Foundations of the General Approach to Psychological Assessment
Psychological assessments must be founded upon specific theoretical premises that guide the assessment process. The history of psychological assessment is quite extensive, resulting in many theoretical stances upon which assessments are based. It is our belief, however, that psychological assessment of adults in mental health settings is based on two founding premises: assessments must be evidence-based and multimodal.
Psychological assessment in mental health settings must be evidence-based. That is, a client's psychiatric symptoms must be systematically assessed in relation to the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV; American Psychiatric Association, 1994) criteria for particular disorders. This criteria analysis is then supplemented with results from empirically validated psychological tests and structured interviews. The client's responses on these measures are used to indicate likely diagnoses and appropriate treatment implications based on empirical research. Consequently, an evidence-based approach to psychological assessment requires empirical support for any conclusions or recommendations, as opposed to relying solely on clinical impression and judgment.
The evidence-based approach was initially practiced in general medicine and recently has been incorporated in psychology. Evidence-based medicine (EBM) integrates clinical expertise with external evidence based on systematic research, while considering the values and expectations of patients or clients (Gambrill, 1999). Within the medical community, EBM is defined as a set of strategies designed to ensure that the clinicians form opinions and base subsequent decisions on the best available external evidence (Geddes, 1997). Thus, decisions pertaining to the client are made in light of the most up-to-date information available. The steps involved in EBM include a precise definition of the clinical problem, an efficient search for the best available evidence, critical appraisal of the evidence, and integration of the research findings with clinical expertise (Geddes, 1997; Olson, 1996). At each stage of EBM, recent developments in clinical findings and information technology are harnessed and utilized (Geddes, 1997). As physicians have acknowledged that no single authority has comprehensive scientific knowledge, the EBM approach is viewed as an improvement over the authoritative knowledge approach to practicing medicine (Kennell, 1999).
Within the domain of psychological assessment, an evidence-based approach emphasizes the importance of systematic observation and the use of rules of evidence in hypothesis testing. Thus, psychologists base their assessments and diagnoses on the best available evidence (Bensing, 2000). This approach to psychological assessment affords the opportunity to integrate real clinical problems with critical evaluation of the psychiatric research literature (Gilbody, 1996). In essence, an evidence-based approach to psychological assessment is premised on obtaining actuarial evidence from both structured interviews and objective measures that have been empirically supported. Empirical and clinical literature suggests patterns of symptoms that are associated with specific diagnoses and provide treatment implications, thereby enhancing the likelihood of making an accurate diagnosis.
The evidence-based approach is distinct from the more established and popular approach based on clinical judgment. Evidence-based actuarial assessments proceed in accordance with a prespecified routine and are based on empirically derived relations between data and the trait or event of interest (Dawes, Faust, & Meehl, 1989; Wiens, 1991). In contrast, clinical judgment consists of decisions made in the clinician's mind. In its most polar form, this distinction is analogous to a dimension with objectivity (evidence-based) on one end and subjectivity (clinical impression) at the other end.
The clinicians who base their assessments on clinical judgment highlight the advantages of their technique. First, certain assessment tools, such as unstructured interviews and behavioral observations, cannot be empirically evaluated or subjected to statistical analyses required by the evidence-based model. In fact, clinical judgment is required to evaluate the results of such tools. The results provide clinicians with a plethora of information, including clinical impressions as to the nature of clients' difficulties and distresses. Second, clinicians' impressions and judgments structure the rest of the assessment and provide a framework around which the client's symptoms and difficulties are conceptualized and understood. Third, many clinicians contend that their clinical impressions and judgments are rarely disputed by empirical test results. Thus, in the interest of conducting an efficient assessment, they rely solely on their judgment gleaned from information obtained from unstructured interviews. Fourth, some clinicians fear that by basing a diagnosis on empirical findings, they will be treating the client nonoptimally through reliance on actual experience (Meehl, 1973). Furthermore, many clinicians often shun actuarial-based data for fear that the data themselves will involve significant error, thereby leading to misdiagnosis of a client. Consequently, reliance on one's own experience and judgment rather than actuarial-based data when making diagnoses and treatment recommendations remains a popular method by which clinicians conduct psychological assessments.
Despite the historical popularity of basing assessments on clinical judgments, the validity of such judgments is often low, thereby placing the client at potential risk for underdiagnosis, overdiagnosis, or misdiagnosis (Faust & Ziskin, 1988; see also chapter by Garb in this volume). Clinical inference and judgment involve probabilistic transitions from clients' observable or reported episodes to their dispositions. Ideally, such inferences should be based upon an extensive actuarial experience providing objective probability statements (Meehl, 1973). However, in reality, this ideal is rarely achieved, because often the conditional probabilities are judged based solely on a clinician's experience, rather than on empirical findings. Consequently, permitting a weak or moderately strong clinical inference to countervail a well-supported set of actuarial data on patients similar to one's client will lead to an increase in erroneous clinical decisions (Meehl, 1973).
Faust and Ziskin (1988) also highlighted some of the disadvantages of clinical judgment. For example, they noted that clinicians often overvalue supportive evidence and undervalue evidence contrary to their hypotheses. They stated that clinicians tend to find evidence of abnormality in those they assess, regardless of whether they have any psychopathology. In addition, they argued that clinicians tend not to receive any outcome information about their clients; therefore, they are unable to learn whether their predictions were accurate and their suggestions were helpful. In summary, although the clinical impression approach has some merits, the validity and utility of the evidence-based approach is making this new format the standard for psychological assessment. Indeed, Hersen, Kazdin, and Bellack (1991) suggested that as the extent of the relevant research increases, the use of actuarial procedures will also increase.
Finally, contrary to popular opinion, clinical judgments and evidence-based models do not generate the same conclusions. Meehl (1973) contends that human judgment and statistical predictions concerning diagnosis, prognosis, and decisions based on the same set of information have a less than perfect correlation. Dawes et al. (1989) reviewed research comparing clinical judgment to actuarial judgment. They pointed out that with the same set of data, different actuarial procedures lead to the same conclusion, whereas differenthumanjudgments may result in several different conclusions. Moreover, Dawes et al. stated that clinicians' diagnoses can fall prey to self-fulfilling prophecy in that their predictions of diagnoses can influence their decisions about symptom prevalence and, later, diagnosis. Moreover, they noted that the mathematical nature of actuarial procedures ensures that each variable has predictive power and is related to the criterion in question (valid vs. invalid association with the criteria). In contrast, clinicians may deal with a limited and unrepresentative sample of individuals; therefore, they may not be able to determine accurate relations between variables. Furthermore, clinical judgment is prone to human error. Neither procedure, however, is infallible. Therefore, the actuarial procedures should be reassessed periodically.
The approach to psychological assessment in mental health settings should also be multimodal. One assessment tool is not sufficient to tap into complex human processes. Moreover, given that empirical support is critical to the validity of a psychological assessment, it is just as essential that there is concordance among the results from the client's history, structured interview, self-report, objective tests, and clinical impression. Because the results and interpretations are obtained from several sources, the multimodal approach increases reliability of the information gathered and helps corroborate hypotheses (Hertzman, 1984). Moreover, this approach draws on the strengths of each test and reduces the limitations associated with each test. Amultimodal approach has the benefit of relying on shared methods and thus minimizing any potential biases associated with specific assessment methods or particular instruments. Finn and Butcher (1991) note that objective tests are imperfect and the results should not be seen as definitive conclusions but, rather, as hypotheses that should be compared with information from other sources. A diagnosis can be made with more confidence when several independent sources of information converge than when inferences are based on a single source. Moreover, the multimodal approach prevents the influence of a single perspective from biasing the results (Beutler, Wakefield, & Williams, 1994).
An evidence-based and multimodal approach to psychological assessment enables the clinician to attain the main goals of assessment, namely clarifying diagnosis and providing treatment recommendations. Whereas other authors have emphasized additional assessment goals such as insight into a client's personality, interpersonal style, and underlying drives, we think that the goals of clarifying diagnosis and guiding treatment are the mainstays of psychological assessment and, in fact, incorporate many of the other goals.
A primary reason for conducting psychological assessments of adults in a mental health setting is to make or clarify a diagnosis based on the client's presenting symptomatology. This is a common issue when the client presents with symptoms that are common to several diagnoses or when there is a concern that the symptoms of one disorder may be masking the symptoms of another disorder (Olin & Keatinge, 1998). Adhering to an evidence-based multimodal approach ensures that cross-validated actuarial evidence is obtained, thereby enhancing the validity of the diagnosis and increasing the clinician's confidence in the diagnosis.
Clinicians are often asked to make a differential diagnosis. However, the either-or implication of differential diagnosis is problematic. Often, clients manifest criteria of several disorders simultaneously, or they may manifest symptoms that do not meet criteria for a specific disorder despite the fact that their behaviors and cognitions are maladaptive (Maloney & Ward, 1976; Westen & Arkowitz-Westen, 1998). Thus, clinicians may find it beneficial to use multiple diagnostic impressions and, if possible, determine which disorder is generating the most distress and requires immediate attention.
Making or clarifying one or more diagnoses can benefit the clinician in many ways. These benefits include the following: enhancing communication between clinicians about clients who share certain features; enhancing communication between a clinician and the client through feedback; helping put the client's symptoms into a manageable and coherent form for the client; giving the client some understanding of his or her distress; guiding treatment; and enhancing research that, in turn, should feed back into clinical knowledge (Westen, 1998). Nonetheless, difficulties of psychological diagnosis should also be mentioned. Gunderson, Autry, Mosher, and Buchsbaum (1974) summarized the controversy associated with making a diagnosis:
Diagnosis, to be meaningful, must serve a function. Too often its function becomes subservient to the process of choosing a label.
Thus, although the intent of diagnosis may be the communication of information in summary form, it may actually convey misinformation if insufficient attention is paid to the complexities and variability of human behavior during the diagnostic process. (p. 22)
According to Kellerman and Burry (1981), diagnosis involves several interconnected features that must be taken into account. These include the potential for shift within any diagnostic formulation, the relationship between the presenting problem and the client's personality, acute versus chronic dimension of the pathology, the presence of various levels and types of pathology and their interconnections, and the impact of diagnostic features on the development of intervention strategies and prognostic formulations. In essence, the diagnosis of the problem is not a discrete final step but, rather, a process that begins with the referral question and continues through the collecting of data from interviews and test results (Maloney & Ward, 1976). Diagnosis is thus a complex process that incorporates a myriad of potential questions and data.
Diagnoses are dependent on meeting DSM-IV criteria for Axis I and Axis II disorders, because the DSM-IV is currently the gold standard by which to diagnose psychopathol-ogy and personality disorders. It is an operational system in which each diagnosis must be met by a necessary and sufficient number of criteria that must occur on multiple dimensions (Hertzman, 1984).
Unfortunately, there are problems inherent in making a diagnosis based on the DSM-IV, because the DSM-IV itself has certain limitations. First, it is based on a medical model and does not consider underlying processes (i.e., it is concerned only with the signs and associations of the disorder) and overall manifestations of disorders. Second, it does not address eti-ological contributions to disorders and how they affect the manifestation and outcome of disorders. Third, the Axis I and Axis II disorder criteria represent a consensual opinion of a committee of experts that labeled a particular pattern of symptoms a disorder. Traditionally, the committee's decision to assign a certain cluster of symptoms to a diagnosable condition has been based on the presence and frequency of symptoms, an empirical analysis of the symptoms' social significance, and the specificity of the symptomatic response to various classes of drugs (Beutler et al., 1994). Thus, the process of developing DSM-IV diagnoses lacked the very characteristic valued in the assessment process: relying on empirical evidence and ensuring the collection of data from a variety of sources. Fourth, the DSM-IV is categorical in nature, requiring a specified number of criteria to meet a diagnosis, even though human nature, mental illness, and mental health are distributed dimensionally.
There are numerous limitations to such a categorical approach in which mental disorders are divided into types based on criteria sets with defining features. It becomes restricted in its clinical utility when diagnostic classes are heterogeneous, when there are unclear boundaries between classes, and when the different classes are not mutually exclusive (DSM-IV-TR; American Psychiatric Association, 2000). In addition, the DSM-IV categories have overlapping boundaries, resulting in multiple diagnoses and the problem of comorbidity (Barron, 1998). Moreover, a categorical approach does not provide as powerful predictions about etiology, pathology, prognosis, and treatment as a dimensional approach (Gunderson, Links, & Reich, 1991). Fifth, the DSM-IV is skewed toward the nomo-thetic end of the spectrum, resulting in static diagnoses whose operational definitions may be inaccurate, unsupported by research findings, and camouflaging questionable construct validity (Barron, 1998). Other criticisms of the DSM-IV include excessive focus on reliability at the expense of validity, arbitrary cutoff points, proliferation of personality disorders, and questionable validity of the personality disorder clusters (Blatt & Levy, 1998).
The American Psychiatric Association has attempted to make the DSM-IV more empirical, accessible, reliable, and useful (Nathan, 1998), as well as to create an optimal balance between a respect for historical tradition, compatibility with the International Statistical Classification of Diseases and Related Health Problems, 10th edition (ICD-10; World Health Organization, 1992), evidence from reviews of the literature, analysis of data sets, results of field trials, and consensus of the field (DSM-IV-TR; American Psychiatric Association, 2000). Furthermore, many diagnostic categories are supported by empirical literature (i.e., data from DSM field trials). In summary, the DSM-IV is a descriptive classifi-catory system, ostensibly unbound to a specific theory of development, personality organization, etiology, or theoretical approach (Barron, 1998). Moreover, it is an official nomenclature that is applicable in a wide number of contexts, can be used by clinicians and researchers from various theoretical orientations, and has been used across psychological settings. The DSM-IV also attempts to address the heterogeneity of clinical presentation of symptoms by adopting a polythetic approach. That is, clients must present with a subset of items from a list of criteria in order to meet a diagnosis. In addition, the DSM-IV includes several axes to take social, medical, and economic factors into account. These merits of the DSM-IV, particularly in the absence of another comprehensive diagnostic system, suggest that assessment of psychological disorders should adhere to this multiaxial system.
The potential problem with the DSM is that it undergoes periodic revision; thus, the clinician relying on this diagnostic system would seem to be continually chasing a moving target or construct. However, except for the changes made from DSM-II to DSM-III, this system does not undergo substantial structural changes with each new version. Moreover, most tests, for example, the MMPI-2, cover most symptoms associated with a variety of syndromes. The changes in DSM from version to version usually involved carving sets of symptoms into different syndromes. Thus, the omnibus inventories designed to assess a variety of psychiatric symptoms are not necessarily affected by these changes, because the fundamental symptoms of most disorders remain captured.
A second and equally important goal of psychological assessments of adults in a mental health setting is to offer a guide for treatment by developing an individualized treatment plan for the client (and family). A psychological assessment offers the opportunity to link symptomatology, personality attributes, and other information with certain treatment modalities or therapeutic targets. Therefore, giving treatment recommendations allows psychologists to proceed past the level of diagnosis and provide suggestions about how to deal with the diagnosed disorder. In fact, diagnosis has most utility when it can be related to treatment. Ideally, an outline of treatment recommendations should include plans to immediately deal with the client's acute symptoms, as well as long-term treatment plans that address the client's chronic symptoms, personality features, coping mechanisms, and interpersonal problems, and stressors within the client's environment (Hertzman, 1984). Moreover, treatment recommendations must provide suggested changes as well as methods for implementing these changes (Maloney & Ward, 1976). In short, treatment recommendations should include short-term and long-term goals, procedures to reach the goals, possible obstacles to treatment, and prognosis of the client.
The process of diagnostic clarification, often the first and primary goal of psychological assessment, often serves as a guide to treatment. Certain treatment protocols are suggested by way of the diagnosis, whereas other treatments may be excluded by virtue of failing to meet criteria for a certain disorder (Hertzman, 1984). However, treatment planning is complicated, because the relationship between diagnosis and treatment is not always simple. Due to the nature of psychiatric difficulties, a client's symptomatology may result from multiple causal pathways, thereby contributing to imprecise treatment (Clarkin & Mattis, 1991). Nonetheless, diagnosis can provide important useful information.
Although diagnosis is often a first step in the treatment planning process, the ability to offer treatment recommendations must go beyond diagnosis and assess a variety of qualities and variables that best describe the client (Halleck, 1991). Treatment planning should take into account information about symptom severity, stage of problem resolution, general personality attributes, interpersonal style, coping mechanisms, and patient resistance. Further sources of information include the client's psychiatric and medical history, psychological mindedness, current levels of stress, motivation levels, and history of prior treatments, as well as physical condition, age, sex, intelligence, education, occupational status, and family situation (Halleck, 1991). This information is relevant to treatment planning in two ways. First, demographic variables and a history of prior treatments can dictate or modify current treatment modalities. Second, other variables might help formulate certain etiological models that can in turn guide treatment (Halleck, 1991). Thus, information from various sources obtained in a psychological assessment can be integrated to provide treatment recommendations as well as to predict the prognosis of the client and expected effects of treatment.
In addition, Clarkin and Hurt (1988) listed several areas of patient functioning that must be evaluated to adequately inform treatment planning. These include patient symptoms, personality traits (strengths and weaknesses) and disorders, cognitive abilities and functioning, patient psycho-dynamics, patient variables that enable the patient to engage in various kinds of treatments, environmental demands, and general therapeutic enabling factors (Clarkin & Hurt, 1988). In particular, patient enabling factors refer to patient dimensions that are important for treatment planning and engaging in particular forms of psychological intervention (Clarkin & Mattis, 1991). For example, the patient's defensive structure, coping style, interpersonal sensitivity, and basic tendencies and characteristics adaptations may dictate the most appropriate psychological intervention (Beutler & Clarkin, 1990; Harkness & Lilienfeld, 1997).
Psychological tests have been widely used to guide treatment. Unfortunately, the information they provide is not necessarily useful in guiding the choice of specific therapeutic modality. However, test scores can guide treatment recommendations. For example, symptom severity, stage of client resolution, recurrent interpersonal themes, level of resistance to treatment, and coping styles can be obtained from various psychological tests, and all serve as indicators for the focus and prognosis of psychotherapeutic procedures (Beutler et al., 1994). In particular, clients' scores on the MMPI-2 (Butcher, Dahlstrom, Graham, Tellegen, & Kaemmer, 1989) validity scales offer predictions about treatment based on factors such as compliance, level of insight, current psychological status, risk of premature termination of therapy, and level of motivation. Both individual scores and profiles of scores on the content and clinical scales, as well as endorsement of critical items, can also be used for treatment planning, including determining needs to be met, issues with which to deal, and structure and style of therapy (Greene & Clopton, 1994). Similarly, the Personality Assessment Inventory (PAI; Morey, 1991) can also guide treatment recommendations by providing information about a client's level of functional impairment, potential for self-harm, risk of danger to others, chemical dependency, traumatic stress reaction, and likelihood of need for medication (Morey & Henry, 1994). Furthermore, the PAI contains a number of scales that serve as either positive or negative indicators of potential for psychotherapy. Positive indicators include level of perceived distress, positive attitude toward treatment, capacity to utilize psychotherapy, availability of social supports, and ability to form a therapeutic alliance. These suitability indicators should then be weighed against negative indicators, including having disorganized thought processes, being nonpsy-chologically minded, and being characterologically unsuited for therapy (Morey & Henry, 1984).
Was this article helpful?