A Multi-Tiered Approach to Addressing Air Quality Concerns on Hopi Lands

Abstract

Introduction: Air pollution remains a significant public health hazard, resulting in over seven million deaths annually. Few studies within solid fuel (biomass, coal) burning communities simultaneously quantify indoor concentrations of PM2.5 and examine their association with acute health effects. The overall goal of this dissertation research is to assess the relationships between indoor air quality and respiratory health among members of the Hopi Tribe in Northern Arizona. Methods: Data were collected as part of the Hopi Environmental Health Project (HEHP). Indoor and outdoor concentrations of PM2.5 were measured for 24-hours at 76 randomly selected homes during the heating and non-heating seasons. Airway inflammation of participants was measured at the end of each 24-hour sampling period using fractional exhaled nitric oxide (FeNO). Univariate and multivariate time-series imputation methods were evaluated to address problems with homes having consecutive periods of missing data for the real-time air monitors. Multivariable regression modeling examined associations among indoor PM2.5 levels, seasonal household, behavioral, and environmental factors and to assess the relationship between indoor PM2.5 and FeNO levels. Results: Univariate methods of Markov, random, and mean imputation were the best performing methods, yielding imputed 24-hour concentrations and minute-by-minute concentrations with the lowest errors across various levels of missingness. During the winter heating season, indoor PM2.5 levels were positively associated with housing type, heating fuel type, presence of pests, indoor relative humidity, and negatively associated with use of a climate control device and the number of interior doors. However, during the non-heating season, different behavioral and household characteristics were associated with indoor PM2.5 concentrations. Indoor smoking and/or burning incense, area of the surrounding environment, and outdoor PM2.5 were positively associated with indoor PM2.5 concentrations. Building size and height, and opening of doors and windows were negatively associated with indoor concentrations. During the heating season, indoor PM2.5 concentration and work location were negatively associated with FeNO. Indoor temperature and heart rate were positively associated (p≤0.05) with FeNO levels during the heating season. Univariable and multivariable regression showed an inverse relationship between indoor PM2.5 and FeNO levels. During the non-heating season indoor temperature, use of non-steroid anti-inflammatory drugs (NSAID’s), and respiratory symptoms were associated with greater FeNO measures. Conclusion: Imputation using univariate methods provided a reasonable solution to addressing missing data for short-term monitoring of air pollutants, especially in resource-limited areas. Within the households tested, we observed seasonal differences in household and behavioral factors associated with indoor PM2.5 concentrations. Homes that burned coal and/or wood, or a combination of coal and/or wood with electricity and/or natural gas had elevated indoor PM2.5 concentrations that exceeded both the EPA ambient standard (35 μg/m3) and the WHO indoor air guidelines (25 μg/m3). Environmental, personal, and physiological factors were associated with FeNO levels and these associations differed by season; however, indoor concentrations of PM2.5 were not associated with greater airway inflammation levels as anticipated.