RESUMO
Newborn screening programs for severe metabolic disorders using tandem mass spectrometry are widely used. Medium-Chain Acyl-CoA dehydrogenase deficiency (MCADD) is the most prevalent mitochondrial fatty acid oxidation defect (1:15,000 newborns) and it has been proven that early detection of this metabolic disease decreases mortality and improves the outcome. In previous studies, data mining methods on derivatized tandem MS datasets have shown high classification accuracies. However, no machine learning methods currently have been applied to datasets based on non-derivatized screening methods. A dataset with 44,159 blood samples was collected using a non-derivatized screening method as part of a systematic newborn screening by the PCMA screening center (Belgium). Twelve MCADD cases were present in this partially MCADD-enriched dataset. We extended three data mining methods, namely C4.5 decision trees, logistic regression and ridge logistic regression, with a parameter and threshold optimization method and evaluated their applicability as a diagnostic support tool. Within a stratified cross-validation setting, a grid search was performed for each model for a wide range of model parameters, included variables and classification thresholds. The best performing model used ridge logistic regression and achieved a sensitivity of 100%, a specificity of 99.987% and a positive predictive value of 32% (recalibrated for a real population), obtained in a stratified cross-validation setting. These results were further validated on an independent test set. Using a method that combines ridge logistic regression with variable selection and threshold optimization, a significantly improved performance was achieved compared to the current state-of-the-art for derivatized data, while retaining more interpretability and requiring less variables. The results indicate the potential value of data mining methods as a diagnostic support tool.
Assuntos
Mineração de Dados/métodos , Triagem Neonatal/métodos , Espectrometria de Massas em Tandem/métodos , Acil-CoA Desidrogenase/classificação , Acil-CoA Desidrogenase/deficiência , Inteligência Artificial , Bélgica , Humanos , Recém-Nascido , Erros Inatos do Metabolismo Lipídico/classificaçãoRESUMO
OBJECTIVE: The aim of this study is to guide healthcare instances in applying process analytics on healthcare processes. Process analytics techniques can offer new insights in patient pathways, workflow processes, adherence to medical guidelines and compliance with clinical pathways, but also bring along specific challenges which will be examined and addressed in this paper. METHODS: The following methodology is proposed: log preparation, log inspection, abstraction and selection, clustering, process mining, and validation. It was applied on a case study in the type 2 diabetes mellitus domain. RESULTS: Several data pre-processing steps are applied and clarify the usefulness of process analytics in a healthcare setting. Healthcare utilization, such as diabetes education, is analyzed and compared with diabetes guidelines. Furthermore, we take a look at the organizational perspective and the central role of the GP. This research addresses four challenges: healthcare processes are often patient and hospital specific which leads to unique traces and unstructured processes; data is not recorded in the right format, with the right level of abstraction and time granularity; an overflow of medical activities may cloud the analysis; and analysts need to deal with data not recorded for this purpose. These challenges complicate the application of process analytics. It is explained how our methodology takes them into account. CONCLUSION: Process analytics offers new insights into the medical services patients follow, how medical resources relate to each other and whether patients and healthcare processes comply with guidelines and regulations.