Psychological Assessment Screening in Medical Settings

Michael D Franzen. The Health Psychology Handbook: Practical Issues for the Behavioral Medicine Specialist. Editor: Lee M Cohen. Sage Publications. 2003.

The presence of a psychologist in specialty medical settings has been common since the late 1960s and 1970s. During that era, there was increasing attention paid to the use of psychological principles to understand or enhance the treatment of cancer patients, cardiac patients, and chronic pain patients. Recently, more attention has been given to the potential role of psychologists in general medical or family practice and pediatric clinics. These health care providers, namely pediatricians, primary practice physicians (sometime known as primary care physicians [PCPs]), and family practice doctors, are the frontline care providers, even in those situations where it might be necessary to involve specialists. For example, most of the antidepressant prescriptions in the United States are written by general practitioners. In addition, even if a diagnosis requires specialized care, the first person to come into contact with the patient is most likely to be the person’s PCP.

Psychological interventions begin with a psychological assessment. Psychological assessment has various forms, from behavioral functional analysis to personality assessment and psychological/psychiatric diagnosis. Because psychological interventions in medical settings focus more on enhancing the provision of medical care than on in-depth psychotherapy, assessment must be tailored to the needs of the setting. The psychologist might be called on to develop a plan to help manage chronic pain from a psychological perspective or to help manage anxiety related to upcoming surgery. In all of these cases, the assessment is aimed at elucidating psychological and environmental factors that could be useful in the provision of adequate medical care. In addition to formulating interventions, the psychologist might be called on to assess the patient for the presence of psychological or cognitive factors that would negatively affect the provision of medical care. These interfering factors may include cognitive impairment, psychological distress, or substance abuse behaviors.

Screening assessment is one type of psychological assessment frequently used in the medical setting. The purpose of screening is to identify important areas that may require more detailed assessment and evaluation. Therefore, the screening target is partially a function of the setting, the base rate of potential problem areas, and the population being seen. For example, the base rate of active psychosis in a family practice is fairly low, and it would make little sense to screen for symptoms of schizophrenia on a regular basis. Alternatively, there is a fairly high comorbidity between depression and certain medical conditions, and it would be eminently sensible to screen for depression in an endocrinology outpatient clinic.

Psychology in the Clinic versus Psychology in the Medical Setting

The practice of psychology in a medical setting has characteristics that differentiate it from the general practice of psychology. For example, the issue of assessment in a medical setting is complicated by the other variables in addition to environmental influences that may impinge on the patient’s behavior. In behavioral assessment, it is assumed that the medical and physiological factors have been ruled out. In personality or traditional psychological assessment, it is assumed that the medical factors are already accounted for. In neuropsychological assessment, it is assumed that environmental variables have been minimized. In screening and evaluating the medical patient in a behavioral framework, none of these assumptions can be reasonably made. In fact, a more accurate assumption would be that all of these variables are playing a role in the current clinical presentation. Therefore, it would be important for the behavioral medicine clinician to consider medical, cognitive, and psychological features. Screening assessment seeks to identify the possibility of an issue such as cognitive impairment. Psychological assessment seeks to identify the construct at issue and provide an estimate of the precise level of that construct.

There are some features of screening that distinguish it from general psychological assessment. First and most obvious, screening is conducted in a more restricted time frame than is general assessment. This is true both for the administration of procedures and for the interpretations and recommendations made on the basis of the assessment results. In general psychological assessment, an appointment may be made for the following week and the typed report may be ready a week after that. In screening, whether for an inpatient setting or an outpatient setting, the person must be screened in the same appointment as the identification of the problem is raised and the interpretation must frequently be provided to the referral source outside the consultation room. A second feature that distinguishes screening from general assessment is the length of the procedures and the sensitivity and completeness with which the target constructs are evaluated.

The three areas of most concern in general medical settings are the possibilities of neu-rologically based cognitive impairment, of psychological or psychiatric disorders, and of substance abuse. Even if a specific psychiatric diagnosis is not appropriate, it may be helpful to screen for the presence of anxiety or depression, either of which can significantly affect medical outcome. The two reasons for screening for these variables are that their presence can negatively affect the medical treatment and that their presence may indicate the need for referral for further evaluation or specialized treatment.

Psychometric Considerations in Screening

Even though psychological screening might not necessarily entail the full range of psychometric complexity that comprehensive or diagnostic assessment might, it is still important to pay attention to the relevant psychometric considerations. The several issues of validity and accuracy that need to be addressed in assessment are simplified somewhat in screening. The most relevant aspect of the psychometric properties of the screening instrument is related to its accuracy in identifying the presence of a pathological state and its utility in the decision-making process leading to a referral for comprehensive assessment. The construct underlying the instrument, whether it be memory or attention, is less important in screening than whether a score above a certain level is indicative of some form of cognitive problem with an organic basis.

In choosing a screening instrument and in setting a cutoff score, it is important to ask what the likelihood is that a certain score would be associated with a correct decision to pursue further evaluation. Similarly, it is important to ask what the likelihood is that a certain score would be associated with a correct decision to not pursue further evaluation. The first question is an issue of sensitivity. The second question is an issue of specificity. Sensitivity is the extent to which the assessment instrument identifies the presence of the target construct. Specificity is the extent to which positive assessment findings do not occur in the absence of the target construct. Positive predictive power is the accuracy with which a positive score predicts the target. Negative predictive power is the accuracy with which a negative score predicts the absence of the construct. In screening, it may be more important in some cases that an instrument be sensitive even when it might not be specific. For example, in high-risk situations where the incorrect decision to not pursue further evaluation could result in missing the presence of a growing brain tumor, the fact that many cases of no tumor are found with high scores (poor specificity) is less important than the fact that cases of tumor are found with high scores (good sensitivity).

The Use of the Interview

The most potent weapon in the armamentarium of the clinician is the interview and history. There are two sources of clinical information derived from the interview. The first is the content information such as whether certain symptoms have been noticed or brought to the attention of the patient. Other content information includes the time course of the symptoms and whether there are any consistent changes in the level of symptoms. A second important source of information comes from the clinical observations made by the clinician. The clinician can note the quality of the verbalizations of the patient. Is the articulation understandable? Is the diction accurate? Is there any word-finding difficulty, paraphasic error, or paucity of speech? Are the station and gait normal? What is the appearance of the person? Is the grooming and hygiene adequate (inadequate hygiene may reflect either depression or cognitive impairment)? These factors were addressed in greater depth in Berg, Franzen, and Wedding (1994).

In addition to using the observations to generate hypotheses regarding which areas to consider for screening purposes, the clinician can use the historical information for that purpose as well. Any changes in usual functions can be a “red flag” that cognitive screening might be useful. A history of recent minor automobile accidents might be an indication to screen for attention or for visual-spatial skills. The clinician may want to also screen for any alcohol or substance abuse. If the patient has had to change jobs or was forced into early retirement, cognitive screening may be in order. A history of multiple sex partners or of high-risk sexual behavior may also indicate the need for screening. Occupational history also gives useful information. If the patient worked in an industrial setting with exposure to solvents, heavy metals, or insecticides, cognitive screening may be indicated.

Cognitive Screening

Cognitive screening is not as simple as it may seem. Although discussion of screening frequently involves only a single instrument, there is no adequate single instrument for screening all populations in all settings. There is a range of decisions that need to be made in the choice of an appropriate instrument, and because of the nature of screening itself, interpretations and conclusions are limited. One of the first decisions is also one of the most difficult ones to make. The clinician needs to decide just which aspects of cognition are suspected of being impaired. Although there is usually considerable correlation among different cognitive skills, this correlation tends to be disrupted under conditions of impairment. It is quite possible for conversational speech to be intact while short-term memory is severely impaired and vice versa. Choosing a screening instrument that is sensitive to memory skills when the person is suspected of having a cerebral vascular accident with expressive language impairment will not be a useful endeavor.

Broad-based screening requires that a range of impairments be considered. There are a few instruments that are sensitive to a broad range of types of impairment. Their sensitivity is derived from the fact that several cognitive skills are required for adequate performance. An example is the Digit Symbol subtest of the Wechsler Adult Intelligence Scale-III (WAIS-III). The Digit Symbol test requires eye-hand coordination, visual scanning, symbolic translation, and motor speed as well as learning skill. Impairment in any one of these areas may result in poor performance on the Digit Symbol test. Other examples of such instruments include Part B of the Trail Making test and the Category test of the Halstead-Reitan Neuropsychological Battery. These instruments tend to be sensitive to psychiatric disturbance and also have significant age-related effects. Interpretation of the results of such an instrument is problematic because poor performance can be the result of any number of impairments. However, this seeming drawback is also the instrument’s strength because when the area of impairment is unknown, it is useful to have a single instrument that can detect impairment in any one of the suspected areas. The lack of specificity is not a tremendous shortcoming in the context of screening because a positive result would be followed by more extensive evaluation in which the nature and extent of the impairment could be delineated more extensively.

To guide the psychologist in the screening procedure, there is little available in terms of training at the graduate level where the emphasis is on either specialized assessment methods (e.g., neuropsychological, marital, career) or traditional intellectual and personality methods. The psychologist who wishes to learn how to screen needs to seek out supervised experience in a medical setting such as might be available at the internship level and through continuing education opportunities. In addition, there is some published material. Two excellent book resources are Clinical Neuropsychology: A Pocket Handbook (Snyder & Nussbaum, 1998) and Screening for Brain Impairment (Berg et al., 1994).

Screening for Dementia

Dementia is broadly defined as any acquired cognitive impairment sufficient to disrupt occupational, social, or adaptive functioning. There can be many causes and different manifestations. However, the most frequent cognitive difficulty evidenced by dementia patients is memory impairment. Therefore, a brief memory screening procedure can be helpful in uncovering dementia. The Digit Span procedure is relatively useless in this situation and has limited utility in screening in general. The maximum span forward in the digit procedure is relatively impervious to acquired impairment until either the late stages of a progressive condition or the severe range of an injury. Suggested procedures for memory screening include the Hopkins Verbal Learning Test (Brandt, 1991), the Rey Auditory Verbal Learning Test (Schmidt, 2000), the Brief Visual Spatial Memory Test (Benedict, 1997), and the Benton Visual Retention Test (Sivan, 1992).

Folstein Mini Mental State Exam (MMSE)

The Folstein Mini Mental State Exam (MMSE) (Folstein, Folstein, & McHugh, 1975) is perhaps the most widely used (and some say the most widely abused) instrument for the quick assessment of cognitive status. The MMSE is nearly ubiquitous among physicians, especially those working in neurological or psychiatric settings. The MMSE screens for different cognitive functions, frequently using only one item for each construct (e.g., one item taps visual-spatial construction by asking the individual to copy a drawing of two overlapping pentagons). The MMSE is heavily weighted toward orientation questions, with 10 of a possible 30 points being directed at orientation to time, place, and date.

The MMSE has been criticized for producing too many false negatives, but part of the problem may exist in the use of suboptimal cut points. The degree of accuracy depends on the eventual diagnosis of the individual (Harper, Chacko, Kotik-Harper, & Kirby, 1992), something that is not usually known at the time of the screening. The MMSE, the Mattis Dementia Rating Scale, and the Neurobehavioral Cognitive Status Exam were found to be roughly equivalent in discriminating patients with Alzheimer’s or vascular dementia from healthy elderly when optimal cut points were used rather than the cut points suggested in the literature. Because the MMSE takes less time to administer, it has an advantage over the other two tests. To increase the clinical utility of the MMSE, it would be helpful to use more extensive norms as well as norms that are sensitive to the differences associated with age and education. Grigoletto (1999) presented norms on 908 healthy Italian elderly persons. More data like this are needed.

Barbarotto, Cerri, Acerbi, Molinari, and Capatani (2000) reported data from a study of 27 patients with a variety of cognitive disorders. These authors concluded that the MMSE was less useful and less reliable when the total scores dip below 10 to 12 points. This is an important consideration in the overall use of the MMSE, but it might not be a substantial problem in a general medical setting where cognitive impairment is likely to be more subtle and total scores of 10 to 12 may be infrequent. Bidzan and Bidzan (2002) reported on a 5-year follow-up study involving 204 individuals over the age of 55 years. Eventually, 19 of these individuals were diagnosed with Alzheimer’s dementia. The Folstein MMSE, the Cognitive subscale of the Alzheimer’s Dementia Assessment Scale, the Instrumental Activity of Daily Living Scale, and the Physical Maintenance Scale were found to contribute to predicting eventual dementia, although a comparison among the instruments was not conclusive. There have been reported demographic effects on total scores of the MMSE, making it obvious that good normative information is necessary. Jones and colleagues (2002) presented normative data regarding the performance of community-dwelling elderly persons. Such data are very helpful in interpreting the scores of older persons.

Clock Drawing Test

The Clock Drawing Test (Freedman et al., 1994) is, as its name suggests, a test in which individuals are asked to draw a clock face with the hands set to “10 after 11.” After the clock is drawn from command, a line-drawing model is provided for the same task. Adunsky, Flessig, Levenkrohn, Arad, and Noy (2002) reported that the Clock Drawing Test is roughly equivalent to the MMSE in identifying impairment. However, the greater variety of items in the MMSE may provide for a broad-based evaluation. The Clock Drawing Test has its greatest utility with the elderly.

Other Cognitive Screening Tests

There are several additional screening tests available for the psychologist. Some were designed specifically for screening, whereas others are part of larger test procedures or shortened from the original versions. The Trail Making test (especially Part B) of the Halstead-Reitan Neuropsychological Battery is very sensitive to any cognitive impairment. It is easy to administer and takes less than 10 minutes to complete. The disadvantages are that it is not very specific and that psychiatric conditions such as depression can affect it. The Digit Symbol subtest of the Wechsler Adult Intelligence Scale-Revised is the most sensitive of all the Wechsler subtests. The corresponding research on the WAIS-III has not yet been conducted, but the modification introduced by the WAIS-III will probably make it more specific as well as more sensitive. Somewhat longer alternatives include the Neurobehavioral Cognitive Status Exam (Northern California Neurobehavioral Group, 1988). This test includes subtests of attention, memory, construction, language skills, and practical problem solving, but the meaning of the various subtest profiles is unclear, and it is best to stick with the total score as an indicator of cognitive impairment.

Intellectual Screening

The utility of intelligence quotient (IQ) scores is largely related to academic planning and the suitability of the person for certain services available from the state. In certain cases, a physician may question the intelligence level of a patient where the documentation of mental retardation may make the patient eligible for government-reimbursed support services. In other instances, there may be questions about the capacity of the person to follow a complicated medical regimen or about the competency of the person to make decisions related to medical care. Screening assessment is insufficient to answer these questions. However, screening for intellectual capacity may guide the clinician in deciding whether to refer the person for a more complete evaluation. Although short forms of the Wechsler Adult Intelligence Scale are available (Kulas & Axelrod, 2002), and there is even a short form available from the publisher (Axelrod, 2002), it appears that a seven-subtest short form is as abbreviated as one can go and still expect reasonable reliability (Axelrod, Ryan, & Ward, 2001). Other options include the use of short tests and procedures that have been shown to correlate acceptably (but not optimally) with longer, more comprehensive intellectual exams such as the WAIS-III and the Stanford-Binet-IV. These short tests include the Slosson Intelligence Test, the Beta-III, and the Test of Nonverbal Intelligence-2.

The Slosson Intelligence Test is a short test based on the Stanford-Binet. Although short and easy to administer, it may have significant limitations in estimating IQ in the lower ranges (Kunen, Overall, & Salles, 1996). The Beta-III (Kellogg & Morton, 2001) is a set of five nonverbal procedures that can be administered in less than 15 minutes and give a reasonably culture-fair estimate of intelligence. The Test of Nonverbal Intelligence-2 (Brown, Sherbenou, & Johnsen, 1990) also provides a reasonable estimate of culture-fair intelligence by assessing visual abstraction skills. It does not have as many different types of tasks as the Beta-III, and it is somewhat shorter in administration time.

Psychiatric Screening

Affective problems are typically brought to the attention of primary care providers (PCPs) first. This is not sufficient unless the PCPs can accurately recognize emotional disorders. Wittchen and colleagues (2002) reported that in a study of more than 20,000 patients, PCPs were able to reasonably well identify the presence of serious psychiatric disorders but were not as accurate in determining the actual diagnoses. The role of psychologists here is to act as a resource to whom PCPs can turn when emotional disorder is suspected to provide correct diagnoses and recommend appropriate treatments.

Derogatis and Dellapietra (1994) discussed screening for psychiatric disorders but did so largely from a perspective of the outpatient psychiatric clinic. They reviewed a variety of screening instruments, including the Symptom Checklist-90 Revised, the brief Psychiatric Rating Scale, the Center for Epidemiological Studies-Depression Scale (CES-D), the Self-Rating Depression Scale, and the Hamilton Anxiety and Depression scales.

The Beck Scales

The Beck Depression Inventory (BDI) (Beck & Steer, 1987) is one of the most widely used self-report instruments for the assessment of depression. Although its authors recommend the BDI as an instrument suitable for evaluating the level of depression as well as sensitive to changes in level, the BDI has great utility as a screening instrument. It is brief, consisting of 21 items that are endorsed at one of four levels. There is an even shorter form that consists of 13 items. There is considerable agreement between these two instruments (Reynolds & Gould, 1981), and the short form might be preferable when there are time constraints.

Beck and his associates developed other self-report instruments that have utility in a screening setting. The Beck Hopelessness Scale (Beck & Steer, 1988) taps the feelings of negative expectations about the future and global cognitions of despair. Although it has been found to be helpful in identifying suicidal risk in clinical populations, at least one study has questioned the utility of the instrument in a more general setting (Steed, 2001). The Beck Anxiety Inventory (Beck & Steer, 1990) has 21 items that are endorsed on a scale from 1 to 3. It correlates well with other instruments to measure anxiety and is fairly accurate in identifying DSM-III (Diagnostic and Statistical Manual of Mental Disorders, third edition) (American Psychiatric Association, 1988) anxiety diagnoses. The Beck Scale for Suicide Ideation (Beck & Steer, 1991) has 19 items that can be helpful in quickly obtaining information regarding the possibility of suicide-related thoughts but that is not particularly accurate in predicting actual suicide attempts.

An even shorter assessment instrument is the Center for Epidemiological Studies-Depression Scale, with only 10 items. The CES-D has been used in multiple settings, including Puerto Rican primary care patients (Robison, Gruman, Gaztambide, & Blank, 2002). It has generally been found to have adequate sensitivity and specificity.

Often, the decision to screen is made after some suspicion regarding a general probability that a disorder might be present. For example, in cases where individuals complain of cardiac symptoms and describe histories of going to emergency rooms because of fear that myocardial infarcts were occurring, the clinicians might suspect a panic disorder, and screening efforts would be directed at this construct using an instrument such as the Autonomic Nervous System Questionnaire, a self-report instrument that contains only five items but has been found to have good sensitivity but low specificity (Stein et al., 1999).

Psychological Adaptation to Illness

Yet another feature of psychological screening is particular to the medical setting. The construct of interest here can be broadly defined as psychological reaction to the medical condition. For example, individuals may vary in the degree to which they develop maladaptive behaviors in response to the medical condition. The Illness Behaviour Questionnaire is an example of an instrument that can evaluate these responses, defined as the inappropriate experience of the state of health (Pilowsky, 1994). Unfortunately, attempts to shorten the test (cf. Chaturvedi, Bhandari, Beena, & Rao, 1996) have not been successful (Bond & Clark, 2002). Another example of an instrument that is somewhat shorter is the Health Anxiety Questionnaire (Lucock & Morley, 1996).

The Illness Behaviour Questionnaire has also been used in patients with chronic pain (Pilowsky & Katsikitis, 1994). But there are also instruments that are more directly focused on pain behaviors and cognitions, including the Pain Anxiety Symptoms Scale, the Fear Avoidance Beliefs Questionnaire, and the Fear of Pain Questionnaire (McCracken, Gross, Aikens, & Carnike, 1996).

Screening for Substance Abuse Behaviors

Individuals who abuse alcohol and/or other substances are overrepresented in psychiatric settings and in general medical settings. A large number of health problems can arise from alcohol abuse, including disorders of the pancreas, stomach, liver, and intestines as well as hypertension. Obviously, a psychologist working in a medical setting will be exposed to alcoholism continuously. There is a need to screen for substance abuse behaviors on a regular basis.

One of the simplest measures that is useful in screening for alcohol abuse is known as the “CAGE” (cut, annoy, guilty, eye) questions. These questions are as follows:

  • Have you ever tried to cut down on your alcohol consumption?
  • Does it make you annoyed when people discuss your alcohol use?
  • Have you ever felt guilty about your alcohol use?
  • Did you ever need a drink as an eye opener?

As simple as this assessment method is, it is also fairly effective at identifying people who are likely to have been abusing alcohol (Bradley, Boyd-Wickizer, Powell, & Burman, 1998; Ewing, Bradley, & Burman, 1998; Nadeau, Guyon, & Bourgault, 1998). It has all the prerequisites of a screening procedure; it is short, simple, and easy to use and score. In addition, it is not generally intrusive and is acceptable to patients. It also possesses the shortcoming common to screening procedures, namely fairly high false positives.

Other procedures for screening for substance abuse behaviors include the Maryland Addiction Questionnaire (O’Donnell, DeSoto, & DeSoto, 1997) and the Michigan Alcoholism Screening Test (MAST) (Seltzer, 1971).

Report Writing

Report writing for psychological consultation in a medical setting follows the same format of the consult notes written by medical practitioners. These reports are brief, concise, and pointed toward answering the referral question. Unfortunately, sometimes the referral question is not well articulated or thought out. In those instances, it would be helpful to have a brief conversation with the referral source to clarify the information needed. This conversation can serve to elucidate the current concerns as well as to educate the referral source as to future consultation requests. Psychologists sometimes have a tendency to “show and tell” all that went into an assessment. This is good and well when the recipient of the report is the patient’s psychotherapist or when the psychologist’s assessment skills are being assessed, but in a medical setting it is the bottom line that is most important. That is not to say that fine points or subtlety should be ignored. However, the concise report is the report that gets read and used.

Specialized Training and Skills Needed

The psychologist who desires to work in a medical setting should first receive supervised clinical training in that setting. Any good clinical psychology graduate training program will provide training in psychological assessment, but much of graduate training in assessment is conducted in an environment very different from the typical medical setting. Graduate psychology training in assessment typically takes place in the university clinic, where entire days might be spent administering psychological tests. Reports are written over the course of several weeks and are lengthy treatises.

In contrast, psychological assessment and screening in a general health care setting takes place at the bedside in a hospital or in an examining room in the outpatient clinic of the medical service. The psychologist is not on his or her home turf and must be prepared to conduct the assessment with whatever materials have been brought to the appointment. The report may include a dictated note that can be up to one page long, but an initial note should be on the chart or given to the referral agent immediately on finishing the administration of assessment instruments. For these reasons, traditional graduate training in assessment is necessary but insufficient. There must also be training in the context of a general health care setting. Interpretation needs to be quick, and the psychologist is often called on to think on his or her feet.

The psychologist clinician should have familiarity with a range of medical disorders and know their basic pathophysiology, etiology, and treatment. For example, it would be important for the psychologist to know that hypertension can be associated with mildly impaired attention and memory and that some antihypertensive medications can cause side effects that mimic depression. In particular, the psychologist should be familiar with the types of patients and disorders seen in that clinic. Cultural sensitivity is a must. If the psychologist is not totally familiar with the particulars of that clinic, he or she should seek out learning experiences. The physician is a good resource for medical information regarding the disorders, and the nursing and support staff are good sources of information regarding the patients and their subculture.

Optimal Characteristics of a Screening Procedure

It would be useful to briefly reiterate the characteristics of a good screening test. It should be accurate. It should be sensitive to the construct under consideration. It should possess at least moderate specificity. It should have moderate positive predictive power and high negative predictive power. It could be administered by paraprofessional staff or self-administered, thereby meeting the final characteristic of using a minimal amount of professional time.