Q&A

Six-item screening tool is sensitive for dementia

J Fam Pract. 2003 January;52(1):12-31
By
Michael Klein, MD
Warren Newton, MD, MPH
Author and Disclosure Information

  • BACKGROUND: Is a six-item screening tool sensitive for dementia? Current screening tools for dementia, including the 30-item Mini-Mental Status Examination (MMSE), are cumbersome and often too time consuming to be used readily in a busy office setting. This study evaluated a new, short screening tool.
  • POPULATION STUDIED: Two populations were studied separately. A community group was recruited from a random sample of a predominantly African American community. The investigators enrolled 344 African Americans who were older than 65 years and had higher-than-average cognitive impairment as established by an epidemiologic screen. The average age was 74.4 years; average education was 10.4 years. Of the 344 subjects, 59.4% were female and the prevalence of dementia was 4.5%. A second group was enrolled from patients referred by family, caregivers, and clinicians to an Alzheimer disease center at a medical center. This sample had 651 subjects who were on average 69.6 years old and had 12.5 years of education. Of this group, 57.1% were female, 16.1% were black, and 53% were demented. The combination of community-dwelling and referral populations, with substantial numbers of African Americans and a wide range of education and cognitive impairment, suggested that the results of this study can be generalized to the typical family practice office setting.
  • STUDY DESIGN AND VALIDITY: This was a prospective study of a 6-item screening tool. The screening tool was derived from the MMSE and measured orientation to year, month, and day of the week and a 3-minute recall of 3 words, for a total of 6 points. All subjects underwent a battery of neuropsychiatric tests and a complete physical examination and evaluation by a geriatric psychiatrist or neurologist. Cognitive assessments included the MMSE, the Cambridge Mental Disorders in the Elderly Examination, and the Consortium for Establishment of Registry of Alzheimer Disease battery, including the Animal Fluency Test, the Boston Naming Test, Constructional Praxis, and the Word List Recall. Whenever possible, a research nurse completed a semi-structured interview to assess activities of daily living and calculate the Blessed Dementia Scale. The experimental screen was compared with the gold standard of expert clinician judgment reviewing all test results and the examination. The area under the response operating characteristic curve was compared for the screen and the MMSE to determine the relative overall accuracy of each test. The performance of the new screen also was compared with the other assessment tools.
  • OUTCOMES MEASURED: The authors measured sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the 6-item screen. Time actually taken and physician satisfaction with the tool were not assessed.
  • RESULTS: In the community sample, sensitivity and specificity for 1 error or more on the 6-item test were 100% and 38.4%, respectively, for dementia, with a PPV of 6.7% and an NPV of 100%. In the referral sample, sensitivity and specificity for 1 error or more on the screener were 96.8% and 53.3%, respectively, for dementia, with a PPV of 70.0% and a NPV of 93.7%. The area under the response operating characteristic curves for the screen and the MMSE were similar, and the scores of the other instruments progressively worsened as the number of errors on the screen increased.
PDF Download

 

PRACTICE RECOMMENDATIONS

This study provides excellent evidence that a 6-item screening tool based on orientation to date and a 3-item short-term recall is sensitive for dementia. Clinicians should consider using this simple screen in an outpatient setting, keeping in mind that the final diagnosis of dementia is a clinical judgment after full assessment and that this screen was not intended for use in following patients over time. Caution also should be exercised in extending these results to patients in the hospital, who were not included in this trial.

Next Article:

What are the indications for urodynamic testing in older adults with incontinence?