Cross-national harmonization of cognitive measures across HRS HCAP (USA) and LASI-DAD (India)

• DOI: 10.1371/journal.pone.0264166
• Publication Date: 2022-02-25T18:32:26Z
• Authors: Jet M. J. Vonk; Alden L. Gross; Andrea R. Zammit; Laiss Bertola; Justina F. Avila; Roos J. Jutten; Leslie S. Gaynor; Claudia K. Suemoto; Lindsay C. Kobayashi; Megan E. O’Connell; Olufisayo Elugbadebo; Priscilla A. Amofa; Adam M. Staffaroni; Miguel Arce Rentería; Indira C. Turney; Richard N. Jones; Jennifer J. Manly; Jinkook Lee; Laura B. Zahodne
• Keywords: Harmonization

Background As global populations age, cross-national comparisons of cognitive health and dementia risk are increasingly valuable. It remains unclear, however, whether country-level differences in function attributable to population or bias due incommensurate measurement. To demonstrate an effective method for comparison studies, we aimed statistically harmonize measures episodic memory language across two population-based cohorts older adults the United States (HRS HCAP) India (LASI-DAD). Methods Data 3,496 HRS HCAP (≥65 years) 3,152 LASI-DAD (≥60 participants were harmonized performance using confirmatory factor analysis (CFA) methods. Episodic variables investigated differential item functioning (DIF) precision. Results CFA models estimating domains based on a priori adjudication comparable items fit data well. DIF analyses revealed that four out ten five twelve measured underlying construct comparably samples. DIF-modified scores showed patterns precision range latent trait each sample. Conclusions Harmonization will facilitate future investigation effects factors, policies, treatments, reducing study-level measurement administrative influences. international aging studies become more widely available, advanced statistical methods such as those described this study central making universal generalizations drawing valid conclusions about population.