Study Protocol Using Cohort Data and Latent Variable Modeling to Guide Sampling Women with Type 2 Diabetes and Depressive Symptoms

Nicole Beaulieu Perez, Gail D'Eramo Melkus, Gary Yu, Janet Brown-Friday, Kathryn Anastos, Brad Aouizerat

Research output: Contribution to journalArticlepeer-review


Background Depression affects one in three women with Type 2 diabetes, and this concurrence significantly increases the risks of diabetes complications, disability, and early mortality. Depression is underrecognized because of wide variation in presentation and the lack of diagnostic biomarkers. Converging evidence suggests inflammation is a shared biological pathway in diabetes and depression. Overlapping epigenetic associations and social determinants of diabetes and depression implicate inflammatory pathways as a common thread. Objectives This article describes the protocol and methods for a pilot study aimed to examine associations between depressive symptoms, inflammation, and social determinants of health among women with Type 2 diabetes. Methods This is an observational correlational study that leverages existing longitudinal data from the Women's Interagency HIV Study (WIHS), a multicenter cohort of HIV seropositive (66%) and HIV seronegative (33%) women, to inform purposive sampling of members from latent subgroups emergent from a prior retrospective cohort-wide analysis. Local active cohort participants from the Bronx study site are then selected for the study. The WIHS recently merged with the Multicenter Aids Cohort Study (MACS) to form the MACS/WIHS Combined Cohort Study. Latent subgroups represent distinct symptom trajectories resultant from a growth mixture model analysis of biannually collected depressive symptom data. Participants complete surveys (symptom and social determinants) and provide blood samples to analyze plasma levels and DNA methylation of genes that encode for inflammatory markers (CRP, IL-6, TNF-). Correlation and regression analysis will be used to estimate the effect sizes between depressive symptoms and inflammatory markers, clinical indices (body mass index, hemoglobin A1C, comorbidities), and social determinants of health. Results The study began in January 2022, and completed data collection is estimated by early 2023. We hypothesize that depressive symptom severity will associate with higher levels of inflammation, clinical indices (e.g., higher hemoglobin A1C), and exposure to specific social determinants of health (e.g., lower income, nutritional insecurity). Discussion Study findings will provide the basis for future studies aimed at improving outcomes for women with Type 2 diabetes by informing the development and testing of precision health strategies to address and prevent depression in populations most at risk.

Original languageEnglish (US)
Pages (from-to)409-415
Number of pages7
JournalNursing research
Issue number5
StatePublished - Sep 1 2023


  • Type 2 diabetes
  • depressive symptoms
  • diabetes
  • latent variable modeling
  • precision health
  • sampling
  • social determinants of health
  • stress
  • women

ASJC Scopus subject areas

  • General Nursing


Dive into the research topics of 'Study Protocol Using Cohort Data and Latent Variable Modeling to Guide Sampling Women with Type 2 Diabetes and Depressive Symptoms'. Together they form a unique fingerprint.

Cite this