Examining a Preclinical Alzheimer’s Cognitive Composite for Telehealth Administration, the tPACC, for Reliability between In-Person and Remote Cognitive Testing
Tugce Duran1, Sarah Gaussoin1, Samuel Lockhart1, Melissa Rundle1, Mark Espeland1, Benjamin Williams1, Timothy Hughes1, Suzanne Craft1, Bonnie Sachs1, James Bateman1
1Wake Forest University School of Medicine
Objective:
To evaluate the concordance of tPACC scores from in-person and remote testing.
Background:
The clinical trial landscape in Alzheimer’s disease (AD) and related dementias is focused on targeting individuals in preclinical stages. The preclinical Alzheimer’s cognitive composite (PACC) was developed for in-person administration to capture subtle cognitive decline in amyloid positive individuals compared to amyloid negative. It is desirable to have a transportable, composite measurement sensitive to detecting cognitive changes across cognitively normal (NC) and impaired (MCI and dementia) participants.
Design/Methods:
We examined in-person cognitive data from 662 adults (70.2±8.1y) from the Wake Forest AD Research Center’s Clinical Core, who received clinical evaluation, cognitive testing, and adjudication. The PACC in-person only was calculated using RAVLT Delayed Recall, Digit Symbol Coding (DSC), semantic fluency, Craft Story Delayed Verbatim, and MMSE total scores using baseline NC participants as a reference. The tPACC used measurements that were available at both in-person and remote visits with modifications from the in-person PACC; Montreal Cognitive Assessment (MoCA) was used in place of MMSE, and DSC was not included. A subset analysis with PACC in-person and tPACC remote from a pilot study examined reliability between in-person and remote testing. We performed correlation analysis and generated Bland-Altman plots.
Results:
Of the 662 adults studied, 434 (66%) were female, 522 (79%) were White, 206 (34%) were
APOE4 carriers, and mean education was 15.9±4.1 years.
At baseline, there is a significant positive relationship between in-person tPACC and PACC (Overall group; r
2=0.94, p=<0.0001, n=640). Although there is a good agreement in both subgroups (Impaired: r
2=0.86, p=<0.0001, n=301 and NC: r
2=0.90, p=<0.0001, n=328), tPACC overestimates cognitive performance compared to PACC for those with lower scores. We also found good agreement between in-person PACC and remote tPACC (r
2=0.83, p=<0.0001, n=38).
Conclusions:
There is generally good agreement between tPACC and PACC for NC and impaired individuals.