Emergency Psychiatry: Rapid Screening for Cognitive Impairment in the Psychiatric Emergency Service: II. A Flexible Test Strategy

The Trail-Making Test (TMT) is a useful cognitive screen because it is accessible, easy to use, and modifiable. The TMT is given in two parts (A and B). Developed by the U.S. Army, the test is considered to be within the public domain and thus may be reproduced without permission. Part A requires the patient to draw lines connecting consecutively numbered circles on a worksheet. Part B tests cognitive flexibility skills and requires patients to shift sets by drawing lines connecting numbered to lettered circles—from 1, to A, to 2, to B, and so on. This format has been modified for use with non-English-speaking populations by using two colors—pink and yellow—rather than letters in the trails (6). In the color version of the test, part A is essentially the same as in the English-language version, but part B requires the patient to match numbers with colors in a progressive and alternating pattern—1 with yellow, 1 with pink, 2 with yellow, 2 with pink, and so on. Cross-cultural studies have shown that this color trails test is very effective in discriminating HIV-positive and -negative persons (6). However, the test's ability to differentiate between HIV-related dementia and other types of dementia has not been investigated.

The TMT and the visual reproduction subtest of the Wechsler Memory Scale (WMS) have been used to evaluate various groups of emergency psychiatric patients. The visual reproduction subtest of the WMS requires the patient to copy four designs after a brief presentation of ten seconds' duration and then approximately 30 minutes later. Use of the TMT and the WMS in combination was found to successfully discriminate among patients with schizophrenia, mood disorder, or adjustment disorder (4). The screen detected greater cognitive impairment among patients with schizophrenia and mood disorder than among patients with adjustment disorder and than among control patients. Moreover, performance on this screen was associated with likelihood of admission to the hospital. Patients who were admitted to the inpatient service demonstrated greater cognitive impairment on the cognitive screen than did patients who were treated and released from the psychiatric emergency service (4).

The California Verbal Learning Test (CVLT) (7) has also been used successfully in the psychiatric emergency service to discriminate among various groups of psychiatric patients. The CVLT is a memory test in which 16 words are presented to the participant aurally. The participant is then instructed to recall the list of words. The list is repeated five times, providing a measure of learning ability. After the fifth trial, 16 distracter words are presented for recall. After short- and long-delay conditions (the long delay is 20 minutes), the participant is instructed to recall the original words. Recognition performance is also measured. A recent study found that CVLT performance profiles discriminated among patients in the psychiatric emergency service who abused cocaine, those with schizophrenia, and those with schizophrenia who abused cocaine (8).

Moreover, the pattern of each group's CVLT-identified performance deficits was better at discriminating between diagnostic groups than were clinical-based assessment ratings of patients' presenting symptoms. The patients with schizophrenia, for example, who abused cocaine had a CVLT profile that was marked by significant forgetting of previously acquired CVLT information compared with the other groups, who retained acquired information (8).

The CVLT may also be useful in the psychiatric emergency service because of its sensitivity in detecting milder forms of dementia. Unlike the Mini-Mental State Examination (MMSE), the CVLT has been shown in laboratory settings to be sensitive in distinguishing cortical from subcortical dementia, including dementia associated with Huntington's disease, Parkinson's disease, progressive supranuclear palsy, Wilson's disease, Pick's disease, multiple sclerosis, and HIV infection. Although cognitive dysfunction among patients with various types of these subcortical dementias varies (9), reliable differential diagnoses of dementing illnesses can be performed (10). For example, one of the primary functional distinctions between cortical and subcortical dementia is that cortical dementia typically involves deficits in verbal and semantic knowledge (11). By contrast, patients with subcortical dementia are found to have general retrieval deficits but no significant language impairment.

On the CVLT, for example, patients with Alzheimer's disease do not show normal improvement in performance when recognition memory rather than recall memory is tested (12). By contrast, patients with Huntington's disease identify more target items during recognition testing than during free-recall trials. Patients with Alzheimer's disease also demonstrate significantly more intrusion errors than do patients with Huntington's disease. Symptomatic HIV-positive patients and patients with Huntington's disease or Parkinson's disease have demonstrated similar performance deficits on the CVLT. For example, these groups demonstrate impaired recall over the five learning trials and show similar patterns of increased intrusion errors during recall performance (13).

Given that Alzheimer's disease is the most common cause of dementia (14), efforts to develop cognitive screening tests have focused on this patient population. Investigations of the use of cognitive screens has led to important contributions in the early detection of Alzheimer's disease, tracking of dementia progression, and assessment of the efficacy of potential drug therapies (14). In particular, the sensitivity of the clock-drawing test (CDT) among patients with Alzheimer's disease supports the wide use of the test as an index of the cognitive decline associated with dementia (15). However, the CDT is also sensitive to a variety of other psychiatric and neurologic impairments (15,16,17,18). In terms of Alzheimer's-related impairment, research using the CDT has shown that patients with Alzheimer's disease demonstrate deficiencies in their attempts to draw a clock face compared with other patients with dementia. Wolf-Klein and colleagues (15), for example, found that 94 percent of patients with probable Alzheimer's disease were impaired in global performance (a score of less than 6 on a scale of 1 to 10, with lower scores indicating more severe impairment), compared with only 75 percent of patients with dementia of nonspecific origin.

An analysis of patients' errors on the CDT provides an additional means of comparison. Whereas a majority of errors made by patients with Alzheimer's disease involve conceptual deficits—difficulty in the production of a clock face—patients with various subcortical dementias, vascular dementias, or Huntington's disease make relatively more spatial or planning deficits, such as leaving large gaps between some numbers on the clock face while crowding other numbers together, or other types of disorganization during reproduction of the clock face (18).

These planning problems are not weighted heavily in the global performance measure and likely account for the lower proportion of impairment seen in the non-Alzheimer's groups. Furthermore, the conceptual errors demonstrated by patients with Alzheimer's disease occur early in the course of their illness and appear to increase over time, resulting in a steady decline in global performance measures on the CDT (15,17). Overall, the CDT provides useful information in discriminating between functional psychoses and dementia and in differentiating between a broad range of the dementing illnesses.

Alone or in combination, individual tests used in the psychiatric emergency service can provide clinicians with a flexible approach for screening for cognitive deficits. Although the potential benefits of implementing routine use of cognitive screens in the psychiatric emergency service are impressive, greater knowledge about the specific cognitive deficits associated with psychiatric disorders would further enhance the utility of these screening tests. For this reason, additional research is needed in the psychiatric emergency service to assess the ability of experimental cognitive screens to differentiate between various clinical populations presenting for treatment. The cognitive screening instruments we highlight here may eventually come to be routine supplements to the MMSE during emergency psychiatric assessment.

Dr. Copersino is affiliated with the department of psychiatry at the Johns Hopkins University School of Medicine in Baltimore. Dr. Serper is affiliated with the department of psychiatry at New York University School of Medicine in New York City and the department of psychology at Hofstra University in Hempstead, New York. Dr. Allen is affiliated with Alcohol, Drug, and Psychiatry Services of Denver Health Medical Center. Send correspondence to Dr. Serper, Department of Psychology, 222 Hauser Hall, Hofstra University, Hempstead, New York 11549-1350 (e-mail, [email protected] or [email protected]). Douglas H. Hughes, M.D., is editor of this column.