Drivers of underreport of memory loss in mild cognitive impairment due to Alzheimer’s and vascular disease

Abstract

AbstractBackgroundWe examined self‐ and informant‐based report of memory loss in mild cognitive impairment due to Alzheimer’s disease (AD‐MCI) and vascular disease (Va‐MCI). We hypothesized that drivers of underreport would differ based on etiology and participant characteristics.MethodThis retrospective cross‐sectional study included participants with AD‐MCI (n = 2874) and Va‐MCI (n = 376) in the National Alzheimer’s Coordinating center dataset (NACC). Report of memory loss by the participant and informant was recorded as a binary (yes/no) value. We examined the role of demographic variables (age, sex, education, race, ethnicity) and neuropsychological variables (anxiety, depression on NPI‐Q) in driving underreporting of memory loss using Fisher’s exact test.ResultCompared to AD‐MCI, Va‐MCI was associated with significantly more underreport of memory loss by participants (24.5% vs 19.7%, p = 0.031) and informants (31.4% vs 21.6%, p<0.0001). Black/African American participants displayed higher rates of underreport in both AD‐MCI (27.1% vs. 18.4%, p<0.0001) and Va‐MCI (37.8% vs. 20.8%, p = 0.0022). Informant underreport was increased in both AD‐MCI and Va‐MCI for participants who were female (AD: 24.3% vs 18.6%, p = 0.0002; Va:38.1% vs 22.2%, p = 0.0011) and Black/African American (AD: 35.1% vs 19.3%, p<0.0001; Va: 48.8% vs. 26.5%, p = 0.0002), and in AD‐MCI for participants who had less than 16 years of education (AD: 23.9% vs 20.3%, p = 0.0262). In both AD‐MCI and Va‐MCI, anxiety associated with less underreport in both participants (AD: 13.1% vs 22.0%, p<0.0001; Va: 15.91% vs. 31.2%, p = 0.0071) and informants (AD: 10.8% vs. 25.5%, p<0.0001; Va: 15.9% vs. 37.4%, p = 0.0002). Underreport was similarly lowered with depression in participants (AD: 14.7% vs 21.4%, p<0.0001; Va: 15.1% vs 28.3%, p = 0.0146) and informants (AD: 11.5% vs 25.2%, p<0.0001; Va: 17.4% vs. 36.8%, p = 0.0008).ConclusionPersons with MCI and who have likely vascular presentations, or who are female, Black/African American, or have lower education are at increased risk for self‐ and/or informant‐based underreport of memory loss. Affective symptoms appear to increase the likelihood of correctly reporting memory loss. These factors may be particularly important in influencing interview‐based MCI detection and would suggest extra vigilance in persons at‐risk for self‐ or informant‐based misreport.