Expanding Our Reach: Sampling, Measurement, and Advocacy for Vulnerable Populations

People who are or have been incarcerated are often excluded from research, leading to less information about this population and their systematic exclusion from studies. In this presentation, we will discuss best data collection methods; strategies for retention that ensure quality longitudinal enrollment; and data cleaning, maintenance and analysis approaches for incomplete data. 

Statisticians play a crucial role in ensuring that research accurately represents and benefits the communities it studies. In this presentation, I share approaches for engaging with Latino community-based partners to enhance the impact and relevance of statistical methods in primary care and health disparities. I discuss strategies for building meaningful collaborations, translating complex statistical findings into actionable insights, and incorporating community perspectives into study design and data interpretation. 

Tribal Nations often lack accurate, granular, and comprehensive information on the true burden of mortality indicators across diverse American Indian/Alaska Native (AI/AN) population structures needed for different use cases. Racial misclassification in State death certificates is 30% higher for the AI/AN population compared to other race groups due to higher proportions of multi-ethnic persons within AI/AN populations. Additionally, US Census reported AI/AN population statistics suffer from severe under-representation at every geographic resolution due to small sample sizes, challenges in data collection, cost, and privacy algorithms. Limitations in data accuracy and coverage for both numerator and denominator data result in distorted estimates of the total burden to AI/AN populations. Importantly, official mortality and population statistics do not comprehensively or accurately enumerate diverse AI/AN populations defined by (1) Self-identification/affiliation with the AI/AN racial classification based on the U.S. Census racial categories (racial category), (2) Tribal citizenship with a Tribal Nation (legal category), and (3) Residence on Tribal lands (geographic category). To address these critical gaps, we propose a Bayesian hierarchical model integrating small-area disease mapping and measurement error methods allowing for customizable model components and integration of multiple data sources. Our results point the way toward more comprehensive assessment and modeling of the multiple sources of structured and unstructured uncertainty involved in fusing information from these heterogeneous sources. This proof-of-concept study will serve as a critical step toward the development of a scalable and generalizable framework for Tribal Nations to generate accurate and actionable health estimates. This work aims to empower Tribal Nations with the data sovereignty necessary to monitor and address health disparities within AI/AN communities.