On The Utility of Stochastic Models of Chronic Obstructive Pulmonary Disease

Chronic obstructive pulmonary disease (COPD) is a progressive and irreversible lung disease characterized by reduced lung function, abnormal inflammatory response, dyspnea, increased sputum production, and costly and debilitating exacerbations. Common treatments for COPD include smoking cessation therapy, short- and long-acting bronchodilators, inhaled corticosteroids, pulmonary rehabilitation, and home oxygen therapy. However, a recent meta-analysis demonstrated that inhaled corticosteroids fail to slow the decline in FEV1, the primary clinical marker of COPD progression. In addition, some evidence exists to suggest that long-term inhaled corticosteroid use may increase bone density loss, leading to osteoporosis. Given that the bone health effects of inhaled corticosteroid use may have detrimental effects on patient health-related quality of life and increase direct medical costs, while having no effect on the primary clinical marker of disease progression, the wisdom of their widespread use among COPD sufferers has become a topic of hot debate in the pulmonary community. We addressed the common clinical questions surrounding the efficacy and side effects of inhaled corticosteroids through the application of two literature-based Markov models of direct medical costs and patient quality of life over short and long simulation periods. The three-year model of disease progression and exacerbations indicates that inhaled corticosteroid use is a cost-saving measure of improving patient quality of life, with cost-savings ranging from $7,500 to $150,000 per quality-adjusted life year gained. The use of inhaled corticosteroids also reduces the fraction of severe exacerbations; however, it appears that their use does not slow progression to more severe forms of COPD. The twenty-year model of osteoporotic fractures in COPD patients demonstrates that, in the worst case, the use of inhaled corticosteroids simultaneously fails to increase patient mortality or decrease the number of quality-adjusted life years lived by increasing the incidence of osteoporotic fractures. The cost-utility ratio in this model ranges from approximately $7,000 to $200,000, depending on the model assumptions, but remains moderate for most simulated treatment effects. We conclude that Markov modelling is an appropriate method of summarizing the COPD literature to reach clinically useful estimates of the effects of inhaled corticosteroids on direct medical costs and patient health-related quality of life