Introduction: Longitudinal testing is necessary to accurately measure cognitive change. However, repeated testing is susceptible to practice effects, which may obscure true cognitive decline and delay detection of mild cognitive impairment (MCI).
Methods: We retested 995 late-middle-aged men in a ∼6-year follow-up of the Vietnam Era Twin Study of Aging. In addition, 170 age-matched replacements were tested for the first time at study wave 2. Group differences were used to calculate practice effects after controlling for attrition effects. MCI diagnoses were generated from practice-adjusted scores.
Results: There were significant practice effects on most cognitive domains. Conversion to MCI doubled after correcting for practice effects, from 4.5% to 9%. Importantly, practice effects were present although there were declines in uncorrected scores.
Discussion: Accounting for practice effects is critical to early detection of MCI. Declines, when lower than expected, can still indicate practice effects. Replacement participants are needed for accurately assessing disease progression.
Keywords: Cognitive change; Longitudinal testing; Mild cognitive impairment; Practice effects; Repeat testing; Serial testing.