Abstract
OBJECTIVE: Studies and settings designed to identify incident cases of dementia can take advantage of imaging, biomarkers, and specialist clinicians to aid in diagnosis. This is not always feasible for large cohort studies.
DESIGN SETTING AND PARTICIPANTS: The Successful AGing after Elective Surgery (SAGES) study is a long-term, observational study of 560 community-dwelling adults aged ≥70, with serial neuropsychological assessments completed up to 72 months postsurgery.
MEASUREMENT: Mild cognitive impairment (MCI) and dementia diagnoses were determined retrospectively using a rigorous multimodal approach with nested subsamples. This included 1) an expert panel consensus based on serial neuropsychological testing in all participants; 2) diagnosis based on chart review in a subsample of participants; and 3) an in-person neurologist examination in a subsample. An expert adjudication panel then used a Delphi approach to assign research diagnoses incorporating information from all available modalities.
RESULTS: During 72 months of follow-up, 63 incident cases of MCI (11.3%) and four cases of dementia (0.7%) were identified. The most frequent multimodal approach involved combined expert panel consensus and chart review in 41% of MCI (n = 26) diagnoses. Twelve MCI (19%) and three dementia (75%) diagnoses used all three modalities. Eight cases (12%) were diagnosed by a single modality. Diagnostic confidence considered both number and type of information sources used, for example, if agreement by more than one source, or diagnosis only by an in-person exam with a neurologist. Of the 63 MCI and four dementia diagnoses, the expert panel rated their confidence as moderate in four cases (all were MCI) and high in the remaining 63 cases (4 dementia and 59 MCI).
CONCLUSIONS: A novel approach that integrates multiple sources of information and a robust adjudication process to determine incident MCI and dementia in a long-term cohort study may be useful for research applications. Such an approach may be useful when access to expert specialists, biomarkers, advanced neuroimaging and other diagnostic tests may be limited.