Diagnosing Dementia and Cognitive Impairment in Symptomatic Patients

How accurate is the clinical judgment of primary care physicians when diagnosing symptomatic patients with dementia and cognitive impairment?

In patients who present with cognitive concerns, the clinical judgment of primary care physicians is sensitive, but not specific, when used to screen for any cognitive impairment. It is specific, but not sensitive, when diagnosing overt dementia.¹ Primary care physicians should not rely solely on clinical judgment when evaluating patients for dementia or cognitive impairment; however, clinical judgment may help inform the choice of additional testing to confirm or exclude the diagnoses of dementia and cognitive impairment. (Strength of Recommendation: B, inconsistent or limited-quality patient-oriented evidence.)

Patients with cognitive concerns commonly visit their primary care physicians for an initial evaluation. The incidence of cognitive impairment increases with age and encompasses a spectrum ranging from mild cognitive impairment to dementia. Approximately 6.9 million people in the United States who are 65 years and older are living with dementia, and this number will increase as the population ages.² Making a diagnosis of dementia and cognitive impairment can be challenging because it requires taking an extensive history in multiple cognitive domains that affect social and occupational functions, plus a mental status evaluation using clinical assessment tools that are inherently flawed. A neuropsychological evaluation may be needed for some patients when the initial assessment is inconclusive. Primary care physicians may also use their clinical judgment to help determine whether someone has dementia or cognitive impairment. The authors of this review sought to determine whether the clinical judgment of primary care physicians is accurate in screening for or diagnosing cognitive impairment and dementia in patients with cognitive concerns.

This Cochrane review included 12 cross-sectional and cohort studies that were published through September 2021.¹ Most of the studies were conducted in Europe, with one in the United States and one in Australia. The review involved 4,287 participants with an average age of 73 to 83 years. Reference standards, including the Diagnostic and Statistical Manual of Mental Disorders, International Classification of Diseases (ICD), etiologic definitions, or expert clinical diagnosis, were used to diagnose dementia and cognitive impairment.

Of eight studies involving 2,790 participants, 826 patients (30%) had dementia by reference standards. Clinical judgment by primary care physicians had a specificity of 89% (95% CI, 79% to 95%) and sensitivity of 58% (95% CI, 43% to 72%) in diagnosing dementia.

A total of 1,497 people were evaluated in the four cognitive impairment studies; 594 participants (40%) had cognitive impairment by reference standards. Clinical judgment by primary care physicians had a sensitivity of 84% (95% CI, 60% to 95%) and a specificity of 73% (95% CI, 50% to 88%) in diagnosing cognitive impairment.

These conclusions should be interpreted with caution due to problems with the original study designs. Differences in the prevalence of dementia between studies make it difficult to compare them; however, there were consistent trends in the estimated accuracy of the test among the studies.

These results suggest that primary care physicians are reasonably good at diagnosing cognitive impairment but are not as skilled at diagnosing dementia using only their clinical judgment. Although most primary care physicians do not rely on clinical judgment alone when evaluating a patient with cognitive concerns, their judgment may help them choose an additional assessment tool to confirm or exclude a diagnosis of dementia. If a physician thinks that a patient with cognitive concerns has dementia, an assessment tool with high specificity may be used to confirm the diagnosis. If a physician thinks a patient with cognitive concerns does not have dementia, an assessment tool with high sensitivity may be used to confirm that the patient does not.