# Detecting Anomalous User Behavior in Remote Patient Monitoring

**Authors:** Deepti Gupta, Maanak Gupta, Smriti Bhatt, Ali Saman Tosun  
**arXiv:** [2106.11844](https://arxiv.org/abs/2106.11844)  
**Date:** 2021-06-22  
**Summary page:** https://www.summarizepaper.com/en/paper/2106.11844v1/

## Key points

- Growth in Remote Patient Monitoring (RPM) services using IoMT devices enhances diagnosis quality and enables timely treatment across various medical conditions
- Increased proliferation of IoMT devices poses risks such as theft of personal information, data breaches, and compromised medical devices
- IoMT devices generate vast amounts of data reflecting user behavior patterns, which may lead to anomalies due to unexpected user behavior or compromised devices
- Establishing a robust framework for securing smart healthcare infrastructure is crucial to detect and mitigate anomalies effectively
- An innovative Hidden Markov Model (HMM)-based anomaly detection model tailored for RPM leveraging IoMT and smart home devices has been developed
- The model achieved over 98% accuracy in identifying anomalies within the context of RPM through comprehensive testing with multiple IoMT devices and home sensors
- Network data captured through tools like Wireshark is used to monitor user presence based on device connectivity patterns
- Behavioral data collection techniques are explored to enhance anomaly detection capabilities within remote patient monitoring scenarios

## Summary

The growth in Remote Patient Monitoring (RPM) services utilizing wearable and non-wearable Internet of Medical Things (IoMT) devices holds great promise for enhancing the quality of diagnosis and enabling timely treatment across a wide range of medical conditions. However, with the increasing proliferation of IoMT devices, there is a heightened risk of malicious activities that could result in severe consequences such as theft of personal information, data breaches, and compromised medical devices, ultimately endangering human lives. IoMT devices generate vast amounts of data reflecting user behavior patterns encompassing personal activities, social interactions, and routine health monitoring. Within this context, anomalies may arise due to various factors including unexpected user behavior, sensor malfunctions, or abnormal values stemming from compromised devices. To effectively address this issue, it is imperative to establish a robust framework for securing smart healthcare infrastructure to detect and mitigate anomalies. In response to this critical need, our research presents an innovative anomaly detection model tailored for RPM leveraging IoMT and smart home devices. We introduce a Hidden Markov Model (HMM)-based anomaly detection approach that scrutinizes normal user behavior within the RPM environment comprising both smart home and smart health devices while pinpointing anomalous user actions. To validate our proposed model, we have designed a comprehensive testbed incorporating multiple IoMT devices and home sensors to gather relevant data. By training the HMM model using network traffic and behavioral data collected from these devices, we have achieved remarkable accuracy exceeding 98% in identifying anomalies within the context of RPM. Furthermore, our study delves into the intricacies of network data captured through tools like Wireshark to monitor user presence based on device connectivity patterns. Additionally, behavioral data collection techniques are explored to enhance anomaly detection capabilities within remote patient monitoring scenarios. Overall, our research underscores the significance of developing advanced anomaly detection mechanisms within IoMT-driven healthcare ecosystems to safeguard against potential threats and ensure the seamless delivery of secure and efficient remote patient monitoring services.