RECENT STUDIES have shown that mental health problems

Transcription

1 405 More Than Meets the Eye: How Examiner Training Affects the Reliability of the MacNeill Lichtenberg Decision Tree in Geriatric Rehabilitation Patients Adam L. Bank, MA, Susan E. MacNeill, PhD, Elizabeth M. Hall, MD, Richard K. Nadjarian, MD, Ann V. Zaccagnini, DO, Peter A. Lichtenberg, PhD ABSTRACT. Bank AL, MacNeill SE, Hall EM, Nadjarian RK, Zaccagnini AV, Lichtenberg PA. More than meets the eye: how examiner training affects the reliability of the MacNeill Lichtenberg decision tree in geriatric rehabilitation patients. Arch Phys Med Rehabil 2002;83: Objective: To examine the effect of examiner training on the test-retest reliability of the MacNeill Lichtenberg Decision Tree (MLDT), a tool for guiding clinicians decision making for referrals for mental health problems. Design: Correlational analyses and chi-squares were used to examine the influence of demographic variables on MLDT performance and the test-retest reliability of its cognitive and affective components. Setting: Rehabilitation unit of a large, freestanding, urban hospital. Participants: In study 1, 39 older, medical rehabilitation patients consecutively referred to the neuropsychology service. In study 2, 57 older, consecutively admitted medical rehabilitation patients. Interventions: In study 1, patients underwent testing with the MLDT by a novice examiner. In study 2, patients were tested by trained examiners. Both sets of results were compared with those obtained by experienced examiners. Main Outcome Measures: The temporal stability of participants performance on the cognitive and affective components of the MLDT was compared between the 2 studies. Results: Training was associated with high test-retest reliability on both the cognitive and affective components. A lack of training was associated with reduced reliability in depression screening. Conclusion: These findings support the use of the MLDT as a mental health triage tool for older adults in inpatient medical settings. Its use is dependent on training and accurate administration. Key Words: Cognition; Rehabilitation; Reproducibility of results; Triage by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation From the Department of Psychology (Bank, Lichtenberg), Institute of Gerontology (Bank, Lichtenberg), Rehabilitation Institute of Michigan (Zaccagnini, Lichtenberg), Wayne State University; Department of Physical Medicine and Rehabilitation, Wayne State University School of Medicine (Hall), Detroit, MI; Division of Neuropsychology, Henry Ford Health System (MacNeill), Detroit, MI; and Department of Rehabilitation Medicine, Sinai-Grace Hospital (Hall), Detroit, MI. Accepted in revised form April 18, No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors(s) or upon any organization with which the author(s) is/are associated. Reprint requests to Adam L. Bank, MA, Institute of Gerontology, Wayne State University, 87 E Ferry, 226 Knapp Bldg, Detroit, MI 48202, ALBANK2000@hotmail.com /02/ $35.00/0 doi: /apmr RECENT STUDIES have shown that mental health problems are strongly associated with poor physical and psychosocial outcomes in geriatric rehabilitation patients. 1-3 It is, therefore, important for health care professionals to detect and treat mental illness to reduce suffering, disability, and health care costs. Several screening instruments have been developed to easily and quickly identify psychopathology in older hospitalized patients; they include the Mini-Mental State Examination 4 (MMSE), the Short Portable Mental Status Questionnaire, 5 and the 7-Minute Screen. 6 However, each of these instruments assesses only personal competencies and lacks any assessment of environmental demand needed to help a clinician decide whether additional evaluation is appropriate. To address this dilemma, MacNeill and Lichtenberg 7 developed a triaging tool to screen for mental health problems in geriatric rehabilitation patients and to guide decision making for referrals. The MacNeill Lichtenberg Decision Tree (MLDT) includes a cognitive screen, an affective screen, and items that evaluate psychosocial demands. The cognitive screen consists of Benton s Temporal Orientation Test 8 (BTOT) and a measure of semantic verbal fluency (Animal Naming Test). 9 The affective screen consists of 3 items from the 30-item Geriatric Depression Scale 10 (GDS). The psychosocial demand items target whether patients live alone, whether they will be doing basic and advanced activities of daily living without supervision, and whether they or their caregivers are aware of any cognitive deficits. The psychosocial items permit the examiner to triage patients more accurately and appropriately. Patients who fail the cognitive screen may or may not benefit from further neuropsychologic evaluation, depending on their psychosocial demands. Patients who pass the cognitive screen are not administered the psychosocial items and are deferred from further cognitive assessment because of the low probability (described later) that they will show a degree of cognitive impairment that would impact key psychosocial domains (eg, ability to live alone, ability to perform self-care activities without supervision). Three studies to date have found evidence of the validity and clinical utility of the MLDT. 7,11,12 In those studies, the MLDT had high sensitivity and specificity in targeting patients with a higher probability of cognitive impairment and in effectively identifying cognitively intact patients. The MLDT also had an excellent capacity for ruling out patients who lack significant levels of depressive symptoms. Equally important, it was effective in clinical practice because it eliminated a substantial number of unnecessary referrals and reduced clinical assessment time by 44%. 12 In noting the substantial increase in the number of older adults undergoing medical rehabilitation, MacNeill and Lichtenberg 7 argued for broad implementation of the MLDT by frontline health care staff, such as physicians and nurses, rather than sole use of the instrument by rehabilitation psychologists. Such a diffuse use would likely provide a more far-reaching effect on appropriately triaging geriatric medical patients.

2 406 EXAMINER TRAINING, Bank However, no studies to date have examined the level of training necessary for physicians, nurses, or allied health practitioners to use the MLDT in a reliable manner. The MLDT appears at first glance to be a straightforward instrument to administer and score, such that training in its use seems superfluous. However, a comprehensive review of research on another brief screening instrument suggests otherwise: variations and inconsistencies in the administration, scoring, and interpretation of the MMSE, the most widely used cognitive screening instrument, were quite prevalent. 13 No study to date has systematically examined whether such inconsistencies impact the reliability of a given screening instrument. This question is an important one because the reliability of any standardized test or screen places an upper limit on the instrument s validity and clinical utility. In this study, we investigated the effects of training on the test-retest reliability of the cognitive and affective components of the MLDT. We hypothesized that the MLDT would show higher levels of test-retest reliability when administered by trained examiners and that its reliability would be substantially compromised when administered and scored by a novice examiner. In study 1, we examined the MLDT s reliability when a novice (ie, untrained) examiner performed the test administration and scoring. In study 2, we tested the MLDT s reliability when trained examiners administered and scored the instrument. METHOD Study 1: Novice Examiner In study 1, a novice examiner (a 4-y college graduate) was presented with the MLDT instrument, book chapter describing the components of the MLDT, as well as the rationale behind it. 14 The novice examiner was not supervised during the administration of the MLDT, nor was he given feedback regarding his performance in administering the instrument. Participants Participants were consecutive referrals to the neuropsychology service of a rehabilitation unit of a large, freestanding, urban hospital. The mean age standard deviation of the sample was years, and the mean level of education was years. Sixty percent of the participants were women, and 69% were black. Twenty-two percent of participants were widowed; 46% were married or living together; 22% were separated, estranged, or divorced; and 11% had never been married. Reasons for admission to inpatient rehabilitation services included debility from multiple medical illnesses (29%), orthopedic and cardiovascular disorders (38%), and cerebrovascular accident (33%). The median length of hospitalization before admission to the rehabilitation unit of the hospital was 7 days (mean, d). Procedure Between December 1998 and April 1999, all English-speaking patients aged 60 years and older who were referred to the neuropsychology service were asked by the attending neuropsychologist (SEM) to participate in the study. None of the patients approached refused to participate, and their informed consent was obtained. The neuropsychologist, an experienced MLDT examiner, administered the initial MLDT to all but 1 patient (2% of the sample), who was tested first by the novice examiner. Each patient was then retested by the novice examiner. Sixty-three consecutively admitted patients met the inclusion criteria (ie, age 60y and English speaking), and testretest data were available for 39 of these patients. Twenty-four patients were excluded for the following reasons: 12 patients were unavailable for testing by the novice because of participation in various rehabilitation therapies or temporary absence from the unit for medical tests; 7 were discharged before the novice could administer the MLDT; for 3 patients, the novice did not administer and/or record scores on the MDLT s cognitive portion; and the novice tested 2 patients who were not tested by the neuropsychologist. Seventy-four percent of the sample was retested within 6 days, and the median interval between the initial and second tests was 3 days (mean, d). Instruments MLDT measures. The MLDT is an examiner-administered instrument that includes a cognitive component, an affective component, and questions targeting psychosocial demands. Its cognitive portion is comprised of the BTOT and the Animal Naming Test. The BTOT provides a quantified measure of a patient s orientation to date, day of the week, and time of day. The Animal Naming Test is a semantic verbal fluency task. Failure on the cognitive portion of the MLDT is defined as either a score greater than 3 on the BTOT or a score less than 10 on the Animal Naming Test. Patients who fail the cognitive portion are evaluated by the examiner on 3 psychosocial items: (1) Does the patient live alone?, (2) Does the patient complete self-care and instrumental activities without supervision?, and (3) Is the patient or the patient s family aware of cognitive deficits? Psychosocial needs are indicated by an affirmative response to either of the first 2 questions or by a negative response to the third question. Further neuropsychologic testing is indicated when the patient fails at least 1 cognitive measure and shows 1 or more psychosocial needs. The GDS-3 screen for clinically significant dysphoric symptomatology consists of 3 items from the GDS-30; it is administered irrespective of a patient s performance on the cognitive screen. Patients who positively endorse at least 1 of the 3 items are considered to have failed the affective screen, and they are referred for a thorough evaluation of depressive disorders (regardless of their responses to the psychosocial questions), including administration of the GDS-30 and a clinical interview. Study 2: Trained Examiners In study 2, 3 physicians (an attending physical medicine and rehabilitation physician, 2 residents under her supervision) underwent brief training on the administration and scoring of the MLDT by the attending neuropsychologist (SEM) and an advanced clinical psychology doctoral student (ALB), both of whom were experienced MLDT examiners. After reviewing an instruction manual detailing the administration and scoring of the instrument (see appendix 1), the physicians participated in a 15-minute training session given by one of the experienced examiners. Their first 2 MLDT administrations were observed by either the neuropsychologist or doctoral student to identify and correct administration and scoring errors. The attendingphysicians and the residents were also observed periodically during the remainder of the study to ensure accurate administration and scoring. Participants Participants were consecutive referrals to an attending physician s service on the rehabilitation unit of a large, freestanding, urban hospital. The mean age of the sample was years, and the mean level of education was years.

3 EXAMINER TRAINING, Bank 407 Table 1: Relationship Between Demographic Variables and Performance on MLDT Components (Study 1: Novice Examiner) Examiner Test Component Variable N 2 r df P Expert (SEM) Cognition Age Education Gender Race Depression Age Education Gender Race Novice Cognition Age Education Gender Race Depression Age Education Gender Race Sixty-seven percent of the participants were women, and 81% were black. Thirty-nine percent of participants were widowed; 35% were married or living together; 16% were separated, estranged, or divorced; and 11% had never been married (percentages do not add to 100 because of rounding). Reasons for admission to inpatient rehabilitation services included debility caused by multiple medical illnesses (37%), orthopedic fracture (30%), cerebrovascular accident (28%), and traumatic brain injury (4%). The median length of hospitalization before admission to the hospital s rehabilitation unit was 6 days (mean, d). Procedure Between March 2000 and July 2000, all patients who were admitted to the rehabilitation unit, were 60 years of age and older, and were English-speakers were administered the MLDT during the history and physical portion of the admission process. The initial administration of the MLDT, which was part of the usual care provided, was conducted by 1 of the 3 physicians. After this initial administration, each patient was asked by either the neuropsychologist or doctoral student to participate, and informed consent was obtained. None of the patients approached refused to participate. Fifty-eight consecutively admitted patients met the inclusion criteria (ie, age 60y, English speaking) and were administered a second MLDT by one of the experienced examiners. One patient was excluded because of an inability to communicate verbally. The final sample included 57 participants. Seventy-five percent of the sample was retested within or at 72 hours, and the median interval between the initial and second administrations was hours (mean, h). Instruments The same instruments were used in study 2 as were used in study 1. RESULTS Study 1: Novice Examiner The data were analyzed by using correlational analyses and chi-squares, as appropriate. First, we examined whether demographic variables influenced performance on the MLDT components. Alpha levels were set at.01 to reduce the likelihood of type I error. Consistent with prior studies, demographic variables were not correlated with performance (pass/fail) on the cognitive or affective screens across both test administrations, as shown in table 1. Table 2 displays the test-retest stability for the cognitive and affective components of the MLDT when a novice examiner administered the instrument. Overall, 44% of the sample passed the cognitive screen at both the initial and second administrations, whereas 38% failed the cognitive screen at both testings. Among the 20 patients who passed the initial cognitive screen, 17 (85%) passed the cognitive screen on retest. Among the 19 patients who failed the cognitive screen at initial administration, 15 (79%) failed the cognitive screen on retest. Twenty-nine percent of the sample passed the depression screen at both test administrations, and 25% failed the depression screen at both times. Among the 9 participants who passed the initial depression screen, 7 (78%) passed the depression screen on retest. Among the 15 participants who failed the depression screen initially, 6 (40%) failed the depression screen on retest. Several types of errors were observed in examining the protocols obtained by the novice examiner. On the BTOT, for example, error points were incorrectly scored and/or summed, correct answers were not routinely recorded next to the patient s response, and patient responses were sometimes not recorded. On the Animal Naming Test, credit was incorrectly given for naming the same animal twice, and credit was not given for some correct responses. More than two Table 2: Stability Over Time of MLDT (Study 1: Novice Examiner) Test Component % Passed* % Failed 2 df P Cognition screen (N 39) Depression screen (n 24) NS Abbreviation: NS, not significant. * Percentage of patients who passed the screen on retest after passing the first screen. Percentage of patients who failed the screen on retest after failing the first screen.

4 408 EXAMINER TRAINING, Bank Table 3: Relationship Between Demographic Variables and Performance on MLDT Components (Study 2: Trained Examiners) Examiner Test Component Variable N 2 r df P Experts (SEM, ALB) Cognition Age Education Gender Race Depression Age Education Gender Race Trained examiners Cognition Age Education Gender Race Depression Age Education Gender Race fifths (44%) of the patients enrolled in study 1 were not administered the depression screen by the novice examiner, in large part because patients who passed in the cognitive screen were often (erroneously) skipped out of the depression component. Last, the psychosocial indicators portion was sometimes administered by the novice examiner despite the patient passing the cognitive screen. Study 2: Trained Examiner The data were analyzed in the same manner as in study 1. The influence of demographic variables on MLDT performance was examined, with levels set at.01 to reduce the likelihood of type I error. Consistent with prior studies, as well as with study 1, demographic variables were not correlated with performance (pass/fail) on the cognitive or affective screens across both test administrations, as shown in table 3. Table 4 displays the test-retest stability for the cognitive and affective components of the MLDT when administered by trained examiners. Overall, 40% of the sample passed the cognitive screen at both the initial and second administrations, whereas 42% failed the cognitive screen at both testings. Among the 25 patients who passed the initial cognitive screen, 23 (92%) passed the cognitive screen on retest. Among the 32 patients who failed the cognitive screen at initial administration, 24 (75%) failed the cognitive screen on retest. Thirty-two percent of the sample passed the depression screen at both test administrations, and 46% failed the depression screen at both times. Among the 24 participants who passed the initial depression screen, 18 (75%) passed the depression screen on retest. Among the 32 participants who failed the depression screen initially, 26 (81%) failed the depression screen on retest. Several common errors were observed in the physicians administration and scoring of the MLDT during the initial portion of their training. On the BTOT, correct answers were not recorded next to patient responses, and patients were not appropriately cued when they gave vague or I don t know responses. On the Animal Naming Test, errors included a lack of careful timing for exactly 1 minute, failure to encourage patients to put forth additional effort when they paused for more than 10 seconds or gave up, and erroneously giving credit for naming different genders of the same animal. On the GDS-3, vague responses were not queried, discomfort was displayed when asking items related to emotions, and the screen was not administered to every patient. On the psychosocial indicators section, some items were not administered or appropriately cued. Compared with the novice examiner, the trained examiners committed fewer total administration and scoring errors, despite administering more test protocols, as shown in table 5. DISCUSSION The effect of examiner training on the test-retest reliability of the MLDT was evaluated in 2 samples of geriatric rehabilitation patients by comparing the results obtained by a novice examiner (study 1) with those obtained by trained examiners (study 2). We found that training had little effect on the temporal stability of patients performances on the cognitive component of the MLDT; moderate to high levels of stability were found in both the untrained and trained examiner conditions. On the other hand, test-retest reliability on the depression component was substantially compromised when the novice examiner administered and scored the MLDT. More specifically, in the trained examiner condition, we found that 81% of patients who failed the initial depression screen also failed the screen on retest. In contrast, in the novice examiner condition, only 40% of the patients who failed the initial depression screen failed the depression screen on retest. Table 4: Stability Over Time of MLDT (Study 2: Trained Examiners) Test Component % Passed* % Failed 2 df P Cognition screen (N 57) Depression screen (n 56) * Percentage of patients who passed the screen on retest after passing the first screen. Percentage of patients who failed the screen on retest after failing the first screen.

5 EXAMINER TRAINING, Bank 409 Table 5. Frequency of Types of Errors Committed by Novice and Trained Examiners Test Component Novice Examiner* Trained Examiners n % n % BTOT Administration error Scoring or summing error Failure to query Animal Naming Test Timing error NA No encouragement when appropriate NA Crediting same animal Scoring or summing error GDS-3 Failure to query Discomfort in administration NA Failure to administer all items to each patient Scoring error Psychosocial indicators Failure to administer all items when cognitive screen is failed Inappropriately administering items to patients who passed the cognitive screen Failure to query NA Total No. of Errors * Percentage based on 39 protocols administered by the novice examiner. Percentage based on 57protocols administered by the trained examiners. Error categories marked NA for the novice indicate that no data was available because the novice was not observed during test administration. In addition, the total number of errors in these categories for the trained examiners represents only those few trials in which the physicians were initially observed and periodically supervised. Because the physicians were not observed during the majority of the test administrations, these figures should be interpreted as estimates of the actual number of errors committed. In addition, examination of the test protocols administered by the novice examiner revealed substantially greater numbers of administration and scoring errors (86) compared with those of the trained examiners (32); the frequency and types of errors committed by the novice and the trained examiners are presented in table 5. The implications of these errors should not be taken lightly, given their potential impact on clinicians treatment decisions. To illustrate this point by using the literature on the broader assessment of cognitive functioning, 1 study 15 found that a group of licensed practitioners made numerous mistakes on the Wechsler Intelligence Scale for Children Revised, to the extent that the changes in the full scale IQ scores could affect labeling and placement decisions. With respect to this study, errors made by the novice would have resulted in 5 inappropriate neuropsychology referrals, 1 inappropriate depression referral, 2 false negatives (1 on the cognitive screen, 1 on the depression screen), and 19 patients unscreened for depression. Errors made by the physicians during their training period would have resulted in 3 inappropriate neuropsychology referrals and 1 false negative on the depression screen. To eliminate these errors and to aid clinicians in appropriately administering, scoring, and using the MLDT, we have included an updated training manual (Appendix 1). We updated the training manual based on our findings from this study; we added examples of how to make scoring decisions, when and how to cue patients, and how to elicit the best performance from each patient. CONCLUSION We found that training individuals in the use of the MLDT was easily accomplished with 1 training session and observation of initial test administrations. Training was essential for accurate, standardized administration, and scoring. Standard, brief training of MLDT examiners was associated with adequate levels of temporal stability in geriatric rehabilitation patients performances on the cognitive and affective screens, but an absence of training led to a large decrement in reliability of the MLDT depression screen, as well as numerous administration and scoring errors. Thus, training and accurate administration and scoring are essential to ensuring the reliability and use of the MLDT s results. Health care professionals who intend to use the MLDT should first carefully review the training manual and undergo a 15-minute training session. They should then have their test administration and scoring observed on at least the first 2 occasions by an experienced examiner or until competency has been shown. In addition, to reduce the likelihood of examiner drift and to maintain the quality of training, MLDT examiners should periodically review the training manual and have another trained examiner observe their test administration and scoring. Two limitations to this study should be noted. First, our focus on older, inpatient rehabilitation patients limits the generalizability of our findings to other groups of older adults. Second, the relatively small sample size included in the novice examiner condition, particularly with respect to the depression component of the MLDT, limits the stability of our finding of low test-retest reliability in study 1. However, the small sample size was a direct result of the study s design; the novice examiner s unfamiliarity with appropriate administration of the entire MLDT led him to discontinue administration of the instrument, primarily among patients who had passed the cognitive portion. Future studies should examine the generalizability of the MLDT to other settings in which older, medically ill adults are often seen, such as primary care facilities and nursing homes. The validity, reliability, and clinical utility of the MLDT should also be examined in these varied settings. Future research should evaluate the reliability of the instrument when it is used by health care providers other than those included in

6 410 EXAMINER TRAINING, Bank this study, such as nurses, occupational therapists, and speech language pathologists. APPENDIX 1: INSTRUCTIONS FOR THE ADMINISTRATION OF THE MACNEILL LICHTENBERG DECISION TREE Background The MacNeill Lichtenberg Decision Tree (MLDT) is an instrument that was developed to quickly and accurately identify older medical patients with both high and low probabilities of cognitive impairment and depressive disorders and to aid clinicians in making more appropriate referrals for mental health services. The MLDT has been well validated for use with older, lower educated, medically ill adults. The instrument is targeted at medical patients aged 60 years and older, and it is intended to be administered by a range of health care professionals, including physicians, nurses, and allied health practitioners, following appropriate training. The MLDT is a triaging tool, and as such, it should be administered as part of each patient s initial evaluation (eg, as part of the History and Physical Examination) to guide clinicians referral decisions. Instructions General Administration: All patients receive parts A (Cognitive Indicators: Benton Temporal Orientation Test [BTOT], Animal Naming Test) and C (Emotional Factors). If the patient fails either portion of Part A, then Part B (Psychosocial Indicators) is given. For all items administered, it is important for the patient to understand the task in order to elicit his/her best performance. If the patient does not understand each task after the examiner has explained it, rephrase the directions. For those patients who do not complete a task, document why this happened (eg, they refused 1 particular question, they did not understand the idea behind naming different animals despite redirection, etc). Ideally, the patient should be tested alone, in a quiet room that is free of distractions. However, this is sometimes not possible in a busy hospital setting. When family members or friends of the patient insist on remaining in the room when the patient is screened, the examiner should explain that he/she will be asking the patient several questions to evaluate their thinking skills and their mood, and that for the test to be accurate, it is important that they do not distract the patient or answer for him/her. Materials: Testing form, pen, and stopwatch or wristwatch with a second hand. Test Instructions: Cognitive Indicators: BTOT and Animal Naming Test. Benton Temporal Orientation Test* Ask the patient not to look around for a calendar or a clock. Begin by saying, Without looking around the room, tell me what today s date is. If the patient does not respond, or does not understand the question, say, Ms. X, what MONTH is it? If the patient is still unable to respond, cue further: Is it January, or February, March, April, May, June, July,... Write down the patient s responses verbatim in the appropriate Response column; in the Correct column, you should write in the full correct answer (eg, August 29, 2000, Tuesday, 3:35 PM) so that the scoring can be rechecked at a later date. Then ask, What is the date? and follow the protocol outlined above for the date, year, day of the week, and time. Thus, if the patient does not respond to the initial question, the examiner should emphasize the key word in the question. If the patient is still unable to respond, the examiner should cue further by giving examples of the category in question. Patients who are unable to respond to the item, What time is it right now? should be encouraged to guess even if they are not sure. Scoring: Score according to the directions on the testing form. Each item is scored by error points, as specified on the testing form. Self-corrections by the patient receive full credit. Animal Naming Test The examiner should have a stopwatch (or a wristwatch with a second hand), and should be ready to begin timing before giving the patient the instructions. Speak slowly and loudly, and say, Ms. X, I have some things I would like you to do. The important thing is for you to try to do your best. I d like you to tell me as many different animals as you can. I will give you 1 minute to tell me as many as you can. Are you ready? Go! If the patient does not understand what she is supposed to do, explain the directions again. Begin timing as soon as you say go, and record the patient s responses verbatim. If the patient speaks very rapidly, write down the first 2 or 3 letters of the animal name until you can catch up, then fill in the entire animal name later. Stop timing at exactly 1 minute. Do NOT give the patient credit for any responses given after the 1 minute has expired. If the patient gives a few animal names and then stops for more than 10 seconds, encourage the patient to keep trying, and provide praise for effort. For example, calmly say, Okay, you re doing great, tell me a few more animals. Or, Good, now tell me some more animals. Keep going, you re doing fine. Remember, the idea is to obtain the patient s best performance within the 1 minute. Patients receive credit for every different animal they name. Credit is not given for naming the same animal at different developmental stages (eg, dog-puppy, cat-kitten), nor is credit given for a general category of animal followed by an example of the category (eg, bird-eagle). Birds, apes, dogs, and fish are often given as broad categories; when a patient gives these as responses, do not give her credit for the broad category if she then goes on to give examples of the category. If a patient says, cat, dog, elephant, lion, tiger, bear, you should give her credit for cat because it likely refers to a housecat, which is distinct from a wild cat. Pig and hog are counted as 2 different animals, but a sow (female pig) and a pig are scored as 1. Insects and worms are also animals. Goat and sheep are different animals. An ewe is a female sheep, a ram is a male sheep, and a lamb is a young sheep. Thus, a patient who says, sheep, ewe, ram, lamb would get credit for 1 animal, not 4. Types of apes include: orangutan, gorilla, chimp, gibbon, baboon. Types of monkeys include: howler monkey, wooly monkey, spider monkey, owl monkey. Below are some examples of correct scoring, with x s by answers that are not given credit. Patient 1: fish (x), goldfish, shark, salmon, minnow, bass, catfish 6 Patient 2: bird (x), rooster (x), chicken, parakeet, eagle, snake (x), cobra 5 Patient 3: ape (x), monkey, chimp, gorilla, baboon, cow, bull (x) 5

7 EXAMINER TRAINING, Bank 411 Psychosocial Indicators The Psychosocial Indicators are given if the patient fails the BTOT or the Animal Naming Test. Thus, these items are not given if the patient passes both of the cognitive indicators. (1) Does the patient live alone? Cue for details: (a) Do you spend long periods of time all alone at home? (b) Does someone sleep over at night? (2) Does the patient complete self-care and advanced activities without supervision? Cue for details: (a) Does anyone help you with walking, dressing, getting in and out of the bathtub or shower, or getting on the toilet? (b) Who helps you out around the house? and then determine whether the patient has help with preparing meals, taking their medication, or handling their finances and bill payments. (3) Is the patient or the patient s family aware of cognitive deficits? Cue for details: (a) Do you have any problems with your thinking skills (eg, memory, concentration, understanding what others say)? (b) Is your family concerned about you being by yourself? Paying the bills? GDS-3 Begin by saying, I want to ask you some yes or no questions about how your mood has been during the past week. Do you feel that your life is empty? Do you often feel downhearted and blue? Do you feel pretty worthless the way you are now? (Some examiners may experience initial discomfort in inquiring about mood in such a direct fashion; however, we have found that asking these questions in a forthright, comfortable manner tends to result in patients answering in a similar fashion.) In general, the examiner should not change the wording of the questions. However, if the patient has significant difficulty responding to the question, the examiner should restate the question so that it closely approximates the original question. Example 1: Do you feel that your life is empty? Restatement: During the past week, have you been feeling like there is nothing in your life? Example 2: Do you often feel downhearted and blue? Restatement: During the past week, have you frequently felt sad and down? Example 3: Do you feel pretty worthless the way you are now? Restatement: During the past week, have you felt like you really don t have any value as a person the way you are right now? * The BTOT is included as part of the MLDT with permission of its author (Arthur Benton, personal communication, Oct 3, 2000). References 1. Diamond PT, Holroyd S, Macciocchi SN, Felsenthal G. Prevalence of depression and outcome on the geriatric rehabilitation unit. Am J Phys Med Rehabil 1995;74: MacNeill SE, Lichtenberg PA. Home alone: the role of cognition in return to independent living. Arch Phys Med Rehabil 1997;78: Nanna MJ, Lichtenberg PA, BudaAbela M, Barth JT. The role of cognition and depression in predicting functional outcome in geriatric medical rehabilitation patients. J Appl Gerontol 1997;16: Folstein MF, Folstein SE, McHugh, PR. Mini-mental state. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12: Pfeiffer E. A short portable mental status questionnaire for the assessment of organic brain deficit in elderly patients. J Am Geriatr Soc 1975;23: Solomon PR, Hirschoff A, Kelly B, et al. A 7 minute neurocognitive screening battery highly sensitive to Alzheimer s disease. Arch Neurol 1998;55: MacNeill SE, Lichtenberg PA. The MacNeill Lichtenberg Decision Tree: a unique method of triaging mental health problems in older medical rehabilitation patients. Arch Phys Med Rehabil 2000;81: Benton AL, des Hamsher K, Varney NR, Spreen O. Contributions to neuropsychological assessment. New York: Oxford Univ Pr; Spreen O, Strauss E. A compendium of neuropsychological tests: administration, norms, and commentary. New York: Oxford Univ Pr; Yesavage JA, Brink TL, Rose TL, et al. Development and validation of the Geriatric Depression Screening Scale: a preliminary report. J Psychiatr Res 1983;17: Bank AL, MacNeill SE, Lichtenberg PA. Cross validation of the MacNeill-Lichtenberg Decision Tree: triaging mental health problems in geriatric rehabilitation patients. Rehabil Psychol 2000;45: Lichtenberg PA, MacNeill SE. Prospective validity study of a triaging method for mental health problems: the MacNeill- Lichtenberg Decision Tree (MLDT). Clin Gerontol 1999;21: Tombaugh TN, McIntyre NJ. The Mini-Mental State Examination: a comprehensive review. J Am Geriatr Soc 1992;40: MacNeill SE, Lichtenberg PA. Screening instruments and brief batteries for assessment of dementia. In: Lichtenberg P, editor. Handbook of assessment in clinical gerontology. New York: Wiley; p Slate JR, Jones CH, Coulter C, Covert TL. Practitioners administration and scoring of the WISC-R: evidence that we do err. J Sch Psychol 1992;30:77-82.