Progressive lung disease is the leading cause of death in individuals with cystic fibrosis (CF). Rapid decline, characterized by accelerated loss of lung function, is common for CF patients, and cannot be explained or predicted by CFTR/gene dysfunction alone. Mapping the environmental exposures and community characteristics (geomarkers) that predict patient-specific rapid decline and providing tools for earlier detection and monitoring at the center level are essential to transforming the precision of CF clinical care, and offer an opportunity to adjust interventions to prevent irreversible lung damage. The translation of these tools into practice is further hindered by continued use of antiquated statistical methods that ignore the interplay between nonlinear lung function and recurrent pulmonary exacerbations in the clinical course of CF, disregard known mortality biases that can lead to inaccurate projections of rapid decline, and do not leverage extant geographic data on geomarkers, such as air quality or neighborhood socioeconomic conditions, to improve prediction of rapid decline. In this proposal, we will utilize comprehensive geocoding algorithms, novel statistical methods and powerful computational medicine tools for integration into clinical algorithms for the detection to drive early intervention of rapid lung disease progression. The overall objective of this research is to leverage a rich CF registry, extant national and local environmental data sources and prospectively collected study data to accurately forecast the onset of rapid decline in individual patients, and to develop a feasible medical- monitoring tool that positively impacts CF point-of-care decision-making. Our overarching hypothesis is that interactive computational medicine tools for dynamic prediction and clinical surveillance of rapid pulmonary decline in CF will enhance local disease monitoring. This will be accomplished by incorporating both established and novel environmental exposures and community characteristics of CF patients.
Our specific aims are to 1) phenotype patients who experience early, rapid pulmonary decline informed by environmental exposures; 2) transform dynamic prediction of rapid lung-function decline and exacerbations in CF patients through high-dimensional, multi-level joint model mapping with environmental factors; 3) design and implement decision support capabilities that monitor real-time lung-function decline and risk of exacerbations for personalized, center-specific CF patient management. Once systems to accurately and precisely predict rapid decline are in place, better prospective treatment decisions will become possible, resulting in better patient outcomes and improved precision medicine/care.
The proposed research is relevant to public health because having clinical tools for early detection and timely treatment of rapid pulmonary decline in cystic fibrosis will improve the real-world effectiveness of clinical interventions. Accounting for environmental influences is ultimately expected to optimize these tools and to appropriately assign treatment and prevention of rapid decline, thereby helping individuals with cystic fibrosis live healthier, longer lives. Thus, the proposed research is relevant to multiple parts of NIH?s mission that pertain to i) enabling rapid translation of discoveries in quantitative biomedicine into clinical practice; ii) considering the environment of each patient to facilitate precision medicine activities.