Related Experiment Video
Updated: Jun 16, 2026
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
Recruitment Mode Selectivity and Interview Mode Measurement Effects in a Population-Based Assessment of Cognitive
Nicholas Hartman1, Emily M Briceño2,3, Roshanak Mehdipanah4
1Department of Biostatistics, School of Public Health, University of Michigan, Ann Arbor, Michigan, USA.
Objectives:
Using data from the Brain Attack Surveillance in Corpus Christi-Cognitive (BASIC-C) study, we investigate recruitment mode selection effects and interview mode measurement effects in population-based applications of the Montreal Cognitive Assessment (MoCA) and Harmonized Cognitive Assessment Protocol (HCAP).
Methods:
We compared participant demographics across in-person (n = 1212), telephone (n = 645), and community (n = 363) samples recruited over 2018-2023 to American Community Survey statistics for Nueces County, Texas. We then implemented unadjusted and adjusted models to assess differences in harmonized MoCA scores, HCAP eligibility and completion, and HCAP cognitive domain measures across recruitment and interview modes.
Results:
The samples had significantly different levels of selection effects, with significant differences in the odds of HCAP eligibility (Odds Ratios (ORs) of 0.51, 1.32, p-value < 0.001). The in-person and telephone interview modes showed significant differences in the cumulative odds of harmonized MoCA score categories (Cumulative OR = 1.58, p-value < 0.001) and several HCAP domain scores, and these differences were attenuated but not eliminated after adjusting for participant demographics.
Conclusions:
The observed differences in MoCA screening outcomes and HCAP eligibility and cognitive domain scores were largely attributable to recruitment mode selection effects. However, statistical adjustments for population characteristics did not fully attenuate interview mode measurement effects.
