Outcomes in Heart Failure: Study of Contemporary Trends in a Multi-Ethnic Population

Heart failure is a growing cause of morbidity in an ageing population. Despite increasing use of clinically proven therapies its overall prognosis remains poor, and our knowledge of outcomes in some patient groups is still very limited. Most of existing evidence is based on clinical trial populations, which often exclude ethnic minorities, women or sicker elderly patients. UK’s South Asian population has been shown to suffer from particularly high rates of cardiovascular disease but data on their clinical outcomes have been lacking. This study aimed to evaluate heart failure outcomes in an unselected population and to test a hypothesis of poorer prognosis among South Asians. Using population-based historical cohort design, this thesis evaluates the longterm survival in a large unselected cohort of 5,789 patients with an initial heart failure admission between 1998 and 2001, on a background of the overall trends in heart failure hospitalisation and fatality between 1993 and 2001. The relative risks linked to main patient groups are estimated using logistic regression and survival modelling and a prognostic model is proposed. The results show a plateau in the rates of hospitalisation in the late 1990s. Despite a 50% improvement in survival between 1993 and 2001, outcomes remain poor with a 40% one year fatality. South Asian patients tend to be younger at first admission (by 8 years) and with higher rates of comorbidity, however, their survival appears to be similar to other groups. The developed models indicate high prognostic value of concomitant conditions, such as stroke and renal failure, but only a moderate effect of diabetes. This is the first large study to describe heart failure outcomes in a multi-ethnic contemporary population with an almost complete follow-up of patients. On a background of higher cardiovascular risk, younger age at first admission and higher rate of hospitalisation among South Asians, their clinical outcomes appear to be similar to white patients. Despite the clear limitations inherent in routine data sources, this study shows clear benefits in developing routine risk assessment models for public health research and health care evaluation.