Health Consumer Usage Patterns in Management of CVD using Data Mining Techniques

Abstract

The Healthcare system is exposed to the increasing impact of chronic diseases including cardiovascular diseases (CVD); it is of much importance to analyze and understand the health trajectories for efficient planning and fair allotment of resources. This work proposes an approach based on mining clinical data to support the exploration of health trajectories related to heart failure hospitalization and readmission. To investigate the potential impact of treatment guideline based on CVD risk factors at a national level requires individual level representation of data with the multi-variable risk factor profiles for the adult population of a country. As these data are highly confidential, we aimed to conduct our experiments using a large, synthetic, longitudinal dataset, constituted to represent the CVD risk factors distribution and temporal sequence of events related to heart failure hospitalization and readmission. Data mining has a high potential for the healthcare industry to use healthcare data efficiently and also system analytics to identify inefficiencies and best practices that improve care and reduce costs. This research work analyses and represents the temporal events or states of the patient's trajectory with the aim of understanding the patient's journey in the management of the chronic condition and its complications by using data mining techniques. This study focuses on developing an efficient algorithm to find cohesive clusters for handling the temporal events. Clustering health trajectories have been carried out by proposing an improved version of the Ant-based clustering algorithm. The application of Ant-based clustering algorithms has been growing over the last few years, and several supervised and unsupervised algorithms have been developed using this bio-inspired approach. It is also important to select a suitable similarity measure to handle the temporal data. Longest common subsequences (LCS) distance measure is one of the most convincing techniques to use for the temporal events. This research work also focuses on interpreting and utilizing the outcome of clustering to predict the future sequences of events including accurately inferring significant disease progression stages. Insights from this study can potentially result in evidence that these approaches are useful in understanding and analyzing patient's health trajectories for better management of the chronic condition and its progression. Several experiments were conducted with the synthetic hospitalization and readmission data set, in which the improved Ant-based clustering algorithm proved to be very effective compared with the standard Hierarchical and K-medoids clustering algorithms. The Ant-based clustering algorithm is useful to discover cohesive clusters based on temporal events. It is also useful for understanding and interpreting the association between the different clusters according to their temporal events. Moreover, it is also used to find the different groups of patients based on the temporal events. Longest common subsequence (LCS) similarity measure seems to be more effective in association with different clustering techniques especially to handle the temporal data. Moreover, the LCS algorithm is also helpful in discovering health consumers patterns to predict the future sequence of temporal events. The accuracy of the prediction process gives reasonable results when compared with the standard hidden Markov model (HMM) and recurrent neural network(RNN). The results of clustering and prediction thus prove to have a positive impact on the understanding of health consumer temporal event patterns.