Utility of the Montreal Cognitive Assessment and Mini-Mental State Examination in Predicting General Intellectual Abilities Michael A. Sugarman, MA*w and Bradley N. Axelrod, PhD*
Objective: To determine whether scores from 2 commonly used cognitive screening tests can help predict general intellectual functioning in older adults.
FSIQ = Full Scale Intelligence Quotient. MMSE = MiniMental State Examination. MoCA = Montreal Cognitive Assessment. TOPF = Test of Premorbid Functioning. WAIS = Wechsler Adult Intelligence Scale.
Background: Cutoﬀ scores for determining cognitive impairment have been validated for both the Montreal Cognitive Assessment (MoCA) and the Mini-Mental State Examination (MMSE). However, less is known about how the 2 measures relate to general intellectual functioning as measured by the Wechsler Adult Intelligence Scale–Fourth Edition (WAIS-IV).
Received for publication July 3, 2013; accepted December 4, 2013. From the *John D. Dingell Department of Veterans Aﬀairs Medical Center, Detroit, Michigan; and wWayne State University, Department of Psychology, Detroit, Michigan. The authors declare no conﬂicts of interest. Reprints: Michael A. Sugarman, MA, Wayne State University, Department of Psychology, 7th Floor, 5057 Woodward Ave, Detroit, Michigan 48202 (e-mail: [email protected]
linicians are often called upon to perform a cognitive screening test as part of their assessment of a patient. Findings from such a test oﬀer insight into the patient’s general cognitive functions of orientation (to person, place, time, and event), thought content, memory, and judgment. Two standardized screening measures that are used commonly in clinical practice are the Mini-Mental State Examination (MMSE) (Folstein et al, 1975) and the Montreal Cognitive Assessment (MoCA) (Nasreddine et al, 2005). Both are now used most often as screening tools for cognitive impairment in adults aged 60 years and older. The MMSE is a 30-point measure comprising questions about orientation, following of a 3-step command, following of a written command, serial mental subtraction, repeating phrases, generating a sentence in writing, copying a simple drawing, and recalling 3 words after a delay. Drawing on the work of Crum and colleagues (1993), Psychological Assessment Resources (Folstein, 2001) published normative data from >18,000 respondents. The norms include cutoﬀ scores for cognitive impairment based on the examinee’s age and education level. Farber et al (1988) studied a sample of patients with Alzheimer disease and demonstrated that their MMSE scores predicted their general intellectual abilities as measured by the Wechsler Adult Intelligence Scale– Revised (WAIS-R) (Wechsler, 1981). Farber et al (1988) provided a regression equation for using the MMSE to estimate WAIS-R Full Scale Intelligence Quotient (FSIQ) in clinical practice. The MoCA is also a 30-point measure, similar in construction to the MMSE but containing several more complex items: drawing a clock, copying a cube, completing a brief task similar to the Trail Making Test Part B (Reitan, 1958), naming objects, generating words that begin with the letter F, repeating a sentence, recalling 5 words (rather than the 3 words on the MMSE), and answering 2 word similarity items. Because fewer points are given for orientation and more points are given for complex concepts, the MoCA tends to be more diﬃcult than the MMSE (Nasreddine et al, 2005). The MoCA has
148 | www.cogbehavneurol.com
Cogn Behav Neurol
Methods: A sample of 186 older adults referred for neuropsychological assessment completed the MoCA, MMSE, and WAIS-IV. Regression equations determined how accurately the screening measures could predict the WAIS-IV Full Scale Intelligence Quotient (FSIQ). We also determined how predictive the MoCA and MMSE were when combined with 2 premorbid estimates of FSIQ: the Test of Premorbid Functioning (TOPF) (a reading test of phonetically irregular words) and a predicted TOPF score based on demographic variables. Results: MoCA and MMSE both correlated moderately with WAIS-IV FSIQ. Hierarchical regression models containing the MoCA or MMSE combined with TOPF scores accounted for 58% and 49%, respectively, of the variance in obtained FSIQ. Both regression equations accurately estimated FSIQ to within 10 points in >75% of the sample. Conclusions: Both the MoCA and MMSE provide reasonable estimates of FSIQ. Prediction improves when these measures are combined with other estimates of FSIQ. We provide 4 equations designed to help clinicians interpret these screening measures. Key Words: Montreal Cognitive Assessment, Mini-Mental State Examination, Wechsler Adult Intelligence Scale, Test of Premorbid Functioning, neuropsychological testing (Cogn Behav Neurol 2014;27:148–154)
Volume 27, Number 3, September 2014
Cogn Behav Neurol
Volume 27, Number 3, September 2014
shown better sensitivity in detecting cognitive impairment (Hoops et al, 2009). Nasreddine et al (2005) recommended a cutoﬀ score of 26 out of 30 in assisting with a diagnosis of mild cognitive impairment or Alzheimer disease. However, more recent research has suggested that this cutoﬀ score might be too high. Waldron-Perrine and Axelrod (2012) recommended a cutoﬀ score of r21 for determining cognitive impairment, while Freitas et al (2013) proposed a cutoﬀ score of r22. No study has yet used the MoCA to predict examinees’ current intellectual abilities, as did Farber et al (1988) using the MMSE. A continuous metric for interpreting the MoCA could yield valuable information to supplement the traditional dichotomous cutoﬀ score. In this study, we used both the MoCA and MMSE to predict patients’ current intellectual abilities as determined by the FSIQ on the current version of the WAIS, the WAISIV (Wechsler, 2008). We compared our MMSE prediction equations to the equation that Farber et al (1988) had generated for the WAIS-R. This study was the ﬁrst analysis of the predictive utility of the MoCA. Our goal in the study was to develop regression-based prediction equations of FSIQ for use in clinical practice with older adults. Further, we sought to determine how well these screening tools could predict FSIQ when we considered the screens in conjunction with 2 other estimates of FSIQ: reading ability and demographic variables.
METHODS Participants We evaluated 186 US military veterans (5 women, 181 men) aged 60 years and older whose physician had referred them for neuropsychological evaluation at the John D. Dingell Veterans Aﬀairs Medical Center in Detroit, Michigan. We analyzed data only from patients
MoCA and MMSE Predicting Full Scale IQ
whom we deemed to have completed a valid neuropsychological assessment. From an initial sample of 208, we excluded patients if they showed evidence of negative response bias, deﬁned as failure on at least 2 of these 3 performance validity measures: A score