A predictive model of hospitalisation and death from chronic obstructive pulmonary disease

Abstract

Rationale: A recent study showed that doctors are excessively pessimistic about the prognosis in patients with COPD and suggested that a simple tool to predict outcome is needed.

Methods: In a prospective observational study, 3343 patients with an FEV1<80% of the predicted value and FEV1/FVC<70% were selected from a clinical network of patients screened for COPD in Tayside, Scotland. Data were collected during annual visits on demography, spirometry, smoking history, medical research council (MRC) dyspnoea scale, body mass index (BMI) and other variables. The main outcome measures were hospitalisations and death secondary to COPD. A proportional hazard model was used to identify significant risk factors.

Results: Increasing age, low BMI, worsening MRC dyspnoea score, decreased FEV1, and prior respiratory or cardiovascular admission hospitalisation were predictors of poor outcome. Influenza vaccination was protective.

Conclusion: We have developed a model that estimates the risk of respiratory hospitalisation and death in patients with COPD.