Local associations between asthma prevalence and socio-ecological determinants across the United States using spatial machine learning
Topics:
Keywords: Asthma prevalence, Behavioral Health, Noncommunicable Disease, Socio–ecological determinants, Spatial Modeling
Abstract Type: Paper Abstract
Authors:
Aynaz Lotfata, Chicago State University
,
,
,
,
,
,
,
,
,
Abstract
Asthma is a cross-disciplinary public health concern. While some studies have established relationships between socio-ecological characteristics and health, they neither evaluate the relative importance of counties’ cross-disciplinary components from chronic health issues to social and ecological characteristics in the increase of asthma prevalence nor, more crucially, how these factors vary geographically. We use Geographical Random Forest (GRF), a recently established non-linear machine learning approach, to analyze each feature’s spatial variation and contribution to explain regional level asthma prevalence in the United States. In the United States, Smoking Prevalence is the most important factor, whereas language barrier is the least important. While poverty is the most important predictor in southwest and north counties, it is the same as smoking prevalence that is the most critical factor to explain asthma prevalence in north and southwest and along west coast. For policy planning and evidence-based decision-making, we found that cross disciplinary factors are associated with asthma prevalence. As a consequence, interventions should be devised and implemented based on regional circumstances rather than generic notions that apply to the United States’ Counties as a whole.
Local associations between asthma prevalence and socio-ecological determinants across the United States using spatial machine learning
Category
Paper Abstract