Racial/ethnic disparities in neighborhood physical activity environments: Comparing Native Hawaiian and Other Pacific Islander and Asian Americans in the United States.

Native Hawaiian and Other Pacific Islander (NHPI) populations have often been overlooked in health equity research due to their aggregation with Asian Americans. This study compares associations of NHPI and Asian population proportions with physical activity environments (PAEs) at the census tract level across the contiguous U.S., Hawaii, and Alaska, using 2020 Census and 2018 neighborhood data. We analyze four PAE domains-built environment, physical facilities, natural environment, and social environment-using regression models controlling for urbanicity and state. Findings reveal diverging patterns. NHPI proportion is negatively associated with favorable built environment (e.g., physical walkability indicators), physical facilities, and social environment in the contiguous U.S., but shows strong positive associations with favorable natural environment in Hawaii. In contrast, Asian proportion is positively associated with favorable built environment, physical facilities, and social environment in the contiguous U.S., but negatively associated with favorable natural environment, with positive social environment associations in Hawaii. Non-Hispanic White co-residence is associated with better natural environment and physical facilities in Asian tracts but worse built environment in the contiguous U.S. Non-Hispanic Black and Hispanic proportions are associated with better built environments but worse in the other three PAE domains, especially in the contiguous U.S. Alaska's results are limited by small sample sizes. Controlling for poverty strengthens NHPI's natural environment associations, suggesting that these patterns are not explained solely by socioeconomic disadvantages. This study shows that aggregated data mask group-specific PAEs, highlighting NHPI's unique PAE profiles and underscoring the value of disaggregated evidence for informing context-sensitive public health research and intervention design.