Understanding the Progression of Congestive Heart Failure of Type 2 Diabetes Patient Using Disease Network and Hospital Claim Data

Abstract

Chronic diseases have increasingly become common and caused most of the burden of ill health in most countries. They have large impacts on quality of life, social and economic conditions. These diseases bring several health risks to those patients suffering from more than one chronic disease at one time (also known as comorbidity of chronic disease). Due to this, governments and healthcare service providers are concerned about the burden of comorbidity of chronic diseases. Understanding the progression of comorbidities can provide vital information for the prevention and better management of chronic diseases. The routinely collected hospital claim data contain semantic information about patients’ health in the form of disease codes. Therefore, these data can be used to understand the progression of chronic disease comorbidities. Most studies in this field are focused on understanding the progression of one chronic disease rather than multiple chronic diseases. In this study, we aim to understand the progression of multiple chronic diseases, i.e., comorbidities that occur when patients of one chronic condition progress towards another. Based on the prevalence of chronic diseases within the Australian population, we have particularly focused on the comorbidity progression of congestive heart failure (CHF) for type 2 diabetes (T2D) patients. In this study, we propose a research framework to understand and represent the progression of CHF in patients with T2D using graph theory and social network analysis. We used hospital claim data drawn from the Australian healthcare context. We constructed two baseline disease networks from two cohorts (i.e., patients with both T2D and CHF and patients with only T2D). A final weighted disease network from two cohorts was then generated by giving more weights to the prevalent comorbidities in patients with T2D and CHF compared to the patients with only T2D. The results show that chronic pulmonary disease, cardiac arrhythmias, valvular disease and renal failure occurred frequently during the progression of CHF for T2D patients. In addition, the final disease network shows the highest transition between electrolyte disorders and renal failure. This indicates that these two diseases may be potential risk factors for the progression towards CHF in patients with T2D for this population cohort. Thus, the proposed network representation can help the healthcare provider to understand high-risk diseases and progression pattern between recurrence of T2D and CHF. Also, it can help in the efficient management of healthcare resources. The proposed framework could be useful for stakeholders including governments and health insurers to adopt appropriate preventive health management program for the patients at high risk of developing multiple chronic diseases.