Introduction COPD continues to cause a substantial symptom, mortality and financial burden in the UK. Current treatment strategies are predominantly reactive as insufficient evidence exists to successfully target clinical resource into pre-emptive ‘early interventions’. The DOSE (dyspnoea, obstruction, smoking status and exacerbation) score has been validated as a risk predictor for mortality, hospitalisation and poorer health status. However, only a small proportion of COPD patients with poor outcomes have high DOSE scores. We sought to establish if clinical characteristics can be used to pre-emptively identify those COPD patients vulnerable to future poor health status by using an electronic database of anonymised patient records-the Hampshire Health Record Analytical Database (HHRA).
Methods Within our HHRA database COPD cohort, we identified a cohort of 6890 patients who fell into the ‘low risk’ category by DOSE score (<4). Within this group, a subset met the criteria for poor COPD outcomes over the next four years, defined as; death (all cause), COPD related hospital admission, a DOSE score increase of ≥2 points or a subsequent DOSE score of ≥4 (high risk). We used logistic regression analysis to examine the association between demographic and clinical characteristics documented by Read code at baseline and those who subsequently fell into the poor outcomes subgroup.
Results In our ‘low risk’ cohort of 6890 COPD patients, 5000 held sufficient data to be included in the analysis. After four years, 2211 (44.2%) of those 5000 patients fell into the poor outcomes subgroup. As shown in Table 1, poor future outcomes were significantly associated with age, high deprivation decile, low BMI, certain comorbidities, a raised eosinophil percentage (≥2%) and the prescription (in the preceding twelve months) of nebulised bronchodilators, inhaled bronchodilators and an ICS/LABA combination inhaler. A BMI of >25 and rhinosinusitis were associated with a lower risk of poor future outcomes.
Conclusions Poor future clinical outcomes appear to be associated with certain clinical characteristics in a COPD database cohort deemed low risk by DOSE score. These findings warrant further validation in a clinical cohort and investigation into the effect of pre-emptive optimisation of these characteristics on health outcomes.