The Columbia University Scale for Psychopathology in Alzheimer's Disease D. P. Devanand, MD; Lisa Miller; Marcus Richards, PhD; Karen Marder, MD, MPH; Karen Bell, MD; Richard Mayeux, MD; Yaakov Stern, PhD • The Columbia University Scale for Psychopathology in describe interrater reliability between a lay interviewer Alzheimer's disease is a new screening instrument devel¬ (L.M.) and a research psychiatrist (D.P.D.), divergent va¬ oped for use by clinicians or trained lay interviewers. Inter- lidity, prevalence of symptoms, and relations between rater reliability was established between a psychiatrist and scale items and severity of dementia as assessed by cog¬ a lay interviewer in 20 patients. In an independent sample nitive function and ability to perform activities of daily of 91 outpatients with very mild to moderate probable life. Alzheimer's disease, caregiver informants reported that de¬ pressed mood was common (46.2%) but rarely persistent SUBJECTS AND METHOD (2.2%), and that sleep disturbance occurred frequently Scale Development and Design (41.8%) but was never severe (0%). There were significant A few items were adapted from existing instruments2 and but weak associations between the presence of specific other items were developed de novo by the principal developer subtypes of delusions and severity of dementia. Although a (D.P.D.) and other research personnel (Y.S. and M.R.). An ear¬ lier of the scale was before the current variety of delusional symptoms were reported, they were version piloted version transient and often the truth if was finalized. The scale takes approximately 10 to 25 minutes to frequently patients accepted administer to an informant, and focuses on corrected the As a result, few met symptoms during by caregiver. patients the past month. There is a single explicit question for each item broad or narrow criteria used to define delu¬ operational to facilitate its use by lay interviewers. sions. Prior studies may have overestimated the prevalence Operational definitions for delusions, hallucinations, and of psychotic features in Alzheimer's disease by not employ¬ illusions were based on standard definitional criteria for these ing standard definitional criteria. The findings also indicate symptoms.8 The informant's report of the patient's behavioral that new methodology such as that employed in this instru¬ response, if any, is taken into account to score psychotic symp¬ ment needs to be evaluated more widely. toms. Psychotic symptoms in Alzheimer's disease are known to (Arch Neurol. 1992;49:371-376) be different from those seen in other psychotic disorders.2,9 More critically, delusions are frequently transient, and patients with AD will often accept the truth if corrected by the caregiver, even behavioral disturbances, and depressive though the same symptom may recur shortly thereafter as a re¬ Psychosis,symptoms occur frequently in patients with Alzhe¬ sult of amnesia. These features can make it difficult to ascertain the of a "true" delusion traditional definitional imer's disease are to and care¬ presence fulfilling (AD),1-2 distressing patients criteria.8 in the the initial for each are often treated with medica¬ Hence, CUSPAD, question givers,3,4 psychotropic of delusion is scored If a tions,2,5 and be associated with the likelihood of type dichotomously (present/absent). may delusional symptom is present, two additional questions are institutionalization.6 Recent instruments that have been asked. The first follow-up question inquires if the delusion is evaluate in AD on the developed to psychopathology rely persistent (more than three times per week) or transient, and the use of trained clinician interviewers.2,7 We developed the second question inquires if the patient accepts correction of the Columbia University Scale for Psychopathology in Alzhei¬ false belief from the caregiver. A "broad" definition of a delusion mer's Disease (CUSPAD), a short semi-structured instru¬ is met when the patient does not accept the caregiver's correc¬ ment that can be administered by a trained lay interviewer tion of the false belief/idea, while a "narrow" definition of a de¬ for use in cross-sectional and studies. lusion is met when the patient does not accept the caregiver's long-term follow-up correction of the false belief/idea and the is Particular was on the use of symptom persistent, emphasis placed specific ie, occurs more than three times week. to establish defini¬ per follow-up questions help operational Hallucinations, if are rated as or "clear." This tions for the of features in AD, since present, "vague" presence psychotic approach to rate hallucinations was a because this this has been in instruments of this We adopted priori lacking prior type. interview relies solely on the informant's report, and patients are known to manifest behaviors that are often difficult to interpret as hallucinations. Illusions are rated to halluci¬ for October 21, 1991. clearly similarly Accepted publication nations. From the Memory Disorders Clinic in the New York State Psychi¬ atric Institute, and the Center for Alzheimer's Disease Research In Wandering, verbal outbursts, violence, agitation, and sun- New York City (Drs Devanand, Richards, Marder, Bell, Mayeux, downing constitute the items for behavioral disturbance that are Stern, and Ms Miller) and the Departments of Psychiatry (Dr scored as present or absent. Items for depression include Devanand) and Neurology (Ms Miller and Drs Richards, Marder, depressed mood, difficulty sleeping and change in appetite, and Bell, Mayeux, and Stern), College of Physicians and Surgeons of are scored on a 5-point scale to ascertain frequency and severity Columbia University, New York, NY. of these symptoms. Since the interview was developed prima¬ Reprint requests to the New York State Psychiatric Institute, 722 rily as a screening instrument, the addition of more items was W 168th St, New York, NY 10032 (Dr Devanand). not entertained. Downloaded From: http://archneur.jamanetwork.com/pdfaccess.ashx?url=/data/journals/neur/16293/ by a Columbia University User on 06/13/2017 Table 1. —Interrater Reliability for Presence/Absence Table 2. —Demographic Characteristics of the Sample of Symptoms (91 Outpatients With AD)* Kappa Coefficient Demographics Mean SD Minimum Maximum I I Conjoint Independent Sex, % female 56 Symptom Category Interviews Interviews Age, y 70.8 8.3 49 86 Delusions Age at onset, y 67.2 8.6 46 84 Present/absent .77 .61 Duration of illness, y 3.5 1.8 1 9 Persistent/transient .90 .30 Presenile onset, % 29% Accepts truth .74 .54 Education, y 13.3 3.7 1 20 Hallucinations 1.0 .63 mMMSE 36.3 6.5 23 51 Illusions 1.0 .62 BDRS 3.8 1.9 0.5 12 Behavioral disturbance .88 .67 *AD indicates Alzheimer's disease; mMMSE, modified Mini- Mental State maximum score and Blessed .80 .73 Examination, 57; BDRS, Depression Dementia Rating Scale part 1 (functional activity component only). Subjects Patient Demographics Subjects were outpatients attending a memory disorders The mean age was 72.09 years (SD, 9.77), 29% of clinic, and met the Diagnostic and Statistical Manual of Mental Dis¬ patients had their onset of illness before age 65 years, and orders, Revised Third Edition criteria for primary degenerative de¬ 65% were women (Table 2). Patients with presenile onset mentia of the Alzheimer type and National Institute of Neuro¬ did not differ from those with senile onset on measures logical Disorders and Stroke-Alzheimer's Disease and Related of function Mini-Mental State Exam¬ Disorders Association criteria for AD.10 Exclusion crite¬ cognitive (modified probable ination maximum or on to ria were alcohol or the 5 to [mMMSE], score, 57)12 ability drug dependence during years prior activities of life Dementia the first clinic visit, primary affective disorder within 1 year prior perform daily (Blessed Rating 1 it must be to onset of dementia, and history or clinical evidence of other Scale part [BDRS]),13 although emphasized causes of dementia excluded by differential diagnostic work-up that the range of severity of dementia was restricted a including detailed neurologic and psychiatric evaluation, com¬ priori in this sample. puted tomographic scan of the brain (or magnetic resonance im¬ aging if clinically indicated), electroencephalogram, relevant blood studies, and cerebrospinal fluid studies if indicated. Reli¬ Divergent Validity the for = or ability was assessed by interviewing informants from a group of For 91 subjects, scores presence (score 1) 20 patients with AD at all stages of severity. As part of a base¬ absence (score = 0) of relevant items were summed to ob¬ line evaluation in a longitudinal follow-up study, the trained lay tain several continuous measures: paranoid delusions, interviewer conducted informant interviews in an independent misidentification delusions, all delusions combined, hal¬ sample of 91 AD patients with mild to moderate dementia (Clin¬ lucinations, all items for psychosis, behavioral distur¬ ical Dementia of 1 or All 91 were free of Rating 2).n patients psy- bance, and the sum of all scale items. Neither medication at the time of evaluation. depression, chotropic behavioral disturbance nor depression correlated signifi¬ RESULTS cantly with any subset of items for delusions (Table 3), and there was a low correlation between paranoid delu¬ Reliability sions and delusions of misidentification (r=.12). The The scale was administered to an informant. After a brief weak correlations observed between item subsets suggest in seven establish initial piloting phase patients to agree¬ that they measure different entities that do not share ment on rating criteria, interrater reliability was evaluated much common variance. between a trained lay interviewer (L.M.) and a research psychiatrist (D.P.D.) experienced in the evaluation of pa¬ tients with AD. First, the lay interviewer conducted 10 in¬ Prevalence of Delusional Symptoms terviews with the psychiatrist present in the room, and the Fifty percent of informants (27/54) who initially re¬ two raters subsequently completed scoring for the scale in¬ ported that the patient had no strange ideas subsequently dependently.