas 11. Sexual and gender-identity disorders such as the social history or family history, in 12. Eating disorders whatever fashion is most engaging for the patient. 13. Sleep disorders With the flexibly structured format the only limit- 14. Impulse-control disorders not otherwise clas- ing factor is that the standard database must be sified (e.g., kleptomania or pathological gam- thoroughly explored by the end of the available bling) time. It is up to the clinician to schedule the inter- 15. Adjustment disorders view creatively. When done well, the interview 354 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT feels unstructured to the patient, yet delineates an This axis also functions in many respects as an accurate diagnosis. important integration center in which diagnostic One diagnostic area that warrants further expla- concerns can be related to psychodynamic princi- nation is the concept of the V code. V codes repre- ples. For instance, the clinician is asked to look sent conditions not attributable to a mental carefully for evidence not only of personality dis- disorder that have nevertheless become a focus of orders but also of maladaptive personality traits. therapeutic intervention. Examples include aca- These traits can also be listed on Axis II. Along demic problems, occupational problems, uncom- similar lines, the clinician may list specific defense plicated bereavement, partner relational problems, mechanisms that may have been displayed during and others. The DSM-IV has specific V Codes for free-format areas of the interview or as methods of various types of abuse including physical and sex- avoiding certain topics raised by the clinician. ual abuse of both children and adults as well as These defense mechanisms may range from those neglect, emphasizing the importance of these areas commonly seen in neurotic disorders, such as for questioning in the initial interview. Sometimes rationalization and intellectualization, to those V codes are used because no mental disorder is seen in more severe disorders, such as denial, pro- present and the patient is coping with one of the jection, and splitting. stressors just listed. They can also be used if the clinician feels that insufficient information is available to rule out a psychiatric syndrome, but in Axis !11 the meantime, an area for specified intervention is On this axis the clinician considers the role of being highlighted. physical disorders and conditions, especially those that are potentially relevant to the understanding or management of the individual's mental disorders. Axis Ii The clinician is asked to view the patient's psychi- atric problems within the holistic context of the Axis II emphasizes the realization that all the impact of these problems on physical health, and Axis I disorders exist in the unique psychological vice versa. This axis reinforces the idea that a milieu known as personality. Many mental health sound medical review of systems and past medical problems are primarily related to the vicissitudes history should be a component of any complete ini- of personality development. Moreover, the under- tial assessment by a mental-health professional. lying personality of the patient can greatly affect In addition, other physical conditions that are the manner in which the clinician chooses to relate not diseases may provide important information to the patient both in the interview and in subse- concerning the holistic state of the patient. For quent therapy. instance, it is relevant to know whether the patient On Axis II the following diagnostic categories is pregnant or is a trained athlete, for these condi- are utilized: tions may point toward germane psychological issues and strengths. 1. Paranoid personality disorder 2. Schizoid personality disorder 3. Schizotypal personality disorder Axis IV 4. Antisocial personality disorder 5. Borderline personality disorder This axis concerns itself with an examination of 6. Histrionic personality disorder the current psychosocial and environmental prob- lems affecting the interviewee. It examines the cru- 7. Narcissistic personality disorder cial interaction between the client and the 8. Avoidant personality disorder environment in which he or she lives. Sometimes 9. Dependent personality disorder interviewers are swept away by diagnostic 10. Obsessive-compulsive personality disorder intrigues and fail to uncover the reality based prob- 11. Personality disorder not otherwise specified lems confronting the patient. This axis helps to (NOS) keep this important area in focus. By way of illustration, on this axis the inter- Mental retardation is also included on Axis II. viewer may discover that, secondary to a job lay- CONTEMPORARY CLINICAL INTERVIEWING 355 off, the home of the patient is about to be This brief review of the DSM-IV system shows foreclosed. Such information may suggest the need that the impact of this new diagnostic system on to help the patient make contact with a specific the interviewing process has been manyfold. The social agency or the utility of a referral to a case multi-axial approach and the thoroughness of the manager. When the clinical task focuses upon cri- diagnostic schema require the clinician to cover a sis-intervention techniques and solution-focused lot of ground, especially during a one-session strategies, as are commonly utilized in managed- intake, as is often necessary in a managed-care set- care settings, Axis IV becomes of primary impor- ting. But the resulting standardized database is an tance, for it points directly towards possible areas illuminating one that highlights a holistic and rig- for immediate intervention and support. orous approach to understanding the patient's problems, strengths, and needs. Moreover, the skilled use of a flexibly structured interview allows this informational base to be gathered in an Axis V empathic and flowing manner. A variety of changes were made in Axis V when the DSM-III was revised into the DSM-III-R. In DSM-III this axis delineated only the highest func- RESEARCH ON INTERVIEWING tioning of the patient over a two-month period in the preceding year. This relatively narrow perspec- It is not an exaggeration to state that it would tive did not provide an abundance of practical require an entire book to review comprehensively information. Consequently, in the DSM-III-R this the vast literature related to interviewing. On a more modest level, an attempt is made in this chap- axis was broadened. It included not only a rating of ter to introduce the reader to the main currents of the highest functioning in the past year, but also a this research area, providing a simplifying schema rating of the current functioning, which provided for categorizing the available literature while refer- immediate data pertinent to treatment planning and encing specific articles that can be used as stepping the decision as to whether hospitalization was war- stones into the categories described. ranted. These ratings were to be made by combin- One of the confusions facing the reader, as he or ing both symptoms and occupational and she attempts to approach the research on interview- interpersonal functioning on a 90-point scale, the ing, is the significant overlap between interviewing Global Assessment Functioning Scale (GAF research and research done with regard to psycho- Scale). therapy. This overlap is a healthy one, for it dem- In DSM-IV the same procedure and scale are onstrates that alliance issues are in some respects utilized (scale range is now 0-100), except the pro- inseparable from data-gathering issues. There is an cess has been streamlined to only require a GAF intimate relationship between the strength of the rating of "current functioning". Other time frames initial alliance and the resulting ability to gather can be added, in which case the additional time valid information and structure the flow of the con- frame is indicated in parentheses after the addi- versation effectively. tional score. Examples would be as follows, 45 On the other hand, there are differences in (highest level in past year) or 70 (at discharge from emphasis between intake interviews and psycho- hospital). therapy sessions. As the degree of standardization Probably of even more practical importance to has increased with the advent of new therapies and the clinician is the window that this axis opens into new diagnostic systems, these differences have the patient's adaptive skills and coping mecha- become more apparent. Eventually, such differen- nisms as reflected in the rating of current function- tiation between interviewing and psychotherapy ing. Looking for strengths to capitalize upon and to will probably be reflected more distinctly in the utilize as foundations for solution-focused problem research literature as increased research occurs on solving is equally important as finding out areas of structuring techniques and validity concerns. With malfunction and pathology. The gifted initial inter- these qualifications observed, the following major viewer is equally adept at uncovering what is right research areas will be discussed: (a) clinician and what is wrong. Both regions of knowledge are response modes, (b) nonverbal behavior and para- critical in order to provide the most rapid and long- language, (c) clinician characteristics as related to lasting relief for the client seeking help. alliance issues and empathic communication, (d) 356 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT reliability and validity concerns as related to struc- cian-response modes in actual clinical practice, tured interviews, and (e) educational techniques. attempting to delineate their possible impact on With regard to the first category, response-mode client engagement or behavior. Longborg demon- research attempts to examine the type of verbal strated an increase in the use of information giving, exchange occurring between clinician and patient, confrontation, and minimal respones (encouragers focusing on styles of response such as open-ended and silence) over the time course of the initial questions, reflections, and interpretations. Stiles interview of counseling trainees (Longborg, (1978) notes that it is important to separate Daniels, Hammond, Houghton-Wenger, & Brace, response-mode research from content research, 1991). Chang (1994) demonstrated a strong direct which focuses on the actual meaning of the words positive correlation between positive feedback in spoken, and research on extra-linguistic areas, such the initial session of a weight-reduction study and as speech characteristics, pauses, and laughter. specific behavioral indicators of client compliance It has been estimated that 20 to 30 response- including return for second session, number of mode systems have been developed (Elliott et al., weekly report sheets completed, and amount of 1987). Much of the pioneering research on time spent on meditation homework. This excel- response modes was done during the 1950s, 1960s, lent study is one of the few that focuses directly on and 1970s (Aronson, 1953; Danish & D'Augelli, positive client-outcome behaviors as opposed to 1976; Goodman & Dooley, 1976; Hackney & Nye, client engagement. 1973; Hill, 1975; Ivey, 1971; Robinson, 1950; Research with respect to nonverbal communica- Snyder, 1945, 1963; Spooner & Stone, 1977; tion, the second major research category, spans a Strupp, 1960; Whalen & Flowers, 1977). variety of perspectives that can best be separated In 1978 Clara Hill developed a system that inte- into three areas known as proxemics, kinesics, and grated many of the best features of the earlier sys- paralanguage. Edward T. Hall (1966) first coined tems. Her system consisted of 14 categories, the term proxemics in his classic book The Hidden including response types such as minimal-encour- Dimension. Proxemics represents the study of how ager, direct-guidance, closed-question, open-ques- humans conceptualize and utilize interpersonal tion, self-disclosure, confrontation, approval-reas- space. Hall was particularly interested in the surance, and restatement. Hill' s system was further impact of culture on an individual's sense of inter- developed to include three supercategories that personal space. Kinesics is the study of the body in focused on the degree of structuring as seen with movement including movements of the torso, low structure (encouragement/approval/reassur- head, limbs, face, and eyes as well as the impact of ance, reflection/restatement, and self-disclosure), posture. The field was pioneered by Ray T. Bird- moderate structure (confrontation, interpretation, whistell (1952) in the book, Introduction to Kine- and provision of information) and high structure sics: An Annotation System for Analysis of Body (direct guidance/advice and information seeking) Motion and Gesture. The final realm of nonverbal (Friedlander, 1982). study is paralanguage, which focuses on how mes- In 1987, six of the major-rating systems were sages are delivered, including elements such as compared when applied to therapy sessions by tone of voice, loudness of voice, pitch of voice, and well-known clinicians such as Albert Ellis and fluency of speech (Cormier & Cormier, 1979). Carl Rogers (Elliott et al., 1987). Interrater reliabil- The impact of these three areas of nonverbal ity was found to be high; when categories in differ- behavior on the issue of social control has received ent rating systems were collapsed to the same level much attention. Ekman has devoted considerable of specificity, moderate to strong convergence was time to the nonverbal constituents of the act of found. Studies such as the above point to a bright lying (Ekman, 1985; Ekman & Friesen, 1974; future for response-mode research when systems Ekman & Rosenberg, 1998). In a concise review of with high interrater-reliability are applied to vari- the literature concerning nonverbal behavior and ous interviewing situations. Different systems social control, including areas such as status, per- appear to shed slightly different light on the data- suasion, feedback, deception, and impression for- base, and the complementary use of various sys- mation, it appears that gaze and facial expression tems will probably become the preferred approach are the most telling factors (Edinger & Patterson, in the future. 1983). Scheflen (1972) has described kinesic In this regard the next logical step was to utilize reciprocals which represent display behaviors response-mode systems to study patterns of clini- between two organisms that convey intent, such as CONTEMPORARY CLINICAL INTERVIEWING 3 5 7 mating rituals, parenting behavior, and fighting have emphasized, the future of nonverbal research behavior, all of which also reflect the role of non- probably lies in an integrative approach combining verbal behavior in social control. paralanguage concerns, such as those delineated by Another area of active research concerns those Matarazzo and Wiens (1972), with other proxemic nonverbal behaviors that can facilitate the thera- and kinesic elements as they have impact on the peutic alliance. Tepper and Haase (1978) empha- interviewing relationship. sized the importance of considering a multichannel The interviewing relationship is further defined approach to understanding this subtle set of rela- by the third major area of research which focuses tionships. In one study they reviewed the impact of on characteristics of the interviewer that affect the various factors including verbal message, trunk therapeutic alliance, such as communication style, lean, eye contact, vocal intonation, and facial race, physical attractiveness, and the ability to con- expression on facilitative concerns such as empa- vey empathy. Because of its broad area of investi- thy, respect, and genuineness. Nonverbal compo- gation, this type of research overlaps with some of nents appeared to play a major role in these the areas already described. For instance, response facilitative processes. Attempts have been made to modes have been used to correlate client percep- determine methods of measuring clinician ability tions of clinician empathy with clinician phrasing, to decode the nonverbal behavior of patients. responses focused on exploration being strongly Rosenthal, Hall, DiMatteo, Rogers, and Archer associated with perceived empathy (Barkham & (1979) developed the Profile of Nonverbal Sensi- Shapiro, 1986). In a similar vein, the child psychi- tivity (PONS) in this regard. The original PONS atrist, Rutter, has developed a system of training consisted of 220 two-second film segments for clinicians to display four distinct styles ranging which subjects were asked to read accurately non- from a "sounding-board" style to a "structured" verbal clues, such as facial expression and tone of style. The impact of these styles on the interview voice. process was then examined (Rutter, Cox, Egert, The issue of decoding nonverbal cues immedi- Holbrook, & Everitt, 1981). ately raises the concept of cross-cultural differ- The concept of empathy has received as much, if ences with regard to interpretation of nonverbal not more, emphasis than any other single clinician behavior. As mentioned earlier, Hall was fasci- characteristic. As mentioned earlier, Rogers was nated by this process and, in more recent times, pivotal in the development of thought related to the Sue has studied these ramifications in detail (Sue, empathic process. Historically, another major con- 1981; Sue & Sue, 1977). tribution was made by Truax and Carkhuff (1967) Further issues concerning the complicated who emphasized qualities such as accurate empa- nature of how clinicians decode nonverbal lan- thy, nonpossessive warmth, and interpersonal gen- guage was more recently addressed by Hill and uineness as critical to the development of a sound Stephany (1990). They studied the presence of therapeutic alliance (Truax & Carkhuff, 1967). nonverbal behaviors, such as speech hesitancies, The Truax scale itself was a popular measure of vertical head movements, horizontal head move- empathy but has been attacked on numerous ments, arm movements, leg movements, postural grounds ranging from a lack of specificity concern- shifts, adaptors, illustrators, and smiles with recog- ing the clinician behaviors in question, to the claim nition by clinicians of moments of therapeutic that the scale may be measuring more than one importance to clients. thing (Cochrane, 1974; Lambert, DeJulio, & Stein, Issues such as paralanguage and temporal- 1978; Wenegrat, 1974; Zimmer & Anderson, speech characteristics have been carefully studied. 1968). Matarazzo and Wiens (1972) have developed a One of the more powerful unifying theories is concrete system of exploring such interactions. the empathy cycle proposed by G. T. Barrett-Len- They have delineated three major temporal-speech nard (1981). The empathy cycle delineates the characteristics: duration of utterance (DOU), empathic process in five specific phases, including response time latency (RTL), and percentage of such processes as the clinician' s ability to perceive interruptions (Wiens, 1983). In conjunction with the patient's feelings and the patient's ability to Harper, these same authors provide an insightful provide feedback that the empathic message has review of nonverbal behavior in Nonverbal Com- been received. The empathy cycle provides a munication: The State of the Art (Harper, Wiens, & framework from which differing components of Matarazzo, 1978). As Tepper and Haase (1978) the empathic process can be studied (Harmon, 358 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT

1986). Over the years, numerous articles and ongoing therapy have become known as significant reviews concerning empathy have spun off from events research. Cummings, Slemon, and Hallberg the works previously described as well as the view- (1993) have produced a good example of such work points espoused by the psychoanalytic community based partially upon their development of the (Berger, 1987; Elliott et al., 1982; Elliott et al., Important Events Questionnaire (Cummings, Mar- 1987; Gladstein, 1983; Smith-Hanen, 1977). tin, Hallberg, & Slemon, 1992). This questionnaire When considering the broad region of the has five questions, such as, "What was the most impact of clinician characteristics on alliance, one important thing that happened in this session for area of progress has been in the development of you?" and "Why was it important and how was it rating forms with regard to patient satisfaction helpful or not helpful?" and can be completed by with the interviewer. In 1975 Barak and LaCrosse both clinician and client. An attempt to uncover developed the Counselor Rating Form which is some underlying general principles in significant- also available in a shortened form (Barak & events research, using grounded theory-research LaCrosse, 1975; Corrigan & Schmidt, 1983). technique, was done by Frontman and Kunkel Other scales have followed that emphasize the alli- (1994). ance as it develops in the psychotherapeutic rela- A major thrust in research dealing with clinician tionship (Alexander & Luborsky, 1986; Marmar, characteristics evolves from the work of Strong. Horowitz, Weiss, & Marziali, 1986). More His work with the interpersonal-influence theory recently, Mahalik (1994) has developed a scale for of counseling has focused attention on the idea that actually measuring client resistance along five counselors who were perceived as expert, attrac- continua: Opposing Expression of Painful Affect, tive, and trustworthy would possess a more effec- Opposing Recollection of Material, Opposing tive means of influencing the behaviors of their Therapist, Opposing Change, and Opposing clients (Paradise, Conway, & Zweig, 1986; Strong, Insight (Client Resistance Scale [CRS]). In the 1968; Strong, Taylor, Bratton, & Loper, 1971). For same paper Mahalik used Hill's Response Modes example, the physical attractiveness of the clini- Verbal Category System (Hill, 1978) to study cor- cian appears to have a positive impact in certain relations between clinician response modes and situations (Cash, Begley, McCown, & Weise, specific forms of resistance. 1975; McKee & Smouse, 1983; Vargas & Degree of alliance has also been creatively Borkowski, 1982). approached by Stiles (1984), who developed the One of the major areas of recent research has idea of measuring the depth of an interview and the been the impact of race and cultural sensitivity to smoothness of the interview with the Session Eval- both client outcome and clinician perception uation Questionnaire (SEQ). Depth was measured (Atkinson, Matshushita, & Yashiko, 1991; Helms, on five bipolar scales: deep-shallow, full-empty, Carter, & Robert, 1991; Paurohit, Dowd, & Cot- powerful-weak, valuable-worthless, and special- tingham, 1982). Tomlinson-Clarke, using an archi- ordinary. The smoothness index is the mean rating val study, does a nice job of delineating some of on the following five bipolar scales: comfortable- the stumbling blocks in research design and inter- uncomfortabe, smooth-rough, easy-difficult, pretation of results, that is inherent in research pleasant-unpleasant, and relaxed-tense. Utlizing focusing upon the bias caused by race (Tomlinson- the SEQ, Tryon (1990) correlated a higher engage- Clarke & Cheatham, 1993). ment with deeper interviews and longer interviews Other characteristics that have been studied as rated by both the client and the counselor. This include religious background (Keating & Fretz, work was also based on her concept of the engage- 1990), body movement (LaCrosse, 1975), sponta- ment quotient (EQ), representing the percentage of neity and fluency of speech (Strong & Schmidt, clients who return to a counselor following the ini- 1970), and the role of displays of accreditation, tial assessment (Tryon, 1985). Using the SEQ as such as diplomas, on the walls of the clinician's well as four other engagement/outcome rating office (Siegel & Sell, 1978). scales, Mallinckrodt (1993) examined the impact In concluding a review of the literature, describ- of session satisfaction and alliance strength over ing the impact of clinician characteristics on alli- the course of time in a brief therapy format. ance, it is natural to mention some of the work Attempts to focus on the specific clinician/client based on the ultimate measure of clinician impact feelings, expectations, reactions, and behaviors as shown by impact on compliance and follow-up. during pivotal moments of the initial interview or A number of issues have been studied, such as the CONTEMPORARY CLINICAL INTERVIEWING 359 impact of the degree of directiveness, counselor Several important interviews have already been gender, and counselor experience, as well as the mentioned during the historical survey earlier in clinician's willingness to negotiate a therapeutic the chapter including the Diagnostic Interview contract. It appears that the ability to convey accu- Schedule (DIS) and the Schedule for Affective rately a sensitive understanding of the patient's Disorders and Schizophrenia (SADS). All of these problem and the ability to negotiate future treat- interview formats were developed with the idea of ment plans flexibly are powerful predictors of increasing the thoroughness, reliability, and valid- compliance (Eisenthal, Koopman, & Lazare, 1983; ity of the database. In some respects, these goals Eisenthal & Lazare, 1977a, 1977b; Epperson, have been at least partially realized. But Sanson- Bushway, & Warman, 1983; Heilbrun, 1974). Fisher and Martin (1981) have emphasized an Finally, two good reviews on process- and out- important point. Because these interviews have come-research techniques have been done by Hill been shown to be reliable, researchers tend to (Hill, 1990; Hill and Corbett, 1993). assume that the interviews will automatically be The fourth major area of interviewing research reliable in the hands of the clinicians working in leaves the arena of interpersonal dynamics and their protocols. This assumption is not necessarily focuses more on the issue of structured and semis- the case. It is important that reliability studies be tructured interviews and their impact on the thor- used at each research site and in an ongoing fash- oughness, reliability, and validity of the database. ion if, indeed, the interview format is to function Whereas much of the process research previously with a high degree of reliability. described has evolved from the fields of counsel- Before leaving the area of structured interviews ing and psychology, a large part of the work on and their impact on reliability and validity con- structured interviews has been undertaken in the cerns, it is important to mention the major role that field of psychiatry. child psychiatrists have had in the development of In many respects structured and semistructured interview formats. A variety of interviews have interviews grew out of the tradition of psychiatric been developed including the Diagnostic Interview epidemiology (Helzer, 1983). Examples include for Children and Adolescents (DICA) (Herjanic & the Home Interview Survey (HIS) used in the Mid- Campbell, 1977; Herjanic & Reich, 1982), the town Manhattan Study (Srole, Langer, Michael, Interview Schedule for Children (ISC) (Kovacs, Opler, & Rennie, 1962) and the Psychiatric Epide- 1983), the Kiddie-SADS (Puig-Antich & Cham- miological Research Interview (PERI) developed bers, 1978), the Diagnostic Interview Schedule for by Dohrenwend (Dohrenwend, Shrout, Egri, & Children (DISC) (Costello, Edelbrock, Kalas, & Mendelsohn, 1980). One of the most influential Dulcan, 1984), and the semistructured interview interviews that dealt directly with psychiatric developed by Rutter and Graham (1968). The symptomatology and diagnosis was the Present development of such interviewing tools has State Examination (PSE) developed by Wing in allowed researchers to address the intriguing ques- England (Wing, Cooper, & Sartorius, 1974). tions concerning the correlation between develop- The PSE combines elements of both the recent mental age and the validity of information psychiatric history and the mental status. It repre- provided by children (Edelbrock, Costello, Dul- sents a semistructured interview which emphasizes can, Kalas, & Conover, 1985). the need for the interviewer to cross-examine in a The fifth, and final major area in interviewing flexible manner when attempting to delineate the research concerns developments in educational presence and severity of a symptom. The PSE has techniques. This field is both exciting and broad, undergone numerous editions, and the ninth edi- with contributions from all disciplines of mental tion can be used in conjunction with a computer health. For the sake of simplicity, it is best to group program, CATEGO, which will delineate a diag- this research into two large areas: the development nosis from the data gathered during the interview. of improved supervision techniques and the devel- The ninth version contains 140 principal items and opment of tools for measuring student learning its phenomenological approach creates a Western with regard to interviewing skills. European feel in the interview format (Hedlund & In the same fashion that there has been a striking Vieweg, 1981). Numerous studies have been evolution in the number of treatment modalities undertaken with regard to the reliability of the PSE now available, there has been an equally remark- (Cooper, Copeland, Brown, Harris, & Gourlay, able advancement in training techniques over the 1977; Wing, Nixon, Mann, & Left, 1977). past several decades. For many years, interviewing 360 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT training seemed to be stuck on the model of indi- technique to the clinical interview in the mental- rect supervision that had evolved from the psycho- health professions and coined the term Interper- analytic tradition. With indirect supervision the sonal Process Recall. In IPR the students are asked trainee sees the patient alone and then reports on to reflect upon their internal feelings, thoughts, and "what happened" to the supervisor. Indirect super- reactions that are associated with specific clinical vision, when done well, can be very effective, pro- situations observed on videotapes of their own viding an intimate and carefully individualized clinical interviews. It is an excellent tool for supervision, but it has obvious limitations. uncovering countertransference issues and other The idea that the supervisor could actually "sit psychodynamic concerns. IPR is also a powerful in" with the patient and the interviewer probably method of helping trainees to recapture fleeting developed from a variety of disciplines. For impressions that would normally be lost or dis- instance, the idea of direct supervision is a popular torted (Elliott, 1986). style of supervision in family therapy. With regard Role playing provides yet another complemen- to interviewing an individual patient, numerous tary and widely accepted avenue for enhancing advantages appear when comparing direct to indi- specific interviewing skills (Canada, 1973; Errek rect supervision (Digiacomo, 1982; Stein, Karasu, & Randolpf, 1982; Hannay, 1980; Hutter et al., Charles, & Buckley, 1975). 1977). It may represent the single most effective Direct supervision removes many of the distort- manner by which to familiarize trainees with vari- ing mechanisms at work with the secondhand ous methods of handling hostile or awkward information provided by indirect supervision. In patient questions. direct supervision the supervisor can more accu- Ward and Stein (1975) pioneered the concept of rately evaluate nonverbal interaction, the structur- group supervision by colleagues. In this format the ing of the interview, and the handling of resistance. patient is interviewed by the trainee while fellow It also provides the trainee with the all-too-rare trainees observe in the same room. It provides a opportunity to model a more experienced clinician, format in which the group identifies emotionally if the supervisor chooses to demonstrate a tech- with both the patient and the interviewer, provid- nique. Rarely does direct supervision appear to ing a unique window into the processes of engage- hamper engagement with the patient significantly. ment and empathy In one study more than twice as many patients with Combining many of the advances just described, direct supervision, compared with indirect supervi- Ivey (1971) developed the innovative process of sion, remained in active treatment or successively microtraining. Microtraining probably represents completed therapy (Jaynes, Charles, Kass, & Holz- one of the most extensively studied and empiri- man, 1979). cally proven of all the training techniques currently On the heels of direct supervision, the closely utilized. In this format, specific skills, such as the related concept of videotape supervision was use of empathic statements or open-ended ques- developed. Such supervision complements both tions, are taught in an individualized fashion with a indirect and direct supervision. Like direct supervi- heavy emphasis on behavioral reinforcement. The sion it provides an excellent opportunity for feed- trainee is videotaped. Immediate feedback is given back on nonverbal and structuring techniques. It and problem areas delineated. Concise, goal- also offers the advantage of helping the clinician to directed reading material, related to a well-circum- develop a more effective observing ego by literally scribed skill, is provided. The trainee then immedi- experiencing the process of observing and analyz- ately performs further role-plays during which the ing his or her own behavior (Dowrick & Biggs, newly acquired skill is practiced until it is per- 1983; Jackson & Pinkerton, 1983; Maguire et al., fected, the trainee constantly being given concrete 1978; Waldron, 1973). feedback from the supervisor. The advent of recording technologies, such as Shea and colleagues would ultimately combine audiotape and videotape, provided the foundation all of the above techniques into an innovative train- for an innovative style of supervision known as ing program in interviewing (Shea & Mezzich, Interpersonal Process Recall (IPR). Bloom (1954) 1988; Shea, Mezzich, Bohon, & Zeiders, 1989). was one of the first to experiment with the tech- The first innovation in the training program was nique in his attempt to explore the thought pro- the idea that a unified block of highly supervised cesses of college students during discussion time, with an emphasis on direct mentorship and sections. Kagan (1975) was the first to apply the observation, should be set aside for trainees in CONTEMPORARY CLINICAL INTERVIEWING 361 which interviewing skills, as opposed to psycho- This body of literature is relatively large and is therapy skills, were intensively studied in an well reviewed by Ovadia, Yager, and Heinrich immediately relevant setting, such as an assess- (1982). A representative example of one such for- ment clinic or emergency room. The second inno- mat is the Queen's University Interview Rating vation was to focus, not only on traditional skills Scale (QUIRS). This rating process was developed such as empathy and engagement, but on utilizing to test the psychiatric-interviewing skills of medi- these skills in conjunction with real-life clinicial cal students as they rotated on third and fourth year demands such as DSM-III-R differential diagnosis clerkships (Jarrett, Waldron, Burra, & Handforth, and suicide assessment and also performed with 1972). The QUIRS consists of 23 items collapsed real-life time limitations. The third innovation con- from a list of 75 skills drawn from the literature. sisted of integrating both theory and supervision The test items are organized into three supercate- techniques (such as facilics, videotaping, direct gories: interview structure, interviewer role, and supervision, role playing, microtraining, mac- communication skills. With regard to medical rotraining, and behavioral self-monitoring) from a interviewing, Brockway (1978) developed an variety of disciplines into a specialized training extensive system for evaluating interviewing package that was designed into an individualized skills. This system includes over 50 items ranging program for each specific trainee. Individualized from process items, such as the use of silence, to learning goals were established as well as match- content items, such as eliciting the patient's ratio- ing the specific training techniques to the needs nale for making an appointment. and preferences of the trainee. Levinson and Roter (1993) demonstrated that The second major area, with regard to research physicians who participated in a two and one-half on interviewing in education, focuses less on the day continuing-medical-education program, when educational techniques themselves and more on compared to physicians who participated in a four methods of evaluating interviewing skills and and one-half hour workshop, showed significantly determining whether or not educational goals have more improvement in interviewing skills. In the been achieved. It is interesting to note that much of study five sequential patient visits were audiotaped the empirical work in this area has been done with one month before and one month after the train- medical-student education as opposed to psychiat- ings. The short-program group showed essentially ric-resident education or mental-health-profes- no improvements. The long-program cohort sional training. showed the use of more open-ended questions, One test technique consists of providing trainees more frequently asked for the patient's opinions, with videotape vignettes followed by three possi- and gave more biomedical information. Levinson ble physician responses. The trainee is asked to and Roter emphasize the value of a patient-cen- select the most appropriate response (Adler, Ware, tered style of interviewing and a corresponding & Enelow, 1968; Cline & Garrard, 1973). A writ- learner-centered style of teaching. ten test that attempts to examine interviewer deci- Kivlighan (1989) demonstrated improvements sion making with regard to the interview process in psychotherapy skill in trainees who completed a has been described by Smith (Smith, Hadac, & course on interpersonal-dynamic therapy, includ- Leversee, 1980). This instrument, called the Help- ing an increase in the use of minimal encouragers ing Relationship Inventory, consists of 10 brief and the reported depth of the sessions as reported patient statements. Each statement is followed by by the clients. The strength of this study was the five alternative responses categorized as under- use of both a good control group and a battery of standing, probing, interpretive, supportive, or eval- scales that have been documented to have good uative. Liston has developed a tool for assessing reliablity and validity, including the Intentions the acquisition of psychotherapy skills known as List, Client Reactions System, Hill Counselor Ver- the Psychotherapy Competence Assessment bal Response Category System (Hill, 1978), and Schedule (PCAS) (Liston & Yager, 1982; Liston, the Session Evaluation Questionnaire. It has not Yager, & Strauss, 1981). been common for researchers studying interview- With regard to assessing the skills demonstrated ing training programs, to utilize well-tested rating in the initial medical or psychiatric interview, the instruments, such as those used by Kivlighan. vast majority of work has moved away from pen- Along these lines, a paper written by Sanson- cil-and-paper tests, focusing instead on direct or Fisher, Fairbairn, and Maguire (1981) provides a videotaped evaluation of actual clinical interviews. good ending point for this section, albeit a some- 3 62 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT what sobering one. In a review of 46 papers deal- ning of the 19th century. In the past, part of the ing with the teaching of communication and prejudice blocking our understanding of human interviewing skills to medical students, the major- nature was created by a stubborn battle over turf ity of the papers revealed methodological flaws. among the various mental health traditions. Con- According to Sanson-Fisher, the future of research temporary interviewing represents an area in in this area should include a consolidated effort which the disciplines can at last join forces to fur- toward the use of standard research techniques ther our understanding of human nature, both as a including control groups, reliability studies, stu- function of psychopathology and as a function of dent characteristics, patient characteristics, and health. more sophisticated statistical analyses.

REFERENCES SUMMARY Adler, L., Ware, J. E., & Enelow, A. J. (1968). Evaluation In this chapter an attempt has been made to pro- of programmed instruction in medical interviewing. vide a sound introduction to the art and craft of ini- Los Angeles: University of Southern California Post- tial assessment interviewing and diagnosis, graduate Division. including its history, core clinical concerns, and Alexander, L. B., & Luborsky, L. (1986). The Penn Help- research. It can be seen that the historical currents ing Alliance Scales. In L. S. Greenberg & W.M. Pin- of initial assessment interviewing are varied and sof (Eds.), The psychotherapeuticprocess-A research rich. These currents include medical traditions handbook. New York: Guilford Press. such as the mental status, diagnostic systems, and American Psychiatric Association. (1980). Diagnostic psychoanalytic techniques. But they also include a and statistical manual of mental disorders (3rd ed.). remarkable array of contributions from nonmedi- Washington, DC: Author. cal fields such as counseling and psychology. American Psychiatric Association. (1987). Diagnostic At the present moment there is a cross-pollina- and statistical manual of mental disorders (3rd ed., tion among fields that is unusually promising. Rev. ed.). Washington, DC: Author. Research teams from different disciplines can be American Psychiatric Association. (1994). Diagnostic assembled to study the interviewing process from a and statistical manual of mental disorders (4th ed.). variety of perspectives. These interdisciplinary Washington, DC: Author. teams can analyze engagement techniques, non- Aronson, M. (1953). A study of the relationships between verbal processes, and structuring principles in the certain counselor and client characteristics in client- context of specific styles of interaction and charac- centered therapy. In W. U. Snyder (Ed.), Pennsylvania terological functioning, as determined by psycho- State College Psychotherapy Research Groups: Group logical testing and diagnosis by DSM-IV criteria. report of a program of research in psychotherapy. For the first time the role of response modes, empathic statements, and nonverbal techniques Atldnson, D., Matshushita, R., & Yashiko, J. (1991). Jap- can be studied in relation to specific psychopatho- anese-American acculturation, counseling style, logical states such as paranoia or to specific com- counselor ethnicity, and perceived counselor credibil- munication resistances as seen with overly ity. Journal of Counseling Psychology, 38, 473-478. loquacious patients. Barak, A., & LaCrosse, M. B. (1975). Multidimensional This chapter began with an historical perspec- perception of counselor behavior. Journal of Coun- tive, and it seems appropriate to end on an histori- seling Psychology, 22, 417-476. cal note as well. In 1806 the psychiatrist Philippe Barkham, M., & Shapiro, D. A. (1986). Counselor verbal Pinel became renowned for his humanistic treat- response modes and experienced empathy. Journal of ment of patients in the French institution known as Counseling Psychology, 33, 3-10. the Asylum de Bicetre. In his book A Treatise on Barrett-Lennard, G. T. (1981). The empathy cycle: Insanity he wrote, "Few subjects in medicine are Refinement of a nuclear concept. Journal of Counsel- so intimately connected with the history and phi- ing Psychology, 28, 91-100. losophy of the human mind as insanity. There are Benjamin, A. (1969). The helping interview. Boston: still fewer, where there are so many errors to rec- Houghton-Mifflin. tify, and so many prejudices to remove." (p. 3). His Berger, D. M. (1987). Clinical empathy. Northvale, NJ: point is as penetrating today as it was at the begin- Jason Aronson. CONTEMPORARY CLINICAL INTERVIEWING 363

Birdwhistell, R. L. (1952). Introduction to Kinesics: An Danish, S. J., & D'Augelli, A. R. (1976). Rationale and annotation system for analysis of body motion and implementation of a training program for paraprofes- gesture. Louisville: University of Kentucky. sionals. Professional Psychology, 7, 38-46. Bloom, B. S. (1954). The thought process of students in Deutsch, F., & Murphy, W. F. (1955a). The clinical inter- discussion. In S. J. French (Ed.), Accent on teaching: view, Vol. 1: Diagnosis. New York: International Experiments in general education. York: Harper. Universities Press. Brockway, B. S., (1978). Evaluating physician compe- Deutsch, F., & Murphy, W. F. (1955b). The clinical inter- tency: What difference does it make? Evaluation and view. Vol. 2: Therapy. New York: International Uni- Program Planning, 1, 211. versifies Press. Canada, R. M. (1973). Immediate reinforcement versus Digiacomo, J. N. (1982). Three-way interviews and psy- delayed reinforcement in teaching a basic interview chiatric training. Hospital and Community Psychia- technique. Journal of Counseling Psychology, 20, try, 33, 287-291. 395-398. Dohrenwend, B. P., Shrout, P. E., Egri, G., & Mendel- Carlat, D. J. (1999). The psychiatric interview: A practi- sohn, F. S. (1980). Nonspecific psychological distress cal guide. New York: Lippincott Williams & and other dimensions of psychopathology. Archives Wilkins. of General Psychiatry, 37, 1229-1236. Cash, T. F., Begley, P. J., McCown, D. Q., Weise, B. C. Donnelly, J., Rosenberg, M., & Fleeson, W. P. (1970). (1975). When counselors are heard but not seen: Ini- The evolution of the mental status~Past and future. tial impact of physical attractiveness. Journal of American Journal of Psychiatry, 126, 997-1002. Counseling Psychology, 22, 273-279. Dowrick, P. W., & Biggs, S. J. (Eds.). (1983). Using Chang, P. (1994). Effects of interviewer questions and video: Psychological and social applications. New York: Wiley. response type on compliance: An analogue study. Edelbrock, C., Costello, A. J., Dulcan, M. K., Kalas, R., Journal of Counseling Psychology, 41, 74-82. & Conover, N. C. (1985). Age differences in the reli- Cline, D. W., & Garrard, J. N. (1973). A medical inter- ability of the psychiatric interview of the child. Child viewing course: Objectives, techniques, and assess- Development, 56, 265-275. ment. American Journal of Psychiatry, 130, 574-578. Edinger, J. A., & Patterson, M. L. (1983). Nonverbal Cochrane, C. T. (1974). Development of a measure of involvement and social control. Psychological Bulle- empathic communication. Psychotherapy: Theory, tin, 93, 30-56. Research and Practice, 11, 41-47. Egan, G. (1975). The skilled helper: A model for system- Cooper, J. E., Copeland, J. R. M., Brown, G. W., Harris, atic helping and interpersonal relating. Belmont, T., & Gourlay, A. J. (1977). Further studies on inter- CA: Brooks/Cole. viewer training and inter-rater reliability of the Eisenthal, S., Koopman, C., & Lazare, A. (1983). Process Present State Examination (PSE). Psychological analysis of two dimensions of the negotiated 7, 517-523. Medicine, approach in relation to satisfaction in the initial inter- Cormier, W. H., & Cormier, L. S. (1979) Interviewing view. Journal of Nervous and Mental Disease, 171, strategies for helpers~A guide to assessment, treat- 49-54. ment, and evaluation. Montery, CA: Brooks/Cole. Eisenthal, S., & Lazare, A. (1977a). Evaluation of the ini- Corrigan, J. D., & Schmidt, L. D. (1983). Development tial interview in a walk-in clinic. Journal of Nervous and validation of revisions in the counselor rating and Mental Disease, 164, 30-35. form. Journal of Counseling Psychology, 30, 64-75. Eisenthal, S., & Lazare, A. (1977b). Expression of Costello, A. J., Edlebrock, C., Kalas, R., & Dulcan, M. K. patient' s requests in the initial interview. Psycholog- (1984). The NIMH Diagnostic Interview Schedule for ical Reports, 40, 131-138. Children (DISC): Development, reliability, and com- Ekman, P. (1985). Telling lies--Clues to deceit in the parisons between clinical and lay interviews. marketplace, politics, and marriage. New York: Cummings, A. L., Martin, J., Hallberg, E., & Slemon, A. W.W. Norton & Company. (1992). Memory for therapeutic events, session effec- Ekman, P., & Friesen, W. V. (1974). Detecting deception tivieness, and working alliance in short-term counsel- from the body or face. Journal of Personality and ing. Journal of Counseling psychoogy, 39, 306-312. Social Psychology, 29, 288-298. Cummings, A. L., Slemon, A. G., & Hallberg, E. T. Ekman, P. & Rosenberg, E. (1998). What the face (1993). Session evaluation and recall of important reveals: Basic and applied studies of spontaneous events as a function of counselor experience. Journal expression using thefacial action coding system. New of Counseling Psychology, 40, 156-165. York: Oxford University Press. 364 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT

Elliott, R. (1986). Interpersonal Process Recall (IPR) as Goodman, G., & Dooley, D. A. (1976). A framework for a psychotherapy process research method. In L. S. help-intended communication. Psychotherapy, The- Greenberg (Ed.), The psychotherapeutic process: A ory, Research, and Practice, 13, 106-117. research handbook. New York: Guilford Press. Grinder, J., & Bandler, R. (1975). The structure of magic Elliott, R., Filipovich, H., Harrigan, L., Gaynor, J., Reim- /. Palo Alto, CA: Science & Behavior Books. schuessel, C., & Zapadka, J. K. (1982). Measuring Hackney, H., & Nye, S. (1973). Counseling strategies response empathy: The development of a multicom- and outcomes. Englewood Cliffs, NJ: Prentice-Hall. ponent rating scale. Journal of Counseling Psychol- Hall, C. S., & Lindzey, G. (1978). Theories of personal- ogy, 29, 379-387. ity. New York: Wiley. Elliott, R., Stiles, W. B., Mahrer, A. R., Hill, C. E., Fried- Hall, E. T. (1966). The hidden dimension. New York: lander, M. L., & Margison, F. R. (1987). Primary ther- Doubleday. apist response modes: Comparison of six rating Hannay, D. R. (1980). Teaching interviewing with sim- systems. Journal of Consulting and Clinical Psychol- ulated patients. Medical Education, 12, 246-248. ogy, 55, 218-223. Harmon, J. I. (1986). Relations among components of the Endicott, J., & Spitzer, R. L. (1978). A diagnostic inter- empathic process. Journal of Counseling Psychology, view: The schedule for affective disorders and schizo- 33, 371-376. phrenia. Archives of General Psychiatry, 35, 837- Harper, R. G., Wiens, A. N., & Matarazzo, J. D. (1978). 844. Nonverbal communication: The state of the art. New Epperson, D. L., Bushway, D. J., & Warman, R. E. York: Wiley. (1983). Client self-terminations after one counseling Havens, L. (1978). Explorations in the uses of language session: Effects of problem recognition, counselor in psychotherapy: Simple empathic statements. Psy- gender, and counselor experience. Journal of Coun- chiatry, 41, 336-345. seling Psychology, 30, 307-315. Havens, L. (1979). Explorations in the uses of language Errek, H. K., & Randolpf, D. L. (1982). Effects of dis- in psychotherapy: Complex empathic statements. cussion and role-play activities in the acquisition of Psychiatry, 42, 40-48. consultant interview skills. Journal of Counseling Psychology, 29, 304-308. Havens, L. (1980). Experience in the uses of language in psychotherapy: Counterprojective statements. Con- Feighner, J. P., Robins, E., Guze, S. B., Woodruff, R. A., Winokur, G., & Munoz, R. (1972). Diagnostic criteria temporary Psychoanalysis, 16, 53-67. for use in psychiatric research. Archives of General Hedlund, J. L., & Vieweg, B. W. (1981). Structured psy- Psychiatry, 26, 57-63. chiatric interviews: A comparative review. Journal of Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). Operational Psychiatry, 12, 3 9-67. "Mini mental state:" A practical method for grading Heilbrun, A. B. (1974). Interviewer style, client satisfac- the cognitive state of patients for the clinician. Journal tion, and premature termination following the initial of Psychiatric Research, 12, 189-198. counseling contact. Journal of Counseling Psychol- Frances, A., First, M. B., & Ross, R. (1995). Getting up ogy, 21, 346-350. to speed on DSM-IV. Journal of Practical Psychiatry Helmchen, H. (1975). Schizophrenia: Diagnostic con- and Behavioral Health, 1, 2-9. cepts in the ICD-8. In M. H. Lader (Ed.), Studies in Friedlander, M. L. (1982). Counseling discourse as a schizophrenia. British Journal of Psychiatry, Special speech event: Revision and extension of the Hill Publication, 10, 1O- 18. Counselor Verbal Response Category System. Jour- Helms, J., Carter, E., & Robert, T. (1991). Relationships nal of Counseling Psychology, 29, 425-429. of white and black racial identity attitudes and demo- Frontman, K. C., & Kunkel, M. A. (1994). A grounded graphic similarity to counselor preference. Journal of theory of counselors' construal of success in the initial Counseling Psychology, 38, 446-45. session. Journal of Counseling Psychology, 41,492- Helzer, J. E. (1983). Standardized interviews in psychia- 499. try. Psychiatric Developments, 2, 161-178. Gill, M., Newman, R., & Redlich, F. C. (1954). The initial Herjanic, B., & Campbell, W. (1977). Differentiating interview in psychiatric practice. New York: Interna- psychiatrically disturbed children on the basis of a tional Universities Press. structured interview. Journal ofAb~ormal Child Psy- Gladstein, G. A. (1983). Understanding empathy: Inte- chology, 5, 127-134. grating counseling, developmental, and social psy- Herjanic, B., & Reich, W. (1982). Development of a chology perspectives. Journal of Counseling structured psychiatric interview for children: Agree- Psychology, 30, 467-482. ment between child and parent on individual symp- CONTEMPORARY CLINICAL INTERVIEWING 365

toms. Journal of Abnormal Child Psychology, 10, Kivlighan, Jr., D. M. (1989). Changes in counselor inten- 307-324. tions and response modes and in client reactions and Hersen, M., & Turner, S. M. (1985). Diagnostic inter- session evaluation after training. Journal of Counsel- viewing. New York: Plenum Press. ing Psychology, 36, 471-476. Hill, C. E. (1975). Sex of client and sex experience level Kovacs, M. (1983). The Interview Schedule for Children of counselor. Journal of Counseling Psychology, 22, (ISC): Interrater and parent-child agreement. 6-11. LaCrosse, M. B. (1975). Nonverbal behavior and per- Hill, C. E. (1978). Development of a counselor verbal ceived counselor attractiveness and persuasiveness. response category system. Journal of Counseling Journal of Counseling Psychology, 6, 563-566. Psychology, 25, 461-468. Lambert, M. J., DeJulio, S. J., & Stein, D. M. (1978). Hill, C. E. (1990). Exploratory in-session process Therapist interpersonal skills: Process, outcome, research in individual therapy: A review. Journal of methodological considerations, and recommenda- Consulting and Clinical Psychology, 58, 288-294. tions for future research. Psychological Bulletin, 85, Hill, C. E., & Corbett, M. M. (1993). A perspective on the 467-489. history of process and outcome research in counseling Langsley, D. G., & Hollender, M. H. (1982). The defini- psychology. Journal of Counseling Psychology, 40, tion of a psychiatrist. American Journal of Psychiatry, 3-24. 139, 81-85. Hill, C. E. & Stephany, A. (1990). Relation of nonverbal Langsley, D. C., & Yager, J. (1988). The definition of a behavior to client reactions. Journal of Counseling psychiatrist: Eight years later. American Journal of Psychology, 37, 22-26. Psychiatry, 145, 469-475. Levinson, W., & Roter, D. (1993). The effects of two con- Hutter, M. J., Dungy, C. I., Zakus, G. E., Moore, V. J., Ott, tinuing medical education programs on communica- J. E., & Favret, A. (1977). Interviewing skills: A com- tion skills of practicing primary care physicians. prehensive approach to teaching and evaluation. Journal of General Internal Medicine, 8, 318-324. Journal of Medical Education, 52, 328-333. Liston, E. H., & Yager, J. (1982). Assessment of clinical Ivey, A. E. (1971). Microcounseling: Innovations in skills in psychiatry. In J. Yager (Ed.), Teaching psy- interviewing training. Springfield, IL: Charles C Tho- chiatry and behavioral science. New York: Gmne & mas. Stratton. Jackson, M. G., & Pinkerton, R. R. (1983). Videotape Liston, E. H., Yager, J., & Strauss, G. D. (1981). Assess- teaching in family practice residencies. Journal of ment of psychotherapy skills: The problem of inter- Medical Education, 58, 434-435. rater agreement. American Journal of Psychiatry, Jarrett, F. J., Waldron, J. J., Burra, P., & Handforth, J. R. 138, 1069-1074. (1972). Measuring interviewing skill-The Queen's Lonborg, S. D., Daniels, J. A., Hammond, S. G., Hough- University Interviewer Rating Scale (QUIRS). Cana- ton-Wenger, B., & Brace, L. J. (1991). Counselor and dian Psychiatric Association Journal, 17, 183-188. client verbal response mode changes during initial Jaynes, S., Charles, E., Kass, F., & Holzman, S. (1979). counseling sessions. Journal of Counseling Psychol- Clinical supervision of the initial interview: Effects on ogy, 38, 394-400. patient care. American Journal of Psychiatry, 136, MacKinnon, R. A., & Michels, R. (1971). Thepsychiatric 1454-1457. interview in clinical practice. Philadelphia: W.B. Kagan, N. (1975). Interpersonal process recall: A Saunders. method of influencing human interaction. (Available Maguire, P., Roe, P., Goldberg, D., Jones, S., Hyde, C., from N. Kagan, 434 Erickson Hall, College of Edu- & O'Dowd, T. (1978). The value of feedback in cation, MSU, East Lansing, Michigan 48824.) teaching interviewing skills to medical students. Psy- Kaplan, H., Freedman, A., & Sadock, B. (Eds.). (1980). chological Medicine, 8, 695-704. Comprehensive textbook ofpsychiatry (3rd ed.). Bal- Mahalik, J. R. (1994). Development of the client resis- timore: Williams & Wilkins. tance scale. Journal of Counseling Psychology, 41, Kato, M. (1977). Multiaxial diagnosis in adult psychia- 58-68. try. Paper presented at the Sixth World Congress of Mallinckrodt, B. (1993). Session impact, working alli- Psychiatry, Honolulu, HI. ance, and treatment outcome in brief counseling. Keating, A., & Fretz, B. R. (1990). Christians' anticipa- Journal of Counseling Psychology, 40, 25-32. tions about counselors in response to counselor's Margulies, A. (1984). Toward empathy: The uses of descriptions. Journal of Counseling Psychology, 37, wonder. American Journal of Psychiatry, 141, 1025- 293-296. 1033. 366 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT

Margulies, A., & Havens, L. (1981). The initial encoun- attractiveness, and gender on perceptions of counse- ter: What to do first? American Journal of Psychiatry, lor attributes. Journal of Counseling Psychology, 33, 138, 421-428. 16-22. Marmar, C. R., Horowitz, M. J., Weiss, D. S., & Marziali, Pascal, G. R. (1983). The practical art of diagnostic inter- E. (1986). The development of the Therapeutic Alli- viewing. Homewood, IL: Dow Jones-Irwin. ance Rating System. In L. S. Greenberg & W. M. Pin- Paurohit, N., Dowd, E. T., & Cottingham, H. F. (1982). sof (Eds.), The psychotherapeutic processmA The role of verbal and nonverbal cues in the formation research handbook. New York: Guilford Press. of first impressions of black and white counselors. Matarazzo, J. D., & Wiens, A. N. (1972). The interview: Journal of Counseling Psychology, 4, 371-378. Research on its anatomy and structure. Chicago: Pinel, P. (1988). A treatise on insanity. Birmingham, AL: Aldine-Atherton. Gryphon Editions. (Original work published 1806) McKee, K., & Smouse, A. D. (1983). Clients' percep- Puig-Antich, J., & Chambers, W. (1978). The schedule tions of counselor expertness, attractiveness, and for affective disorders and schizophrenia for school- trustworthiness: Initial impact of counselor status and aged children. Unpublished manuscript, New York weight. Journal of Counseling Psychology, 30, 332- State Psychiatric Institute, New York, NY. 338. Reik, T. (1952). Listening with the third ear. New York: Meyer, A. (1951). The collected papers of Adolf Meyer Farrar, Strauss. (Vol. 3). E.E. Winters (Ed.). Baltimore: John Hopkins Richardson, S. A., Dohrenwend, B. S., & Klein, D. Press. (1965). Interviewing: Its forms and functions. New Mezzich, J. E. (1985). Multiaxial diagnostic systems in York: Basic Books. psychiatry. In H. I. Kaplan & B. J. Sadock (Eds.), The Robins, L. N., Helzer, J. E., Croughan, J., & Ratcliff, K. comprehensive textbook of psychiatry (4th ed.). Bal- S. (1981). National Institute of Mental Health Diag- timore: Williams & Wilkins. nostic Interview Schedule--Its history, characteris- Mezzich, J. E., Dow, J. T., Ganguli, R., Munetz, J. R., & tics, and validity. Archives of General Psychiatry, 10, Zettler-Segal, M. (1986). Computerized initial and 41-61. discharge evaluations. In J. E. Mezzich (Ed.), Clinical Robinson, D. J., & Chapman, B. (1997). Brain calipers: care and information systems in psychiatry. Wash- A guide to a successful mental status exam. Gratiot, ington, DC: American Psychiatric Press. MI: Rapid Psychler Press. Mezzich, J. E., Dow, J. T., Rich, C. L., Costello, A. J., & Himmelhoch, J. M. (1981). Developing an efficient Robinson, F. R. (1950). Principles and procedures in stu- clinical information system for a comprehensive psy- dent counseling. New York: Harper. chiatric institute. II. Initial evaluation form. Behavior Rogers, C. R. (1951). Client-centered therapy. Boston: Research Methods & Instrumentation, 13, 464-478. Houghton-Mifflin. Morrison, J. (1993). The first interview: A guide for cli- Rogers, C. R. (1959). A theory oftherapy, personality and nicians. New York: Guilford Press. interpersonal relationships as developed in the client- Othmer, E., & Othmer, S. C. (1989). The clinical inter- centered framework. In S. Koch (Ed.), Psychology: A view using DSM-III-R. Washington, DC: American study of a science: Formulations of the person and the Psychiatric Press, Inc. social context (Vol. 3). New York: McGraw-Hill. Othmer, E., & Othmer, S. C. (1994). The clinical inter- Rosenthal, R., Hall, J. A., DiMatteo, M. R., Rogers, P. L., view using DSM-IV, Vol. 1: Fundamentals. Washing- & Archer, D. (1979). Sensitivity to nonverbal commu- ton, DC: American Psychiatric Press, Inc. nication: The PONS test. Baltimore: Johns Hopkins Othmer, E., & Othmer, S. C. (1994). The clinical inter- University Press. view using the DSM-IV, Vol. 2: The difficult patient. Rutter, M., Cox, A., Egert, S., Holbrook, D., & Everitt, B. Washington, DC: American Psychiatric Press, Inc. (1981). Psychiatric interviewing techniques IV. Ottosson, J. O., & Perris, C. (1973). Multidimensional Experimental study: Four contrasting styles. British classification of mental disorders. Psychological Journal of Psychiatry, 138, 456-465. Medicine, 3, 238-243. Rutter, M., & Graham, P. (1968). The reliability and Ovadia, A. B., Yager, J., & Heinrich, R. L. (1982). validity of the psychiatric assessment of the child: Assessment of medical interview skills. In J. Yager Interview with the child. British Journal of Psychia- (Ed.), Teaching psychiatry and behavioral science. try, 11, 563-579. New York: Grune & Stratton. Rutter, M., Shaffer, D., & Shepherd, M. (1975). A mul- Paradise, L. V., Conway, B. S., & Zweig, J. (1986). tiaxial classification of child psychiatric disorders. Effects of expert and referent influence, physical Geneva, Switzerland: Word Health Organization. CONTEMPORARY CLINICAL INTERVIEWING 367

Sanson-Fisher, R., Fairbairn, S., & Maguire, P. (1981). Snyder, W. U. (1945). An investigation of the nature of Teaching skills in communication to medical stu- nondirective psychotherapy. Journal of General Psy- dent--A critical review of the methodology. Medical chology, 33, 193-223. Education, 15, 33-37. Snyder, W. U. (1963). Dependency in psychotherapy: A Sanson-Fisher, R. W., & Martin, C. J. (1981). Standard- casebook. New York: Macmillan. ized interviews in psychiatry: Issues of reliability. Sommers-Flanagan, R., & Sommers-Flanagan, J. (1999). British Journal of Psychiatry, 139, 138-143. Clinical interviewing (2nd ed.). New York: John Scheflen, A. E. (1972). Body, language and social order. Wiley & Sons, Inc. Englewood Cliffs, NJ: Prentice-Hall. Spitzer, R. L., Endicott, J., & Robins, E. (1978). Research Shea, S. C. (1988). Psychiatric interviewing: The art of diagnostic criteria. Archives of General Psychiatry, 35, 773-782. understanding. Philadelphia: W.B. Saunders. Spitzer, R. L., & Williams, J. B. W. (1983). DSM-III Shea, S. C. (1995). Psychiatric interviewing. In M. H. SCID Manual. New York: New York State Psychiat- Sachs, W. H. Sledge, & C. Warren (Eds.), Core read- ric Institute, Biometrics Research Department. ings in psychiatry. Washington, DC: American Psy- Spooner, S. E., & Stone, S. C. (1977). Maintenance of chiatric Press, Inc. specific counseling skills over time. Journal of Coun- Shea, S. C. (1998). The chronological assessment of sui- seling Psychology, 24, 66-7 1. cide events: A practical interviewing strategy for the Srole, L., Langer, T. S., Michael, S. T., Opler, M. K., & elicitation of suicidal ideation. Journal of Clinical Rennie, T. A. C. (1962). Mental health in the metrop- Psychiatry, 59, (Suppl.), 58-72. olis: The Midtown Manhattan Study (Vol. 1). New Shea, S. C. (1998). Psychiatric interviewing: The art of York: McGraw-Hill. understanding (2nd ed.). Philadelphia: W.B. Saun- Stein, S. P., Karasu, T. B., Charles, E. S., & Buckley, P. ders Company. J. (1975). Supervision of the initial interview. Shea, S. C. (1999). The practical art of suicide assess- Archives of General Psychiatry, 32, 265-268. ment. New York: John Wiley & Sons, Inc. Stiles, W. B. (1978). Verbal response modes and dimen- Shea, S. C., & Mezzich, J. E. (1988). Contemporary psy- sions of interpersonal roles: A method of discourse chiatric interviewing: New directions for training. analysis. Journal of Personality and Social Psychol- Psychiatry: Interpersonal and Biological Processes, ogy, 7, 693-703. 51,385-397. Stiles, W. B. (1984). Measurement of the impact of psy- Shea, S. C., Mezzich, J. E., Bohon, S., & Zeiders, A. chotherapy sessions. Journal of Consulting and Clin- (1989). A comprehensive and individualized psychi- ical Psychology, 48, 176-185. atric interviewing training program. Academic Psy- Strong, S. R. (1968). Counseling: An interpersonal influ- chiatry, 13, 61-72. ence process. Journal of Counseling Psychology, 15, Sheehan, D., & Lecrubier, Y., et al. (1999). Appendix 1, 215-224. The Mini-International Neuropsychiatric Interview Strong, S. R., & Schmidt, L. D. (1970). Expertness and (M.I.N.I.). Journal of Clinical Psychiatry, 60, (Suppl. influence in counseling. Journal of Counseling Psy- chology, 17, 81-87. 18) 39-62. Strong, S. R., Taylor, R. G., Bratton, J. C., & Loper, R. Siassi, I. (1984). Psychiatric interview and mental status (1971). Nonverbal behavior and perceived counselor examination. In G. Goldstein & M. Hersen (Eds.). characteristics. Journal of Counseling Psychology, Handbook of psychological assessment. New York: 18, 554-561. Pergamon Press. Strub, R. L., & Black, W. W. (1979). The mental status Siegel, J. C., & Sell, J. M. (1978). Effects of objective evi- examination in neurology. Philadelphia: F.A. Davis. dence of expertness and nonverbal behavior on client- Strupp, H. H. (1960). Psychotherapists in action: Explo- perceived expertness. Journal of Counseling Psy- rations of the therapist's contribution to the treatment chology, 25, 188-192. process. New York: Gmne & Stratton. Smith, C. K., Hadac, R. R., & Leversee, J. H. (1980). Sue, D. W. (1981). Counseling the culturally different. Evaluating the effects of a medical interviewing New York: Wiley. course taught at multiple locations. Journal of Medi- Sue, D. W., & Sue, D. (1977). Barriers to effective cross- cal Education, 55, 792-794. cultural counseling. Journal of Counseling Psychol- Smith-Hanen, S. S. (1977). Effects of nonverbal behav- ogy, 24, 420-429. iors on judged levels of counselor warmth and empa- Sullivan, H. S. (1970). The psychiatric interview. New thy. Journal of Counseling Psychology, 24, 87-91. York: W. W. Norton Company. 368 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT

Tepper, D. T., Jr., & Haase, R. F. (1978). Verbal and non- Ward, N. G., & Stein, L. (1975). Reducing emotional dis- verbal communication of facilitative conditions. tance: A new method of teaching interviewing skills. Journal of Counseling Psychology, 25, 35-44. Journal of Medical Education, 50, 605-614. Tomlinson-Clarke, S. & Cheatham, H. E. (1993). Coun- Wenegrat, A. (1974). A factor analytic study ofthe Truax selor and client ethnicity and counselor intake judg- Accurate Empathy Scale. Psychotherapy: Theory, ments. Journal of Counseling Psychology, 40, 267- Research and Practice, 11, 48-5 1. 270. Whalen, C. K., & Flowers, J. V. (1977). Effects of role Truax, C. B., & Carkhuff, R. R. (1967). Toward effec- and gender mix on verbal communication modes. tive counseling and psychotherapy: Training and Journal of Counseling Psychology, 24, 281-287. practice. Chicago: Aldine. Whitehom, J. C. (1944). Guide to interviewing and clin- Tryon, G. S. (1985). The engagement quotient: One ical personality study. Arch&es of Neurology and index of a basic counseling task. Journal of College Psychiatry, 52, 197-216. Student Personnel, 26, 351-354. Wiens, A. N. (1983). The assessment interview. In I. B. Tryon, G. S. (1990). Session depth and smoothness in Weiner (Ed.), Clinical methods in psychology. New relation to the concept of engagement in counsel- York: Wiley. ing. Journal of Counseling Psychology, 37, 248- Wing, J. K., Cooper, J. E., & Sartorius, N. (1974). Mea- 253. surement and classification of psychiatric symptoms. Trzepacz, P. T. & Bakev, R. W. (1993). The psychiat- Cambridge, MA: Cambridge University Press. ric mental status. New York: Oxford University Wing, J. K., Nixon, J. M., Mann, S. A., & Leff, J. P. Press. (1977). Reliability of the PSE (ninth edition) used in Vargas, A. M., & Borkowski, J. G. (1982). Physical a population study. Psychological Medicine, 7, 505- attractiveness and counseling skills. Journal of 516. Counseling Psychology, 29, 246-255. Wing, L. (1970). Observations on the psychiatric section von Cranach, M. (1977). Categorical vs. multiaxial of the International Classification of Diseases and the classification. Paper presented at the Seventh World British Glossary of Mental Disorders. Psychological Congress of Psychiatry, Honolulu, HI. Medicine, 1, 79-85. Waldron, J. (1973). Teaching communication skills in Zimmer, J. M., & Anderson, S. (1968). Dimensions of medical school. American Journal of Psychiatry, positive regard and empathy. Journal of Counseling 130, 579-591. Psychology, 15, 417-426. CHAPTER 14

STRUCTURED INTERVIEWS FOR CHILDREN AND ADOLESCENTS

Craig Edelbrock Amy Bohnert

INTRODUCTION from children, including those too young to com- plete paper-and-pencil questionnaires. Interviewing is a universal method of assessment As universal as interviewing is, it is perhaps one in all areas of mental-health research and clinical of the least rigorous and most fallible assessment practice. Face-to-face interviewing of people is a procedures. The flexibility of most interviews is a natural, and arguably indispensable, means of double-edged sword, allowing us to adapt assess- gaining information about emotional and behav- ments to individual respondents, but opening the ioral functioning, physical health, and social rela- door to numerous uncontrolled sources of variation tionships-both past and present. Part of the in the assessment process. Simply put, interview- appeal of interviewing is that it is a "low tech" ers differ widely in what they ask and how they ask assessment method that is adaptable to many dif- it. Given free reign, interviewers choose different ferent purposes. It is highly flexible and can be lines and styles of questioning. They cover differ- quickly adapted to a broad range of target phenom- ent material, in different ways. They project differ- ena, or alternatively to probe in-depth in a specific ent verbal and nonverbal cues to the respondent, area. Interviewing provides unparalleled ability to not to mention the fact that interviewers differ in insure that respondents understand questions, to how they rate, record, interpret, and combine inter- evoke rich and detailed examples, and to document viewees' responses. Such broad variations in con- chronicity of events. tent, style, level of detail, and coverage make Compared to other assessment methods, such as interviewing highly suspect from a measurement psychological testing and direct observation, inter- point of view. In the language of measurement, viewing can be efficient and cost-effective in terms interviewing is prone to high "information vari- of professional time and training. Interviewing is ance"--variability in what information is sought also usually readily accepted by both research sub- and elicited from respondents--which is blamed as jects and clinical clients and is typically expected a major cause of low reliability in the assessment to be the "default" assessment technique. Inter- and diagnostic process (see Matarazzo, 1983). viewing is, of course, ubiquitous as a means not A simple experiment effectively illustrates the only of obtaining assessment information, but of problem: Suppose there was a pool of subjects who "breaking the ice" and establishing rapport were absolutely identical in every way. Clinical between the interviewer and interviewee. It also interviews would not elicit identical information represents a potential way of obtaining information from such clones. Different interviewers would ask

369 3 70 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT

different questions in different ways, and would Many interview schedules have parallel formats rate and record the responses idiosyncratically. for interviewing adults (usually parents) about Moreover, interviews conducted by the same inter- children, and a separate format for direct interview viewer might yield quite different information due of children themselves. Viewing the child as a to variations in interviewing style and content from valuable source of information about themselves one clone to the next. If this hypothetical example was a revolutionary change in assessment theory were a study of diagnostic reliability, the subject and practicemand one that created numerous chal- variance would eliminated, since all subjects are lenges. identical. The criterion variance~variability due The assessment of child psychopathology has to use of different diagnostic criteria---could also traditionally depended upon reports and ratings by be eliminated if one diagnostic system were used. adults, particularly parents. This makes sense But information variance would remain as a major because parents are the most common instigators threat to reliability. Given the freedom of unstruc- of child mental-health referrals and they are almost tured interviews, differences in the information always involved in the assessment process. Par- obtained would undoubtedly arise and the reliabil- ents' perceptions are often crucial in the imple- ity of diagnoses would be less than perfect. mentation of child interventions and the evaluation How can the advantages of interviewing be of child outcomes. For many decades, direct inter- maintained while making it more scientifically rig- view of the child was not considered a useful orous as a measurement technique? The answer, at endeavor. Psychodynamic theories postulated that least in part, involves standardizing the interview children lack insight into their own problems. process. Standardizing in this sense means impos- Child developmentalists argued that young chil- ing some structure~literally limiting variability in dren are not cognitively mature enough to under- the question-answer interactions between inter- stand life history or symptom-oriented interviews. viewer and respondent. This is accomplished in These assumptions have been increasingly ques- three ways. The first is by defining the phenomena tioned, and numerous interview schedules have to be assessed. Differences between interviewers been developed for directly interviewing the can be reduced considerably by establishing what child~not always with successful results. The the interview does (and does not) cover. Second is challenges of interviewing children, however, do by limiting to some degree the order and wording create theoretical and practical problemsmmany of questions to be asked. Individual differences of which remain to be solved. between interviewers are thus further reduced by restricting how the target phenomena are covered. Third is by standardizing how responses are rated, Historical Foundations recorded, combined, and interpreted. Structuring the interview process in these ways addresses both The historical development of structured inter- the criterion variance and information variance views for children and adolescents owes much to inherent in any assessment process. There is less precedents set in the adult area--especially the criterion variance in structured versus unstructured epoch-making development of the Diagnostic interviews because the range and coverage of the Interview Schedule (DIS) and the Schedule for interview is set, and in the case of diagnostic inter- Affective Disorders and Schizophrenia (SADS). In views, the diagnostic system and specific diagnos- fact, two early interviews for childrenmthe Diag- tic criteria are specified. There is less information nostic Interview for Children and Adolescents variance, as well, because the order and wording of (DICA) and the Kiddie-SADS (K-SADS)minher- items is predetermined and there is a standard for- ited much of their format, style, and mode of mat for translating interviewee's responses into administration to their respective adult forerun- objective data. ners. But interviewing children has a history of its Numerous interview schedules have been devel- own and appears to trace out two distinct lines of oped, beginning with those designed for adults to influence: one diagnostic and the other descriptive. report about themselves. Researchers in the child The diagnostic line of development corresponds areas were quick to follow suit and develop inter- to the emergence of more differentiated taxono- view schedules for child and adolescent popula- mies of childhood disorders. Prior to 1980 and the tions. Many of these interviews were spin-offs or publication of the third edition of the Diagnostic downward extrapolations of adult interviews. and Statistical Manual (DSM-III) of the American STRUCTURED INTERVIEWS FOR CHILDREN AND ADOLESCENTS 371

Psychiatric Association (APA) (1980) there was for interviewing mothers about their children's little need for diagnostic interview schedules that behavior. This had the obvious advantage of yield- provided precise, detailed, and reliable assess- ing more objective data than an unstructured clini- ments of child psychopathology. During the era of cal interview and it insured that direct comparisons the first edition of the DSM (APA, 1956) there could be made between subjects assessed by differ- were only two diagnostic categories for children: ent interviewers. Moreover, the goal was to Adjustment Reaction and Childhood Schizophre- describe children's behavioral problems, rather nia. Adult diagnoses could be applied to children, than to detect prespecified syndromes and disor- but the vast majority of children seen in psychiatric ders. The unresolved questions about the existence clinics were either undiagnosed or were labeled and definition of specific childhood disorders were adjustment reactions (Rosen, Bahn, & Kramer, thus circumvented. 1964). More differentiated taxonomies of child- Interviews were conducted with 482 mothers hood disorders were provided by the Group for the and 193 children ages 6 to 12 years. The interview Advancement of Psychiatry (GAP) (1966) and the comprised 200 questions and took about 90 min- second edition of the DSM (APA, 1968), but both utes to complete. Most items had a yes/no response systems lacked explicit diagnostic criteria and format, but some involved rating the frequency or operational assessment procedures. Not surpris- intensity of the target behavior. ingly, the reliability of both systems was mediocre Several findings from this landmark study were (Freeman, 1971; Sandifer, Pettus, & Quade, 1964; replicated by later researchers. Reinterviews with Tarter, Templer, & Hardy, 1975). mothers, for example, indicated high test-retest In 1980, however, the DSM-III provided a dif- reliability for items such as thumb sucking, bed ferentiated taxonomy of "Disorders Usually First wetting, and stuttering. But reliability was low for Evident in Infancy, Childhood, or Adolescence" items such as fears and worries and for items that had more explicit diagnostic criteria. The need requiring precise estimates of frequency (e.g., for more reliable and valid ways of assessing diag- number of temper tantrums). Mother-child agree- nostic criteria was a primary stimulus for the ment was low for most behaviors, but was higher development of structured interview schedules for for behaviors such as bed-wetting, temper tan- children and adolescents. More impetus was trums, and biting fingernails. Mothers tended to gained from the successes in the adult area. report more behavior problems that are irksome to Although adult psychiatric disorders had explicit adults (e.g., bed-wetting, restlessness, overactiv- diagnostic criteria, refined through decades of ity), whereas children tended to report more prob- trial-and-error tinkering, reliability of adult diag- lems that are distressing to themselves (e.g., fears, noses was too low for research purposes, such as worries, nightmares). These findings have been epidemiologic surveys and clinical trials. This replicated many times over the years. prompted the development of structured interview In another pioneering effort, Rutter and Graham schedules, such as the Diagnostic Interview Sched- (1968) developed structured procedures for ule (Robins, Helzer, Croughan, & Ratcliff, 1981) directly interviewing the child. This was a major and the Schedule for Affective Disorders and departure from the prevailing thought and clinical Schizophrenia (Endicott & Spitzer, 1978), which practice of the time. In clinical settings, direct substantially reduced information variance and interview of the child was used primarily as a ther- boosted diagnostic reliability (see Matarazzo, apeutic rather than as an assessment technique. 1983). Researchers interested in child and adoles- Moreover, the assessment uses of the interview cent psychopathology were quick to follow suit were largely restricted to uncovering unconscious and, in fact, many interview schedules for children wishes, fears, conflicts, and fantasies (see Group are downward extrapolations of adult interviews. for the Advancement of Psychiatry, 1957). In con- Apart from diagnostic purposes, there had long trast, the interview procedures developed by Rutter been a need for obtaining descriptive data on chil- and Graham were aimed at descriptive assessment dren's emotional, behavioral, and social problems, of the child's emotional, behavioral, and social but standardized assessment procedures were lack- functioning and were based on direct questioning ing. In the spring of 1955 Lapouse and Monk of the child (and parent). (1958) undertook a survey to determine the preva- The parent and child versions of this interview lence and patterning of problem behaviors in a schedule differ somewhat, but parallel one another community sample. A standard format was used in content and rating procedure. Both are 3 72 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT semi-structured interviews designed for clinically symptoms and behaviors, but some are focused on trained interviewers. The exact order and wording specific syndromes and disorders, such as child- of questions is not prescribed. Instead, areas of hood depression. Lastly, there has been increasing functioning, such as school performance, activi- specialization in interviewer training and qualifi- ties, and friendships are listed, along with symp- cations. Some are designed for clinically sophisti- tom areas such as antisocial behavior, anxiety, and cated interviewers, whereas others are designed for depression. The parent version has more detail as lay interviewers having only interview-specific to duration, severity, action taken, presumed cause, training. and expected course of problems reported. The rat- ing of many items requires clinical judgment. Par- ent and child are interviewed separately. After Summary each interview, the interviewer rates the child's mental status and determines if the child has no Development of structured clinical interviews psychiatric impairment, some impairment, or defi- for children and adolescents can be traced along nite or marked impairment. two historical lines. First, emergence of differenti- Two findings from this early work have been ated taxonomies of childhood disorders with more replicated by later studies. First, higher reliabilities explicit diagnostic criteria necessitated more accu- were obtained for ratings of global psychiatric sta- rate, precise, comprehensive, and reliable diagnos- tus than for ratings of specific symptoms, syn- tic interviewing procedures. Diagnostic interview dromes, and disorders. Rutter and Graham (1968), schedules were therefore developed for purposes for example, found high interrater reliability of differential diagnosis of children already identi- (r=-.84) for the overall ratings of psychiatric fied as cases. Second, standard interview proce- impairment based on separate interviews of the dures for assessing children's emotional, child by different interviewers. But reliabilities behavioral, and social functioning were needed for were mediocre for items pertaining to attention and descriptive, developmental, and epidemiological hyperactivity (r=-.61), social relations (r=-.64), and studies. Interview schedules aimed at obtaining anxiety and depression (r=-.30). Second, as illus- descriptive information about children's function- trated by these results, reliabilities were generally ing were developed primarily for use with nonre- higher for problems, such as hyperactivity and ferred populations. antisocial behavior, than for problems such as The pioneering studies by Lapouse and Monk fears, anxiety, and depression. (1958) and Rutter and Graham (1968) broke new ground and introduced several innovations in inter- viewing, including (a) structuring the content of Recent Trends the interview around specific target phenomena, (b) providing prespecified formats for rating and Structured interview schedules for children and recording responses, (c) focusing on the child's adolescents have proliferated in the last 20 years as functioning rather than psychodynamic states, (d) the need for descriptive and diagnostic assessment directly interviewing the child, and (e) using paral- tools increased. There are now many well-devel- lel interview schedules for parent and child. oped interview schedules that are widely used in Research in the past 20 years has amplified and research and to a lesser extent in clinical practice. improved upon these methodological innovations. A major trend has been toward increasing special- A broad range of interview schedules for children ization of interview schedules. Specialization of and adolescents is now available and these sched- purpose, for example, has resulted in different ules are widely used in research. These interview interview schedules for screening nonreferred pop- schedules have become more specialized in pur- ulations versus differential diagnosis of identified pose, age range, coverage, and training require- cases. Specialization in age range has resulted in ments. Additionally, the child is now viewed as a different interview schedules for preschool-aged potentially important source of information, so children, grade schoolers, and adolescents. Inter- interview schedules have been developed specifi- view schedules have also become more specialized cally for interviewing children about their own in coverage and focus. Most cover a broad range of functioning. STRUCTURED INTERVIEWS FOR CHILDREN AND ADOLESCENTS 373

THEORETICAL UNDERPINNINGS tioned. It is less clear, however, that parents can provide reliable and valid information about covert Theory has played little part in the development behaviors that may be intentionally hidden from of descriptive interviews. The lack of a consensual adults, such as truancy, alcohol and drug abuse, theory of child psychopathology leaves researchers stealing, and vandalism; or about private phenom- with little guidance about what phenomena are ena such as fears, worries, and anxiety. Con- important to assess. Not surprisingly, interview versely, children and adolescents seem items are selected primarily on the basis of face unimpeachable as sources of information about validity--not theoretical importance. their own feelings and covert behaviorsmeven Any taxonomy is an implicit theory about what though a minimum level of cognitive maturity and disorders exist and how to diagnose them. As such, degree of insight may be required. Whether chil- diagnostic interviews are operationalizations of the dren can see certain behaviors such as disobedi- prevailing taxonomic theory~which in the United ence, inattentiveness, and stubbornness, as States has been the DSM. The DSM is evolving~ symptoms and report them during an interview or perhaps (as some critics would assert) just remains controversial. changing~rapidly in the child area. Recent revi- Lastly, the age and developmental level of the sions (DSM-III-R, DSM IV) barely resemble the child being interviewed has created thorny prob- epoch-making edition of 1980 (DSM-III). At one lemsmboth theoretical and practical. Theoreti- time, low diagnostic reliability could have been cally, issues involve how best to adapt interview blamed on inadequate assessment procedures, but procedures to abilities of the child--abilities that now such inadequacy appears at least equally due vary widely by age and developmental level. How to limitations of the underlying taxonomy itself. to make interviews more developmentally attuned No diagnostic procedure can be expected to yield has been a major source of debate and empirical more valid diagnoses than the taxonomy will trial-and-error. The practical offshoot is obvious: allow. There has been considerable taxonomic to what age range can various interview schedules progress in the child area, but the validity of many be used? This is often thought of as: "What is the diagnostic categories has been questioned, and it is lower age limit, or youngest-aged child, to which a not yet clear if the criteria and diagnostic thresh- given interview can be administered?" But the olds proposed in the ever-changing DSM are cor- opposite is of concern as well: "Is there an upper rect. It is not clear that such changes are really age limit for which an interview is deemed appro- taxonomic improvements as opposed to mere per- priate?" mutations and preferences of the DSM commit- Many developmentalists have expressed caution tees. about administering structured interviews to young In a broader historical view, each version of the children, on the grounds that they do not have the DSM must be seen as provisional, subject to revi- cognitive skills or language abilities to understand sion and refinement. It is ironic, however, that or respond correctly to complex and abstract ques- about the time research results can address the tions~especially those about psychological phe- validity of diagnostic categories, the diagnostic nomena. Indeed, questions designed to categories and criteria are revised. The changes operationalize DSM diagnostic criteria are neces- have not been subtle. Some child diagnoses have sarily complex, if only because the diagnostic cri- disappeared completely, many new diagnoses have teria themselves are complex. appeared, and many others have been radically Procedures for making interviews more amena- reformulated. Researchers have been in a seem- ble to young children have been advocated. Bier- ingly endless race of trying to "catch up" with such man (1984) was particularly articulate in stressing rapid revisions and interject empirical research the importance of reducing task complexity, by results into the process of revision. using familiar vocabulary, simple sentences, and The design and use of structured interviews is clear response options. Questions pertaining to not without assumptions of a theoretical nature. It time and frequency of past events have proven par- is a major assumption, of course, that informants ticularly vexing. Fallon and Schwab-Stone (1994) can provide valid information about children's found that reliability was lower for the Diagnostic emotional, behavioral, and social functioning. That Interview Schedule for Children (DISC) questions parents can report on their own children's overt requiring children to delineate time, compared to behavioral and social functioning is rarely ques- questions that either specified a time frame or did 3 74 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT not refer to time at all. Some interviews employ a DESCRIPTION visual time line as an aid to children's recall, and/ or try to anchor recall to significant events (e.g., A structured interview is a list of target behav- before the child' s last birthday, after school started, iors, symptoms, and events to be covered, guide- during last summer). But the value of such proce- lines for conducting the interview, and procedures dures is not well established. for recording the data. Interview schedules differ Empirical results can also address these issues. widely in degree of structure. A crude, but useful, With a highly structured diagnostic interview, distinction can be made between highly structured Edelbrock, Costello, Dulcan, Kalas, & Conover and semi-structured interviews. Highly structured (1985) found that reliability of child reports interviews specify the exact order and wording of increased rapidly over the age range from 6 to 18 questions and provide explicit rules for rating and years, and was low for children ages 6 to 9 years recording the subject's responses. The interviewer (average=.43), moderate for those ages 10 to 13 is given very little leeway in conducting the inter- years (average = .60), and moderately high for view and the role of clinical judgment in eliciting those ages 14 to 18 years (average = .71). For many and recording responses is minimized. In fact, the symptom areas, reliabilities for the younger group interviewer is seen as an interchangeable part of were unacceptably low, prompting the suggestion the assessment machinery. Different interviewers of ages 10 to 11 years as a practical lower limit for should ask exactly the same questions, in exactly interviews of this type. Interview data from parents the same order, and rate and record responses in proved quite reliable across the age range from 6 to exactly the same way. Semistructured interviews, 18 years in this study. Fallon and Schwab-Stone on the other hand, are less restrictive and permit (1994) also found that reliability of child reports the interviewer some flexibility in conducting the increase with age, whereas parents are more highly interview. The interviewer plays more of a role in reliable regardless of the child's age. determining what is asked, how questions are Schwab-Stone and colleagues have also found that phrased, and how responses are recorded. Differ- children were particularly unreliable in reporting ent interviewers should cover the same target phe- about time factors such as symptom duration and nomena when using a semistructured interview, onset (Schwab-Stone, Fallon, Briggs, & Crowther, but they may do so in different ways. 1994). These findings have supported the notion A high degree of structure does not necessarily that only parents should be interviewed for chil- yield consistently better data. Each type of inter- dren below age 10 or so; after ages 10 to 11 years- view has its advantages. Highly structured inter- both parents and children should be interviewed. views minimize the role of clinical judgment and As reasonable as this sounds, it is a disappointing typically yield more objective and reliable data. compromise, since a primary motivation for devel- But they are rigid and mechanical, which results in oping structured interviews was to provide a a stilted interview style that cannot be adapted to means of obtaining data from children them- the individual respondent. Alternatively, semis- selves--especially younger children. tructured interviews try to capitalize on expert clin- Results of a recent study are also quite discon- ical judgment and permit a more spontaneous certing. This study involved interviewing children interview style that can be adapted to the respon- ages nine, 10, and 11, then debriefing them after- dent. Of course, such flexibility allows more infor- wards to determine their level of understanding of mation variance to creep into the assessment the interview items. The findings were dramatic process, which compromises reliability to some and discouraging. The majority of children (more degree. The key unresolved issues are how highly than 60 percent) did not understand the interview structured clinical interviews should be and how items. Questions involving time and frequency much they should depend on clinical judgments by judgments were even more poorly understood interviewers. These are complex issues, of course, (20-30 percent correctly understood). Unfortu- and there may be no simple answer. The more nately, their poor comprehension of the questions appropriate questions may be: When would it be did not prevent these children from answering. best to minimize clinical judgment by highly struc- Almost without exception, they responded to ques- turing the interview, and when would it be best to tions: It is the meaning or potential value of such capitalize upon the expertise of clinically trained responses that must now be seriously questioned. interviewers by providing less structure? STRUCTURED INTERVIEWS FOR CHILDREN AND ADOLESCENTS 3 75

Structured clinical interviews for children and not at all to very extreme. Each section has a series adolescents differ in other ways besides degree of of model questions (e.g., Have you felt sad? Have structure. Most interview schedules have been you cried?) that serve as guidelines for the inter- developed for interviewing parents about their viewer. Interviewers are free, however, to ask as children, but parallel versions for directly ques- many questions as necessary to substantiate their tioning the child are becoming more common. symptoms ratings. Interview schedules also differ in length, organiza- The K-SADS also embodies a skip structure tion, time requirements, age appropriateness, whereby sections can be omitted if initial screen- amount and type of interviewer training, and diag- ing questions or "probes" are negative. If nostic coverage. depressed mood is not evident, for example, subse- quent questions in that section can be skipped. This reduces interviewing time substantially, but little Semi-Structured Interviews information is lost. Following the section on psychiatric symptoms, In the following section, we will briefly the interviewer rates 11 observational items (e.g., review several semi-structured interviews appearance, affect, attention, motor behavior) and including, the Kiddie-SADS, the Child Assess- rates the reliability and completeness of the entire ment Schedule, and the Child and Adolescent interview. Finally, the interviewer completes a glo- Psychiatric Assessment. bal-assessment scale reflecting overall degree of psychiatric impairment. The K-SADS yields information on presence The Kiddie-SADS and severity of about 50 symptom areas (depend- ing on the version of the interview). Most of the The Kiddie-SADS or K-SADS (Puig-Antich & core areas concern depressive disorder, but somati- Chambers, 1978) is a semi-structured diagnostic zation, anxiety, conduct disorder, and psychosis interview schedule for children ages 6 to 17 are also tapped. Additionally, there are 12 sum- years, modeled after the Schedule for Affective mary scales: four hierarchically related depression Disorders and Schizophrenia (SADS), an inter- scales, five depression-related scales (e.g., suicidal view schedule for adults developed by Endicott ideation), and scales reflecting somatization, emo- and Spitzer (1978). The K-SADS is designed to tional disorder, and conduct disorder. The assess current psychopathology. It is focused on K-SADS data can also be translated into Research affective disorders but also covers conduct disor- Diagnostic Criteria (RDC) and DSM diagnostic der, separation anxiety, phobias, attention defi- criteria for major depressive disorder, conduct dis- cits, and obsessions-compulsions. The K-SADS is order, and neurotic disorder. Diagnoses are based administered by clinically sophisticated inter- on the clinician's overview of the interview viewers having intensive training using the inter- responses, rather than computer algorithms applied view schedule and expert knowledge about the directly to the K-SADS data. DSM diagnostic criteria. An epidemiological version of the K-SADS The parent is usually interviewed first about the (K-SADS-E) is also available for assessing life- child. Then the child is interviewed and any dis- time psychopathology (Orvaschel, Puig-Antich, crepancies between parent and child reports are Chambers, Tabrizi, & Johnson, 1982). It parallels addressed. The interviewer may confront the child the K-SADS, but most questions are phrased as about discrepancies and attempt to resolve them "Have you ever done or had X?" As a preliminary before making final ratings. The interviews begin test of validity, 17 subjects having previous with an unstructured section aimed at establishing depressive episodes were reinterviewed six months rapport, obtaining a history of the present prob- to two years later. For all but one subject, the lems, and surveying current symptoms. Onset and K-SADS-E detected the same diagnosis that was duration of the current disorder and type of treat- made previously, suggesting accurate retrospective ment received are then recorded. The interviewer recall of previous psychiatric disturbances. then moves on to more structured sections cover- The K-SADS is widely used in clinical research ing specific symptoms. Each section includes an and there is a growing body of findings supporting item (e.g., depressed mood) to be rated by the its reliability and validity. Short-term test-retest interviewer on a seven-point scale ranging from reliability has been evaluated on 52 disturbed chil- 3 76 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT dren and their parents (Chambers et al., 1985). followed by feelings and behaviors, and ending Reliabilities averaged .55 (intraclass correlation, with items about delusions, hallucinations, and range: .09 to .89) for individual items, and aver- other psychotic symptoms. After interviewing the aged .68 (range: .41 to .81) for the 12 summary child, the interviewer rates 53 items (e.g., insight, scales. Internal consistency for the 12 summary grooming, motor behavior, activity level, speech). scales has averaged .66 (alpha statistic, range: .25 The CAS was intended to facilitate evaluation of to .86). For diagnoses, agreement over time ranged child functioning in various areas and to aid in the from .24 to .70 (kappa statistic). Parent-child formulation of diagnostic impressions. It is less agreement has averaged .53 (intraclass correlation, structured than other interview schedules, provid- range: .08 to .96) for individual items. ing a simple outline of target phenomena to be The K-SADS was developed primarily to iden- assessed, suggested questions, and a simple format tify children with major affective disorders. Since for recording the presence/absence of symptoms. it is designed to assess diagnostic criteria, the The CAS yields scores in 11 content areas (e.g., validity of the K-SADS depends upon the validity school, friends, activities, family) and nine symp- of the diagnostic system (currently the tom areas (e.g., attention deficits, conduct disorder, DSM-III-R). In a sense, the K-SADS has strong overanxious, oppositional). A total score reflecting content validity because it directly operationalizes total number of symptoms is also obtained. DSM criteria. On the other hand, the DSM is Clinical interpretation of the CAS is also flexi- evolving rapidly in the child area and the validity ble and requires considerable expertise. The inter- of many child psychiatric diagnoses is question- view was not originally designed to yield DSM able--so the validity of the K-SADS is necessarily diagnosis, although many items correspond to limited. Nevertheless, the K-SADS serves it DSM criteria. A diagnostic index has been devel- intended purpose well. It has proven to be very oped indicating the correspondence between CAS useful in selecting homogeneous subgroups of items and DSM criteria. To address DSM criteria depressed children from heterogeneous clinic pop- more fully, a separate addendum to the interview ulations (e.g., Puig-Antich, Blau, Marx, Greenhill, has been developed for assessing symptom onset & Chambers, 1978). Preliminary investigations and duration. This complicates the interview some- also suggest that the K-SADS is useful in research what, but provides more adequate coverage of aimed at elucidating the biological correlates of DSM criteria for diagnosis of attention deficit dis- childhood depression (Puig-Antich, Chambers, order, conduct disorder, anxiety disorders, opposi- Halpern, Hanlon, & Sachar, 1979) and some core tional disorder, enuresis, encopresis, and affective depression items are sensitive to treatment effects disorders. Diagnosis are based on clinical over- (Puig-Antich, Perel, Lupatkin, Chambers, Shea, view of CAS responses, rather than explicit algo- Tabrizi, & Stiller, 1979). rithms. Interrater reliability based on independent rat- ings of 53 videotaped child interviews was r=-.90 The Child Assessment Schedule for total symptom score, and averaged r=-.73 for content areas, and r=-.69 for symptom areas. Reli- The Child Assessment Schedule (CAS) is a abilities were somewhat higher for hyperactivity semi-structured interview for children and adoles- and aggression (average r=-.80) than for fears, wor- cents ages 7 to 12 years (Hodges, McKnew, ries, and anxiety (average r=-.60). Interrater reli- Cytryn, Stern, & Kline, 1982; Hodges, Kline, abilities averaged kappa= .57 for individual items. Stern, Cytryn, & McKnew, 1982). It was originally Test-retest reliability has been high for quantitative designed for directly interviewing the child only, scores in both diagnostic areas and content areas but a parallel version for interviewing parents has for an inpatient sample (Hodges, Cools, & McK- been developed. The CAS is designed for clini- new, 1989). Diagnostic reliability has been moder- cally trained interviewers and requires about 45 to ately high for that inpatient sample for many 60 minutes to administer to each informant (parent diagnoses (range: kappa=.56-.1.00), but lower for and child). It comprises 75 questions about school, ADHD (kappa=.43) and Overanxious Disorder friends, family, self-image, behavior, mood, and (kappa-.38). thought disorder. Most item responses are coded The validity of the CAS has been supported by Yes/No. The interview is organized thematically several findings. Total symptom score discrimi- beginning with questions about family and friends, nated significantly between inpatient, outpatient, STRUCTURED INTERVIEWS FOR CHILDREN AND ADOLESCENTS 377 and normal children and correlated significantly young as seven years of age and it has a very sim- (r=-.53, p<.001) with total behavior problem score ple format. Development of a parallel form for par- derived from the Child Behavior Checklist. Using ents, a diagnostic index, and an addendum referral for either inpatient or outpatient services as covering symptom onset and duration are useful the criterion for psychopathology, the CAS additions even though they complicate the inter- achieved a sensitivity of 78 percent and a specific- view and extend the interviewing time required. ity of 84 percent, based on discriminant analysis The CAS depends upon clinical inferences to a (Hodges, Kline, et al., 1982). Combining CAS large extent, but is relatively easy for interviewers scores and CBCL scores in one discriminant anal- to learn. ysis boosted sensitivity to 93 percent and specific- ity to 100 percent (no false positives). This suggests that combining parent and child data (or The Child and Adolescent alternatively, interview and rating-scale data) may Psychiatric Assessment yield better discriminative power. Scores on the CAS overanxious scale have correlated signifi- A relatively new structured interview is the cantly (r=-.54, p<.001) with scores on the Child and Adolescent Psychiatric Assessment State-Trait Anxiety Scale for Children. CAS (CAPA) (Angold, Prendergast, Cox, Harrington, depression scores have also correlated signifi- Simonoff, & Rutter, 1995). The CAPA is designed cantly (r=.53, p<.001) with scores on the Child to assess both DSM and International Classifica- Depression Inventory. tion of Diseases (ICD) criteria for a range of core Concordance between the CAS and the K-SADS diagnoses including affective and anxiety disor- has also been explored (Hodges, McKnew, Bur- ders, and disruptive behavior disorders. The CAPA bach, & Roebuck, 1987). Thirty clinically referred is moderately structured, with questions designed children ages 6 to 17 years and their parents were to be administered exactly as written, and a series interviewed separately using either the CAS and of follow-up questions designed to determine if the the K-SADS, then reinterviewed the next day with respondent meets clinical criteria for a specific the other interview schedule. Order of interviewing symptom. Test-retest reliability has been evaluated was counterbalanced so about half of the subjects with 77 psychiatric patients ages 10 to 18 years were interviewed first with the CAS, whereas the (Angold & Costello, 1995). Surprisingly, higher other half were interviewed first with the K-SADS. reliabilities were obtained for affective and anxiety Concordance between the two interview schedules disorders (kappa=.74-.90) than for disruptive was determined in four DSM-III diagnostic areas: behavior disorders (kappa=.55-.64). Further ADD, conduct disorders, anxiety disorders, and research following up on these promising findings affective disorders. Diagnoses were also made is warranted. based on (a) the child only, (b) the parent only, (c) parent or child, and (d) parent and child consensus. Concordance between the CAS and K-SADS was Highly Structured Interviews moderately high for interviews with parents (aver- age kappa =.62, range: .51 to .75), but lower for The following section reviews two highly struc- interviews with children (average kappa = .44, tured interviews: the Diagnostic Interview for range: .36 to .52). Concordance was lower for anx- Children and Adolescents and the Diagnostic iety disorders than other areas (kappa = .37 for the Interview Schedule for Children. child interviews and .51 for the parent interviews). Taking all diagnoses from parent or child inter- views reduced concordance slightly (average The Diagnostic Interview for kappa = .54). Requiring parent-child consensus on Children and Adolescents diagnoses reduced concordance even more (aver- age kappa = .46). Nevertheless, these results sug- The Diagnostic Interview for Children and Ado- gest moderately high concordance between the lescents (DICA) was one of the first structured CAS and the K-SADS, particularly for parent interviews for children and it has been widely used interviews. in clinical and epidemiological research. The orig- Overall, the CAS is a useful descriptive tool and inal version, developed in 1969, was patterned diagnostic aid. It can be used with children as after the Renard Diagnostic Interview and keyed to 378 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT the ICD and Feighner diagnostic criteria (see Wel- separately from the parent interview and the child ner, Reich, Herjanic, Jung, & Amado, 1987 for a interview. review). The DICA was revised in 1981 along the To test interrater reliability, 10 interviewers lines of the NIMH Diagnostic Interview Schedule independently coded two videotaped interviews (Robins, Helzer, Croughan, & Ratcliff, 1981) and with children. Agreement on symptom items aver- was keyed to then new DSM-III criteria. Research aged 85 percent (Herjanic & Reich, 1982). using the earlier version (e.g., Herjanic, Herjanic, Test-retest reliability has been determined by hav- Brown, & Wheatt, 1975; Herjanic & Campbell, ing five psychiatrists code the same videotaped 1977) was pioneering in many ways, but is proba- interview twice over a two-to-three-month inter- bly obsolete, at least with respect to the reliability val. Agreement over time averaged 89 percent and validity of the later DSM-III and DSM-III-R (range: 80 percent- 95 percent) for individual versions. symptom items. In another study, 27 children The revised DICA is highly structured and pro- admitted to an inpatient psychiatric unit were inter- vides the interviewer with specific wording of viewed twice by two different interviewers, one to questions and explicit categories for response cod- seven days apart (Welner, Reich, Herjanic, Jung, & ing. Most symptom items are coded 1(No), 2(Yes), Amado, 1987). Inter-interviewer agreement on the or 3(Uncertain). Responses coded "Uncertain" can presence/absence of specific diagnoses was quite be clarified by subsequent sub-questions and high (kappas ranged from .76 to 1.00). recoded either "Yes" or "No." The role of clinical Mother-father agreement was tested for a sample inference in conducting the interview and making of 74 children (Sylvester, Hyde, & Reichler, symptoms ratings has been minimized, so the 1987). Agreement regarding the presence of any diagnosis was moderately high (kappa = .54). DICA can be administered by clinicians or lay Agreement was higher for Oppositional/Conduct interviewers. A moderate amount of instru- Disorder and Attention Deficits (range: .54-.61) ment-specific training is required, however. than for Anxiety Disorders and Depression (range: Parallel interview schedules have been devel- .33-.39). Parent-child agreement has also been oped for interviewing the child (DICA-C) and par- determined using a sample of 84 children referred ent (DICA-P) about the child. The parent version for outpatient services and their parents (Welner et covers demographic-background information, al., 1987). For five diagnostic groupings (ADD, pregnancy and childbirth, and medical and devel- conduct disorders, affective disorders, enuresis, opmental history. A long section covers specific oppositional disorder), parent-child agreement on symptoms organized by diagnostic area (e.g., the presence/absence of the diagnosis averaged .62 Attention Deficit Disorder, Conduct Disorder, (kappa statistic, range: .49 - .80). This represents Separation Anxiety Disorder). For each diagnosis, much higher parent-child agreement than has been one or more questions have been written to cover found in previous studies (e.g., Reich, Herjanic, each diagnostic criterion. The interview also Welner, & Gandhy, 1982). includes questions about possible disorders in sib- Validity of the original DICA was supported by lings and a brief family medical and psychiatric its ability to discriminate significantly between history. The child interview parallels the symp- matched samples of pediatric and psychiatric refer- toms portion of the parent interview. Although the rals (Herjanic & Campbell, 1977). Validity of the symptoms sections of the interviews are quite long, DICA-C was tested for 27 inpatients by comparing a skip structure is employed to reduce interviewing DICA diagnoses with independent discharge diag- time if few symptoms are present. noses formulated by clinicians (Welner et al., The DICA yields information on the presence/ 1987). Agreement was moderate for three diagnos- absence of more than 150 specific symptoms, as tic groupings: attention deficit disorders well as their severity, onset, duration, and associ- (kappa=.50), conduct disorders (.43), and affective ated impairments (see Herjanic & Reich, 1982). disorders (.52), but was low for anxiety/phobic dis- Diagnoses are made by directly comparing item orders (.03) and adjustment disorders (-.18). responses to DSM criteria for symptoms, severity, Two other recent studies have addressed validity onset, and duration. All DSM diagnoses applied of the DICA. In one study, agreement between the to children and adolescents are covered. Unlike DICA and best-estimate clinical diagnoses were the K-SADS where parent and child responses are determined for a sample of 30 children receiving first reconciled, DICA diagnoses are formulated inpatient services (Carlson, Kashani, Thomas, STRUCTURED INTERVIEWS FOR CHILDREN AND ADOLESCENTS 379

Vaidya, & Daniel, 1987). For six diagnostic areas children (DISC-P). The child interview takes about (ADD, conduct disorder, oppositional disorder, 40 to 60 minutes to complete with clinically affective disorder, overanxious disorder, and sepa- referred children, whereas the parent version takes ration anxiety), agreement with the best-estimate about 60 to 70 minutes. A time frame of the last diagnoses was low-moderate for the DICA-C year is used for most items and specific informa- (average kappa=.38, range: .15 -.75) and the tion about onset and duration is sought for many DICA-P (average kappa=.40, range: .05 -.66). In symptom items. the other study, the DICA was compared with The DISC covers a broad range of symptoms as scores on the Personality Inventory for Children well as their severity and chronicity. Most items (PIC), a measure of child personality completed by are coded 0-1-2 where 0 corresponds to no or parents (Sylvester, Hyde, & Reichler, 1987). never, 1 corresponds to somewhat, sometimes, or a Scores greater than T=65 on certain PIC scales little, and 2 corresponds to yes, often, or a lot. (e.g., hyperactivity) were used to categorize chil- Descriptions and examples offered by the respon- dren, then these categorizations were compared to dent are recorded verbatim for later editing. The their corresponding diagnosis (e.g., attention defi- DISC was originally keyed to the DSM-III and cit disorder) derived from the DICA-C. This is a covers most psychiatric diagnoses applicable to very stringent test of convergence, since it involves children and adolescents. Some diagnoses (e.g., comparing a parent-completed personality inven- pica, autism) are derived from the parent interview tory with the child-completed diagnostic inter- alone. Diagnoses are generated by computer algo- view. There were significant relationships between rithms applied to edited DISC data. Diagnoses are the two instruments in many areas, although the derived separately from the DISC-C and DISC-P. degree of convergence was fairly low (average Both interviews also yield quantitative symptoms kappa=.28, range: .11 to .48). scores in symptoms areas (e.g., overanxious, con- In sum, the DICA has broad diagnostic coverage duct disorder, attention deficits). and its moderate training requirements make it Interrater reliability has been tested by having suitable for large-scale epidemiological surveys three lay interviewers independently code video- involving many nonprofessional interviewers. taped interviews of 10 children (Costello, Edel- Recent studies have generally supported the reli- brock, Dulcan, Kalas, & Klarie, 1984). ability and validity of the DICA. Reliabilities averaged .98 for symptom scores (range: .94 to 1.00), indicating very little rater dis- agreement in how responses are coded. Test-retest The Diagnostic Interview Schedule for Children reliability has been determined on 242 clinically referred children and their parents (Edelbrock et NIMH has sponsored the decade-long develop- al., 1985). Parents and children were interviewed ment of the Diagnostic Interview Schedule for twice at a median interval of nine days. For parent Children (DISC) for use in epidemiological studies interviews, test-retest reliability was .90 (intraclass of child and adolescent psychopathology (see Cos- correlation) for total symptom score and averaged tello, Edelbrock, Kalas, Kessler, & Klarie, 1982). .76 for symptom scales (range: .44 to .86). For The DISC is similar in design and purpose to the child interviews, reliability was strongly related to Diagnostic Interview Schedule (DIS) used in epi- age. For total symptom scores, reliabilities were demiological research on adult disorders (Robins .39, .55, and .81 for children ages 6-9, 10-13 and et al., 1981) and its offspring the DICA (see 14-18, respectively. For symptom scores, reliabili- description above). The DISC is a highly struc- ties also increased with age and averaged .43, .60, tured diagnostic interview in which the order, and .71 for children ages 6-9, 10-13, and 14-18 wording, and coding of all items is specified. Like years, respectively. For 21 DSM-III diagnoses its predecessors, the DISC employs a skip structure having sufficient prevalence, test-retest reliability to reduce interviewing time with children having for the parent interview averaged kappa = .56 few symptoms. Since it was designed for (range: .35 to .81). Reliabilities of diagnoses large-scale epidemiologic studies, the DISC can be derived from the child interviews averaged .36 administered by lay interviewers having two to (range:. 12 to .71). three days of instrument-specific training. Parallel Parent-child agreement on child symptom scores versions have been developed for separately inter- has also been examined for 299 parent-child dyads viewing children (DISC-C) and parents about their (Edelbrock, Costello, Dulcan, Conover, & Kalas, 380 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT

1986). Only a moderate degree of agreement was disorder, and depression/dysthymia and scales found overall (average r = .27), but agreement was labeled hyperactive, delinquent, and depressed, higher for behavior/conduct symptoms than affec- respectively. These relations were generally linear. tive/neurotic symptoms and was higher among An increasing score on the scale corresponded to older than younger children. Regardless of the an incrementally higher probability of obtaining child's age, parents reported significantly more the diagnoses. No "diagnostic threshold" was behavior/conduct problems than their children apparent. However, children scoring above the reported about themselves. Children reported sig- normative range (T > 70) on the scales were much nificantly more affective/neurotic problems and more likely to receive the diagnosis than children drug and alcohol abuse than their parents reported scoring within the normative range. This suggests about their children. Similar results were obtained substantial convergence between two different using a Spanish version of the DISC with a com- ways of assessing child psychopathology. munity sample in Puerto Rico (Rubio-Stipec, In addition, relations between the DISC and the Canino, Shrout, Dulcan, Freeman, & Bravo, 1994); K-SADS has been determined in a community and with the CAS interview with an inpatient sam- sample of children (Cohen, O'Conner, Lewis, ple (Hodges, Gordon, & Lennon, 1990). So despite Velez, & Malachowski, 1987). One hundred chil- generally low levels of agreement, pattern of dis- dren ages 9 to 12 years who had been interviewed agreement seem consistent. with the DISC were reinterviewed with the Validity of the DISC interviews has been sup- K-SADS three to four months later. Significant, ported by several lines of evidence. Costello, Edel- although moderate, levels of agreement were brock, and Costello (1985) compared matched obtained for many diagnoses. DISC diagnoses, samples of pediatric and psychiatric referrals and however, have been shown to agree very poorly found that symptoms scores computed from both (kappa=.03-.17) with independent diagnoses by the DISC-P and DISC-C discriminated signifi- clinicians (Weinstein, Stone, Noam, Grimes, & cantly between these criterion groups. Total symp- Schwab-Stone, 1989). This is probably more an tom score derived from the DISC-P provided the indictment of the reliability of the clinicians than best discrimination (p<.001). In a multiple dis- the validity of the DISC, but sources of disagree- criminant analysis, symptoms scores derived from ment between the two sources are worth further both parent and child interviews contributed sig- investigation. nificantly to the equation which correctly classi- In the largest study to date, NIMH sponsored a fied 77 of the 80 children. Based on the DISC-P, multisite evaluation of the test-retest reliability of the psychiatric referrals obtained 51 diagnoses of the DISC v2.1 on 97 clinically referred subjects severe disorders, compared to only two diagnoses and 278 non-referred community subjects (Jensen, in the pediatric group. Roper, Fisher, & Piacentini, 1995). This study pro- Symptoms scores derived from the DISC-C and duced many important findings. First of all, there DISC-P have also been shown to correlate signifi- was wide variability in reliability across the three cantly with other measures of child psychopathol- participating sites. For ADHD, for example, ogy, such as the parent and teacher versions of the test-retest reliability ranged from a high of Child Behavior Checklist (CBCL). Total symptom kappa=.72 to a low of kappa=.38 for clinic cases. score derived from the DISC-P, for example, has For conduct disorder, reliabilities ranged from a correlated r=-.70 with total behavior-problem score high of kappa=.90 to a low of kappa=-.11! Sec- from the CBCL (Costello et al., 1984). The ond, for most diagnoses, reliabilities were higher DISC-C has shown weaker, but significant rela- for parent interviews than child interviews, and tions to the CBCL (r=-.30). Costello, Edelbrock, this was true for both clinic and community sub- and Costello (1985, p. 591) also have found signif- jects. Third, for all diagnostic categories, reliabili- icant convergence between severe diagnoses from ties were lower for the community sample, the DISC-P and CBCL scores above the normative compared to the clinic sample (Jensen, Roper, range. Fisher, & Piacentini, 1995, p. 66). For example, for Edelbrock and Costello (1988) also have the category "Depression and/or Dysthymia," explored the relationship between DISC diagnoses kappa=.70 for the clinic sample, but only and specific scales of the Child Behavior Profile. kappa=.26 for the community sample. This is con- They found considerable convergence between sistent with generally higher reliabilities found diagnoses of attention deficit disorders, conduct with clinic samples (e.g., Schwab-Stone et al., STRUCTURED INTERVIEWS FOR CHILDREN AND ADOLESCENTS 381

1993). This is very disconcerting, however, given taker can provide for the child's physical and that the DISC was designed for use in epidemio- emotional needs. logical studies of non-referred community sam- An interview called the Adolescent Adaptive ples. Process Scales have also been developed to assess competence and adaptive behavior in adolescence (Beardslee, Jacobson, Hauser, Noam, & Powers, Other Interviews 1985). This clinical interview covers areas such as relationships with others, performance of The range of interview schedules available for age-appropriate tasks, thinking, and impulse con- use has been expanded in several ways. The trol. Scores derived for the interview data appear to Teacher Interview for Psychiatric Symptoms have good reliability and correlate with an inde- (TIPS) (Kerr & Schaeffer, 1987), for example, is a pendent measure of ego development (Beardslee, semi-structured interview designed to obtain diag- Jacobson, Hauser, Noam, Powers, Houlihan, & nostic information from teachers. Modeled after Rider, 1986). the K-SADS and ISC, the TIPS comprises 46 ques- Lastly, the Child and Adolescent Services tions about psychiatric symptoms that might be Assessment (CASA) is an interview-based mea- evident in school (e.g., general anxiety, attention sure of amount and type of mental health services deficits). The TIPS takes about 45 minutes to com- received by children and adolescents ages 8-18 plete and can be administered over the telephone. years (Farmer, Angold, Burns, & Costello, 1994). The interview begins with questions about the teacher's own teaching experience and style, then moves on to specific child symptoms which paral- RESEARCH FINDINGS lel the Interview Schedule for Children (ISC) (Kovacs, 1982) in format. The teacher is then Taken as a whole, the last 20 years of research asked 11 questions about the child's grooming, on structured diagnostic interviews for children social popularity, school performance, and family and adolescents indicate some very sobering con- problems. clusions. First of all, reliability of most interview Most research has focused on diagnostic inter- schedules has been mediocre. Across all diag- views, but several interview schedules have been noses, test-retest reliabilities have averaged about developed to assess non-diagnostic aspects of chil- .40- .50 (kappa statistic). Reliabilities have typi- cally been somewhat higher (.50 -.60) for Disrup- dren's adaptation, social adjustment, and utiliza- tive Behavior Disorders, and lower (.20- .40) for tion of mental health services. The Social affective/anxiety disorders. It has been rare for any Adjustment Inventory for Children and Adoles- study to achieve test-retest reliability above cents (SAICA) is a new interview schedule for kappa=.75 for any diagnosis--a commonly cited assessing children's adaptive functioning in sev- cutoff point for "excellent" reliability (Landis & eral domains (John, Gammon, Prusoff, & Werner, Koch, 1977). Most studies have achieved "poor" to 1987). The SAICA is a semi-structured interview "fair" reliabilities for most diagnoses. covering children's social and adaptive function- Second, most interviews have achieved about ing in school, in their spare time, and with peers, the same (mediocre) level of reliability. This is siblings, and parents. It can be used to interview somewhat surprising, given the structural and pro- children and adolescents directly or to interview cedural differences between interview schedules. parents about their children. The SAICA is a very It appears that wide variations in administration useful supplement to diagnostic assessment proce- procedure, item structure and wording, level of dures. interview training, and so on., have little net The Child and Adolescent Functional Assess- impact on diagnostic reliability. One exception to ment Scale (CAFAS) was designed to determine this general pattern of findings is the higher-than- the extent to which a psychiatric disorder is disrup- typical reliability for Internalizing diagnoses tive to the child's normal functioning (Hodges, obtained by the CAPA interview (Angold & Cos- 1994). The CAFAS covers five areas: role perfor- tello, 1995). These findings need to be replicated, mance, thinking, behavior towards others, moods/ but they might represent one benefit of trying to self harm, and substance abuse. Interview ques- determine the clinical significance of reported tions also assess degree to which the child's care- symptoms during the interview process. 382 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT

Third, there has been little improvement in reli- low. If that is the case, then further tinkering with ability in the last 15 years, despite extensive efforts interview schedules is a waste of time, and more to refine the interview schedules themselves. This creative work on getting more and better informa- point is controversial, with some investigators see- tion from human respondents will be needed. ing more improvement than others (see Costello, Burns, Angold, & Leaf, 1993; Hodges, 1994; Shaf- fer, 1994). But the weight of evidence suggests that SUMMARY increases in reliability, if there are any, are mini- mal to say the least. In the last fifteen years, more work has been Fourth, the "information yield"--literally what done on the development and testing of structured one gets out of the assessment process--has been interviews for children and adolescents than all disappointingly low for most interviews. Many previous years. Nevertheless, structured interview- investigators have found that interviews require ing with children is relatively new and most inter- lengthy interviewer training, they are very view schedules are still evolving. Research on the time-consuming, require extensive data manage- reliability and validity of structured interviews is ment and exceedingly complex algorithmic scor- still needed. Many interview schedules are reliable ing. This effort is typically not repaid with a rich enough for making global distinctions among database, but rather disappointingly crude diag- groups. Whether or not they are reliable enough for noses of general categories (e.g., "Any Anxiety idiographic description and diagnosis remains to Disorder," or "Disruptive Behavior Disorders," or be seen. alternatively, symptom scores for global syn- Validation efforts have increased dramatically in dromes such as "Internalizing" and "Externaliz- the past few years. The most common approaches ing." Only at the end of a study does an to testing validity have been: (a) comparing crite- investigator realize that they could have obtained rion groups such as clinically referred and non- equivalent information using simple, quick, and referred samples; (b) determining convergent rela- inexpensive symptom checklists and behavioral tions with other indices of child psychopathology, rating scales. Structured interviews are often particularly child-behavior rating scales; and (c) selected for use with the implicit hope and expec- determining convergence between different inter- tation of yielding more richness, detail, and con- view schedules. Overall, most interview schedules textual depth, than expedient paper-and-pencil have performed quite well, certainly well enough measures. But most interview schedules yield only to warrant continued development and testing. crude symptom counts and diagnosesmnot rich A range of applications has also been explored, descriptive data. To use Raymond Cattell's term, including screening, description, and diagnosis. As structured interviews have a very high "dross rate." screening tools, structured interviews are more Most of the information collected is not used. costly and time-consuming than checklists and rat- Finally, there have been few innovations in ing scales. Their screening performance is also structured interviewing in the modern era. Most usually no better and often much worse than much child interviews represent downward extrapola- cheaper paper-and-pencil assessment techniques. tions of adult interviews, which themselves have As descriptive tools, structured interviews are not changed much over the decades. Much of the roughly comparable to checklists and rating scales effort to revise and refine child interviews has been in terms of reliability and information yield. How- at the level of exact item-wording or in structural ever, they lack the psychometric development and details such as skip structure, and so on. It seems normative standardization of many rating scales safe to say from our historical vantage point, that and are probably not the best choice if the goal is this has been misplaced precision. description only. The advantage of interviews lies Taking these general conclusions together, it primarily in their diagnostic applications. Unlike appears that the limiting factor in interviewing lies most checklists and rating scales, many interview not in the details of the interview schedules them- schedules are keyed to specific diagnostic criteria selves but in the human respondent. Perhaps and cover not only symptom presence and severity almost any interview schedule will quickly hit the but also the onset, duration, and associated impair- maximum yield, reliability, and validity of infor- ments necessary for formal diagnoses. mation provided orally by another human being. Even so, no single interview schedule can be Perhaps, as it appears, that level is disappointing recommended for diagnostic assessments. Rather, STRUCTURED INTERVIEWS FOR CHILDREN AND ADOLESCENTS 383 different types of interviews seem suited to differ- with children. One study (Edelbrock et al., 1985) ent purposes. Both the K-SADS and ISC were found that reliability of child reports was low for developed to select subjects for research on child- children below the age of 10, but increased to mod- hood depression and they serve that purpose very erately high levels through middle childhood and well. Both are focused on affective symptoms and adolescence. Parent-child agreement has also been provide precise and detailed information about low in most studies, although this depends upon symptom severity and chronicity. These interviews many factors, such as the area being assessed, the are semi-structured and are intended for clinically age of the children, and the clinical status of the sophisticated interviewers having extensive instru- respondents. Low parent-child agreement is not ment-specific training. To the extent that they tap necessarily an indictment of the interview sched- symptoms in other areas, the ISC and the K-SADS ules, since they may be accurately reflecting true can also be recommended more generally for pur- differences in the way parents and children view poses of differential diagnosis among clinically child functioning. However, low agreement does referred samples. raise the complex issue of how to deal with dispar- The DISC and the DICA are at variance with the ate data from different informants. Researchers ISC and K-SADS. They are highly structured have begun to try different strategies for integrat- interviews in which the role of clinical judgment ing data from parent and child interviews, particu- has been minimized. Both cover a broad range of larly when trying to formulate diagnoses (see symptoms and disorders and are suitable for Young, O'Brien, Gutterman, & Cohen, 1987). large-scale studies employing lay interviewers. For A final issue involves taxonomic progress these reasons, they seem more useful for describ- within child psychiatry. The diagnostic interviews ing symptom prevalence and distribution among are tied to the prevailing taxonomy of child disor- non-referred populations, rather than for purposes ders (i.e., the DSM). Validity of the interviews is of differential diagnosis of identified cases. simultaneously built upon and limited by the valid- ity of the taxonomic system. Wholesale changes in diagnostic interviews were mandated by the advent Future Directions of the DSM-III-R and DSM-IV (American Psychi- atric Association, 1987, 1994), which embodies Research on the reliability and validity of struc- many substantive changes in the categories and cri- tured interviews for children will undoubtedly con- teria applied to children and youth. Some diagnos- tinue for many years. It seems unlikely that many tic interviews (e.g., DISC, DICA, K-SADS) were new interview schedules will be developed, but rekeyed to the DSM-III-R, thento the DSM-IV, but rather that research will concentrate on the handful research on their performance is not yet available. of interviews already available. The more highly developed and tested interview schedules, such as the K-SADS, CAS, DICA, and DISC, will become REFERENCES standard assessment and diagnostic tools in clini- cal and epidemiologic research. Although such Angold, A., & Costello, E. J. (1995). A test-retest reli- studies will not be directed at testing the interviews ability study of child-reported psychiatric symp- themselves, their results will certainly contribute to toms and diagnoses using the Child and the evaluation and ultimate refinement of the inter- Adolescent Psychiatric Assessment (CAPA-C). view schedules. Psychological Medicine, 25, 755-762. Researchers using structured interviews will Angold, A., Prendergast, M., Cox, A., Harrington, R., also have to face many unsolved problems and Simonoff, E., & Rutter, M. (1995). The Child and issues. A key question is whether children are reli- Adolescent Psychiatric Assessment (CAPA). Psy- able and valid reporters of their own social, emo- chological Medicine, 25, 739-753. tional, and behavioral functioning. The ability to American Psychiatric Association. (1956). Diagnos- directly question children about themselves is a tic and statistical manual of mental disorders (lst major strength of interviewing and was one of the ed.). Washington, DC: Author. major stimuli to the development of structured American Psychiatric Association. (1968). Diagnos- interviews. However, many studies have obtained tic and statistical manual of mental disorders (2nd disappointingly low reliability from interviews ed.). Washington, DC: Author. 384 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT

American Psychiatric Association. (1980). Diagnos- Center for Epidemiologic Studies. National Insti- tic and statistical manual of mental disorders (3rd tute of Mental Health. ed.). Washington, DC: Author. Costello, A. J., Edelbrock, C., Kalas, R., Kessler, M. American Psychiatric Association. (1987). Diagnos- D., & Klaric, S. H. (1982). The NIMH Diagnostic tic and statistical manual of mental disorders (3rd Interview Schedule for Children (DISC). Unpub- ed. rev.). Washington, DC: Author. lished interview schedule. Pittsburgh: Department American Psychiatric Association. (1994). Diagnos- of Psychiatry, University of Pittsburgh. tic and statistical manual of mental disorders (4th Edelbrock, C., & Costello, A. J. (1988). Convergence ed.). Washington :D.C.:Author. between statistically derived behavior problem Beardslee, W., Jacobson, A., Hauser, S., Noam, G., & syndromes and child psychiatric diagnoses. Jour- Powers, S. (1985). An approach to evaluating ado- nal of Abnormal Child Psychology, 16, 219-231. lescent adaptive processes: Scale development Edelbrock, C., Costello, A. J., Dulcan, M. K., and reliability. Journal of the American Academy Conover, N. C., & Kalas, R. (1986). Parent-child of Child Psychiatry, 24, 637-642. agreement on child psychiatric symptoms Beardslee, W., Jacobson, A., Hauser, S., Noam, G., & assessed via structured interview. Journal of Child Powers, S., Houlihan, J., & Rider, E. (1986). An Psychology and Psychiatry, 27, 181-190. approach to evaluating adolescent adaptive pro- Edelbrock, C., Costello, A. J., Dulcan, M. K., Kalas, cesses: Validity of an interview-based measure. R., & Conover, N. C. (1985). Age differences in Journal of Youth and Adolescence, 15, 355-375. the reliability of the psychiatric interview of the Bierman, K. (1984). Cognitive development and clin- child. Child Development, 56, 265-275. ical interviews with children. In B. Lahey and A. Endicott, J., & Spitzer, R. (1978). A diagnostic inter- Kazdin (Eds.), Advances in clinical child psychol- view: The Schedule for Affective Disorders and ogy (Vol. 6, pp. 217-250). New York: Plenum. Schizophrenia. Archives of General Psychiatry, Carlson, G. A., Kashani, J., Thomas, M., Vaidya, A., 35, 837-844. & Daniel, A. E. (1987). Comparison of two struc- Fallon, T., & Schwab-Stone, M. (1994). Determi- tured interviews on a psychiatrically hospitalized nants of reliability in psychiatric surveys of chil- population of children. Journal of the American dren aged 6-12. Journal of Child Psychology and Academy of Child Psychiatry, 26, 645-648. Psychiatry, 35, 1391-1408. Chambers, W., Puig-Antich, J., Hirsch, M., Paez, P., Farmer, E., Angold, A., Burns, B., & Costello, E. J. Ambrosini, P., Tabrizi, M. A., & Davies, M. (1994). Reliability of self-reported service use: (1985). The assessment of affective disorders in Test-retest consistency of children's responses to children and adolescents by semi-structured inter- the child and adolescent services assessment view: Test-retest reliability of the K-SADS. (CASA). Journal of Child and Family Studies, 3, Archives of General Psychiatry, 42, 696-702. 307-325. Cohen, P., O'Connor, P., Lewis, S., Velez, N., Mal- Freeman, M. (1971). A reliability study of psychiatric achowski, B. (1987). Comparison of DISC and diagnosis in childhood and adolescence. Journal K-SADS interviews of an epidemiologic sample of Child Psychology and Psychiatry, 12, 43-54. of children. Journal of the American Academy of Group for the Advancement of Psychiatry. (1957). Child Psychiatry, 26, 662-667. The diagnostic process in child psychiatry. New Costello, E. J., Bums, B., Angold, A., & Leaf, P. York: Author. (1993). How can epidemiology improve mental Group for the Advancement of Psychiatry. (1966). health services for children and Adolescents? Psychopathological disorders in childhood: The- Journal of the American Academy of Child Psy- oretical considerations and a proposed classifica- chiatry, 32, 1106-1113. tion. New York: Author. Costello, E. J., Edelbrock, C., & Costello, A. J., Herjanic, B., & Campbell, W. (1977). Differentiating (1985). The validity of the NIMH Diagnostic psychiatrically disturbed children on the basis of a Interview Schedule for Children. Journal of structured interview. Journal of Abnormal Child Abnormal Child Psychology, 13, 579-595. Psychology, 5, 127-134. Costello, A. J., Edelbrock, C., Dulcan, M. K., Kalas, Herjanic, B., Herjanic, M., Brown, F., & Wheatt, T. R., & Klaric, S. H. (1984). Development and test- (1975). Are children reliable reporters? Journal of ing of the NIMH Diagnostic Interview Schedule Abnormal Child Psychology, 3, 41-48. for Children in a clinic population. Final report Herjanic, B., & Reich, W. (1982). Development of a (Contract # RFP-DB-81-0027). Rockville, MD: structured interview for children: Agreement STRUCTURED INTERVIEWS FOR CHILDREN AND ADOLESCENTS 385

between child and parent on individual symptoms. Kestenbaum, C.J., & Bird, H. (1978). A reliability Journal of Abnormal Child Psychology, 10, 307- study of the Mental Health Assessment Form for 324. school age children. Journal of the American Hodges, K. (1994). The Child and Adolescent Func- Academy of Child Psychiatry, 7, 338-347. tional Assessment Scale: Instructions for scoring. Kovacs, M. (1982). The Interview Schedule for Chil- Test manual available from the author, Eastern dren (ISC). Unpublished interview schedule. Michigan University, Psychology Dept., Ypsi- Department of Psychiatry, University of Pitts- lanti, MI 48197. burgh. Hodges, K. (1994). Debate and argument: Reply to Landis, J. R., & Koch, G. G. (1977). The measure- David Shaffer: Structured interviews for assessing ment of observer agreement for categorical data. children. Journal of Child Psychology and Psychi- Biometrics, 33, 159-174. atry, 35, 785-787. Lapouse, R., & Monk, M. A. (1958). An epidemio- Hodges, K., Cools, J., & McKnew, D. (1989). logic study of behavior characteristics of children. Test-retest reliability of a clinical research inter- American Journal of Public Health, 48, 1134- view for children: The Child Assessment Sched- 1144. ule. Psychological Assessment, 1, 317-322. Matarazzo, J. D. (1983). The reliability of psychiatric Hodges, K., Gordon, Y., & Lennon, M. P. (1990). and psychologic diagnosis. Clinical Psychology Parent-child agreement on symptoms assessed via Review, 3, 103-145. a clinical research interview for children: The Orvaschel, H., Puig-Antich, J., Chambers, W., Child Assessment Schedule (CAS). Journal of Tabrizi, M. A., & Johnson, R. (1982). Retrospec- Child Psychology and Psychiatry, 31,427-436. tive assessment of prepubertal major depression with the Kiddie-SADS-E. Journal of the American Hodges, K., Kline, J., Stern, L., Cytryn, L., & McK- Academy of Child Psychiatry, 21,392-397. new, D. (1982). The development of a child Puig-Antich, J., Blau, S., Marx, N., Greenhill, L., & assessment interview for research and clinical use. Chambers, W. (1978). Pre-pubertal major depres- Journal of Abnormal Child Psychology, 10, 173- sive disorder: A pilot study. Journal of the Amer- 189. ican Academy of Child Psychiatry, 17, 695-707. Hodges, K., McKnew, D., Burbach, D. J., & Roe- Puig-Antich, J., & Chambers, W. (1978). The Sched- buck, L. (1987). Diagnostic concordance between ule for Affective Disorders and Schizophrenia for the Child Assessment Schedule (CAS) and the school-aged children. Unpublished interview Schedule for Affective Disorders and Schizophre- schedule. New York State Psychiatric Institute, nia for School-Age Children (K-SADS) in an out- New York, NY. patient sample using lay interviewers. Journal of Puig-Antich, J., Chambers, W., Halpern, F., Hanlon, the American Academy of Child Psychiatry, 26, C., & Sacher, E. (1979). Cortisol hypersecretion in 654-661. prepubertal depressive illness: A preliminary Hodges, K., McKnew, D., Cytryn, L., Stern, L., & study. Psychoneuroendocrinology, 4, 191-197. Kline, J. (1982). The Child Assessment Schedule Puig-Antich, J., Perel, J. M., Lupatkin, W., Cham- (CAS) diagnostic interview: A report on reliability bers, W., Shea, C., Tabrizi, M. A., & Stiller, R. L. and validity. Journal of the American Academy of (1979). Plasma levels of imipramine (IMI) and Child Psychiatry, 21,468-473. desmethylimipramine (DSI) and clinical response Jensen, P., Roper, M., Fisher, P., & Piacentini, J. in prepubertal major depressive disorder: A pre- (1995). Test-retest reliability of the Diagnostic liminary report. Journal of the American Academy Interview Schedule for Children (DISC 2.1). of Child Psychiatry, 18, 616-627. Archives of General Psychiatry, 52, 61-71. Reich, W., Herjanic, B., Welner, Z., & Gandhy, P. R. John, K., Gammon, G.D., Prusoff, B., & Werner, V. (1982). Development of a structured psychiatric (1987). The Social Adjustment Inventory for Chil- interview for children: Agreement on diagnosis dren and Adolescents (SAICA): Testing of a new comparing parent and child. Journal of Abnormal semistructured interview. Journal of the American Child Psychology, 10, 325-336. Academy of Child and Adolescent Psychiatry, 26, Robins, L., Helzer, J. E., Croughan, J., & Ratcliff, K. 916-921. S. (1981). National Institute of Mental Health Kerr, M.M., & Schaeffer, A.L. (1987). Teacher Inter- Diagnostic Interview Schedule: Its history, char- view for Psychiatric Symptoms (TIPS). Pittsburgh, acteristics, and validity. Archives of General Psy- PA: Author. chiatry, 38, 381-389. 386 HANDBOOK OF PSYCHOLOGICAL ASSESSMENT

Rosen, B. M., Bahn, A. K., & Kramer, M. (1964). Shaffer, D. (1994). Debate and argument: Structured Demographic and diagnostic characteristics of interviews for assessing children. Journal of Child psychiatric clinic patients in the USA. American Psychology and Psychiatry, 35, 783-784. Journal of Orthopsychiatry, 34, 455-468. Sylvester, C. E., Hyde, T. S., & Reichler, R. J. (1987). Rubio-Stipec, M., Canino, G., Shrout, P., Dulcan, M., The Diagnostic Interview for Children and Per- Freeman, D., & Bravo, M. (1994). Psychometric sonality Inventory for Children in studies of chil- properties of parents and children as informants in dren at risk for anxiety disorders or depression. child psychiatry epidemiology with the Spanish Journal of the American Academy of Child and Diagnostic Interview Schedule for Children Adolescent Psychiatry, 26, 668--675. (DISC.2). Journal of Abnormal Child Psychology, Tarter, R., Templer, D., & Hardy, C. (1975). Reliabil- 703-720. 22, ity of the psychiatric diagnosis. Diseases of the Rutter, M., & Graham, P. (1968). The reliability and Nervous System, 36, 30-31. validity of the psychiatric assessment of the child: Weinstein, S., Stone, K., Noam, G., Grimes, K., & Interview with the child. British Journal of Psy- Schwab-Stone, M. (1989). Comparison of DISC chiatry, 11,563-579. with clinicians' DSM-III diagnoses in psychiatric Sandifer, M. G., Pettus, C. M., & Quade, D. (1964). A inpatients. Journal of the American Academy of study of psychiatric diagnosis. Journal of Nervous 53-60. and Mental Disease, 139, 350-356. Child and Adolescent Psychiatry, 28, Welner, Z., Reich, W., Herjanic, B., Jung, K., & Schwab-Stone, M., Fallon, T., Briggs, M., & Crowther, B. (1994). Reliability of diagnostic Amado, H. (1987). Reliability, validity, and par- reporting for children aged 6-11 years: A ent-child agreement studies of the Diagnostic test-retest study of the Diagnostic Interview Interview for Children and Adolescents (DICA). Schedule for Children-Revised. American Journal Journal of the American Academy of Child Psy- of Psychiatry, 151, 1048-1054. chiatry, 26, 649--653. Schwab-Stone, M., Fisher, P., Piacentini, J., Shaffer, Young, G., O'Brien, J., Gutterman, E., & Cohen, P. D., Davies, M., & Briggs, M. (1993). Journal of (1987). Research on the clinical interview. Jour- the American Academy of Child and Adolescent nal of the American Academy of Child and Ado- Psychiatry, 32, 651-657. lescent Psychiatry, 26, 613-620.