Network analytics and machine learning for predictive risk modelling of cardiovascular disease in patients with type 2 diabetes

•Develop models to predict cardiovascular disease for diabetic patients.•Apply network analytics and machine learning on health data for disease modelling.•More informed decisions can be made for healthcare management.•Illustrate an important use of administrative claim data. A high proportion of ol...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Expert systems with applications 2021-02, Vol.164, p.113918, Article 113918
Hauptverfasser: Hossain, Md Ekramul, Uddin, Shahadat, Khan, Arif
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:•Develop models to predict cardiovascular disease for diabetic patients.•Apply network analytics and machine learning on health data for disease modelling.•More informed decisions can be made for healthcare management.•Illustrate an important use of administrative claim data. A high proportion of older adults with type 2 diabetes (T2D) often develop cardiovascular diseases (CVD). Diagnosis and regular monitoring of their multimorbidity is clinically and economically resource intensive. The interconnectedness of their health data and disease progression pathways can potentially reveal the multimorbidity risk if carefully analysed by data mining and network analysis techniques. This study proposed a risk prediction model utilising administrative data that uses network-based features and machine learning techniques to assess the risk of CVD in T2D patients. For this, two cohorts (i.e., patients with both T2D and CVD and patients with only T2D) were identified from an administrative dataset collected from the private healthcare funds based in Australia. Two baseline disease networks were generated from two study cohorts. A final disease network was then generated from two baseline disease networks through normalisation. This study extracted some social network-based features (i.e., the prevalence of comorbidities, transition patterns and clustering membership) from the final disease network and some demographic characteristics directly from the dataset. These risk factors were then used to develop six machine learning prediction models to assess the risk of CVD in patients with T2D. The classifiers accuracy ranged from 79% to 88% shows the potential of the network- and machine learning-based risk prediction model utilising administrative data. The proposed risk prediction model could be useful for medical practice as well as stakeholders to develop health management programs for patients at a high risk of developing chronic diseases.
ISSN:0957-4174
1873-6793
DOI:10.1016/j.eswa.2020.113918