[Zurück]

I. Azimi, T. Pahikkala, A. Rahmani, H. Niela-Vilén, A. Axelin, P. Liljeberg:
"Missing data resilient decision-making for healthcare IoT through personalization: A case study on maternal health";
Future Generation Computer Systems, 96 (2019), S. 297 - 308.

Remote health monitoring is an effective method to enable tracking of at-risk patients outside of conventional clinical settings, providing early-detection of diseases and preventive care as well as diminishing healthcare costs. Internet-of-Things (IoT) technology facilitates developments of such monitoring systems although significant challenges need to be addressed in the real-world trials. Missing data is a prevalent issue in these systems, as data acquisition may be interrupted from time to time in long-term monitoring scenarios. This issue causes inconsistent and incomplete data and subsequently could lead to failure in decision making. Analysis of missing data has been tackled in several studies. However, these techniques are inadequate for real-time health monitoring as they neglect the variability of the missing data. This issue is significant when the vital signs are being missed since they depend on different factors such as physical activities and surrounding environment. Therefore, a holistic approach to customize missing data in real-time health monitoring systems is required, considering a wide range of parameters while minimizing the bias of estimates. In this paper, we propose a personalized missing data resilient decision-making approach to deliver health decisions 24/7 despite missing values. The approach leverages various data resources in IoT-based systems to impute missing values and provide an acceptable result. We validate our approach via a real human subject trial on maternity health, in which 20 pregnant women were remotely monitored for 7 months. In this setup, a real-time health application is considered, where maternal health status is estimated utilizing maternal heart rate. The accuracy of the proposed approach is evaluated, in comparison to existing methods. The proposed approach results in more accurate estimates especially when the missing window is large.

http://dx.doi.org/10.1016/j.future.2019.02.015