Most current machine learning predictors lack the contextual consideration required for real world application. Chronic kidney disease (CKD) is a long-term disease which remains under diagnosed in early stages despite a growing number of patients requiring medical intervention. This paper aims to explore the practical efficacy of a machine learning predictor within a clinical workflow using an open-source medical dataset. The data set used is from the UC Irvine machine learning repository consisting of 400 patient records. Using a Logistic Regression algorithm to predict patient outcomes the algorithm was able to achieve a 98% accuracy with an F1 score of 97%. In addition to the overall prediction, a study emphasizing user heuristics and user accessibility was conducted using third-party evaluation software to identify areas of compliance and further improvement for Nielsen heuristics as well as WCAG guidelines. Along with this data visualisation has been emphasised using Python specific libraries such as Plotly were used to visualise patient data and compare attributes using scatterplots as well as compare user inputted data to the imputed CSV data used to train the model using histograms. These considerations were used to identify the efficacy for machine learning applications within real world application.