Three brief cognitive assessments often used in primary care settings to identify patients with cognitive impairment who could benefit from a full diagnostic workup for dementia are often inaccurate, new research shows.
The three tests are the Mini–Mental State Examination (MMSE), which assesses orientation to time and place and the ability to remember words; the Memory Impairment Screen (MIS), which focuses on the ability to remember words; and Animal Naming (AN), which involves naming as many animals as possible in 60 seconds.
“Our study found that all three tests often give incorrect results that may wrongly conclude that a person does or does not have dementia,” study author David Llewellyn, PhD, of the University of Exeter Medical School, United Kingdom, said in a news release.
The study also found that each test has a different pattern of biases, so people are more likely to be misclassified by one test than another, depending on factors such as their age, education, and ethnicity.
“While these results are at first concerning, knowing the specific limitations for each test will help clinicians decide which is the most appropriate for their patient,” lead author Janice Ranson, doctoral researcher in clinical epidemiology at the University of Exeter Medical School, told Medscape Medical News.
“There are many available brief tests, which all have some limitations and biases, and there is currently not strong enough evidence to suggest one particular test is best for everyone. From our findings, it appears that the best test depends on the clinical context and patient characteristics,” said Ranson.
The study was published online November 28 in Neurology Clinical Practice, a journal of the American Academy of Neurology.
Huge Need for Better Tests
The study included 824 adults (mean age, 82 years) from the population-based Aging, Demographics and Memory Study (ADAMS) who underwent a comprehensive workup for dementia. The workup included physical examination, genetic testing for the APOE gene, psychological testing, and comprehensive memory and thinking tests. On the basis of these results, 35% of the patients were found to have dementia, and 65% were found not to.
Armed with this information, the researchers then had participants take the three brief cognitive assessment tests. They found that 35.7% of participants were wrongly classified by at least one test, 13.4% were misclassified by two or more tests, and 1.7% were misclassified by all three tests. Overall dementia misclassification rates for the MMSE, the MIS, and the AN were 21%, 16%, and 14%, respectively. These rates included both false positive and false negative results.
Years of education predicted higher rates of false negative results (odds ratio [OR], 1.23; 95% confidence interval [CI], 1.07 – 1.40) and lower rates of false positive results (OR 0.77; 95% CI, 0.70 – 0.83) on the MMSE.
Nursing home residency predicted lower rates of false negative results (OR, 0.15; 95% CI, 0.03 – 0.63) and higher rates of false positive results (OR, 4.85; 95% CI, 1.27 – 18.45) on the AN.
Across all tests, not having an informant (such as a family member or friend) weigh in on the patient’s memory was associated with increased risk for misclassification.
“Each test is biased in different ways, so the accuracy of each test varies with the characteristics of the patient,” said Ranson.
“There is clearly great potential for improvement of this initial stage in the diagnostic pathway for dementia. We desperately need more accurate and less biased ways of detecting dementia swiftly in clinic,” he added.
This reinforces that we don’t have one simple test.
“Not Surprising”
Reached for comment, Steven DeKosky, MD, McKnight Brain Institute, Florida Alzheimer’s Disease Research Center, Gainesville, said he’s not surprised by the data.
“We have known about this for a long time, and kudos to this group for doing this careful analysis of all three tests. The fact that only 1.7% of the cases were misdiagnosed when all three tests were used is testimony to the fact that the diagnosis is proportional to the amount of time that you spend testing a patient,” said DeKosky, a fellow of the American Academy of Neurology.
“Everybody is looking for the one test that will tell you whether the patient has Alzheimer’s disease or whether they are impaired, and the human brain is a little too complicated for that.
“Dementia and cognitive impairment of normal aging are multidimensional, continuous processes, and we are trying to nail a state out of what is a bunch of declining lines or stable lines of cognition. This reinforces that we don’t have one simple test. There will always be some people that are missed if you use just one test,” said DeKosky.
He said the study also highlights the importance of having the patient’s partner provide information on whether there is memory loss or not.
“Patients will often tell you that they don’t have memory loss, which is either their denial or possibly loss of insight because they are developing one, or because they just don’t want to believe it. So often, the partner is a more accurate source,” said DeKosky.
The study was supported by the Halpin Trust, the Mary Kinross Charitable Trust, the Engineering and Physical Sciences Research Council, and the UK National Institute for Health Research. The study authors and Dr DeKosky have disclosed no relevant financial relationships.
Neurol Clin Pract. Published online November 28, 2018. Abstract
