Modeling longitudinal data of congestive heart failure patients: a case study at Wachemo University Nigist Ellen Mohammed memorial Hospital

Abstract

Congestive heart failure (CHF) is a complex clinical syndrome that can result from any functional or structural cardiac disorder that impairs the ventricle’s ability to fill with or eject blood. There are different vital signs of CHF from those most commonly Heart rate, respiration rate, and weight monitoring in the follow-up to assess the progression of congestive heart failure disease. These markers are correlated and needed to ensure an accurate evaluation of them since each has its own limitations and could be influenced by demographical and physiological characteristics of the patient. The objective of the study: The main objective of this study was modeling longitudinal data of congestive heart failure patients in a Case study at Wachemo University Nigist Ellen Mohammed memorial Hospital. Methods: In this study secondary data was used from Wachemo University Nigist Ellen Mohammed memorial Hospital in CHF Outpatient Clinic. The study consists of 154 CHF patients, measured repeatedly at minimum three and maximum nineteen times on each patient who is 18 years old or older for those visited Hospital from December 2015 to January 2018. The linear mixed model was applied in this study to model the three outcomes of CHF. Results: The baseline mean and standard deviation of Pr, Rr, and Wh are 106.16 and 25.37, 31.53 and 11.44 and 64.68 and 10.12 respectively. From the different correlation structure for the separate, bivariate, and multivariate model; modeling with autoregressive order one correlation structure is appropriate for CHF data in addition to unstructured covariance structure for random effects to consider within and between patients variations. Conclusion: Finally a multivariate model was considered as best to study the joint evolution and identify the potential risk factors affecting the three end-points.