The healthcare industry has generated a huge amount of data in diverse formats. The big data in healthcare is leading the revolution in healthcare. Collecting data at the operational level is the starting point for the big data-driven healthcare revolution. By analyzing the operational level big data, healthcare organizations can gain the business intelligence for further strategy development, for example how to improve the healthcare quality, how to provide better long-term care, and how to empower the patients. This chapter discusses this process as operations-intelligence-strategy (OIS) process in healthcare. Objectives are understanding how to gain business intelligence from sensor data mining in healthcare, biomedical signal analysis, and biomedical image analysis, and exploring the applications and impacts of the OIS process, with a focus on the sensor data mining in healthcare.
TopSensor Data Mining In Healthcare
Sensor refers to a device, module, or subsystem that is used to detect changes in its environment and then send the information to other processing devices or cloud platforms (Dargie & Poellabauer, 2010). Examples of common sensors in daily life include breath analyzer, CO2 sensor, CO detector, smoke detector and etc. Sensor-based monitoring is applied for data collection purposes as sensors provide immense amounts of data. In healthcare, it is critical to use sensors to monitor some cases at the time of treatments. There are different types of sensors are used in medical informatics. They can be categorized into four groups: physiological sensors, wearable activity sensors, human sensors, and contextual sensors (Sow, Turaga, & Schmidt, 2013). Physiological sensors such as ECG sensors, SpO2, temperature sensors in the intensive care unit (ICU) measure patient vital signs or physiological statistics. Wearable activity sensors measure attributes of gross user activity, different from narrowly focused vital sign sensors. Today, there are various wearable devices such as cell phones. Human sensors refer to human bodies for physicians examination, self-reporting, social media sharing. Examples of these sensors are care-provider notes, entries in the Electronic Medical Records (EMRs). Contextual sensors are embedded in the environment around the user to measure different contextual properties. For example, radio frequency identification (RFID) sensors linked with care providers, video and cameras.
Using data collected by sensors has various benefits for healthcare (Darwish & Hassanien, 2011). It facilitates the design of sophisticated clinical decision support systems capable of better observing patients’ physiological signals, helps provide situational awareness to the bedside, and promotes insight on the inefficiencies in the healthcare system that may be the root cause of surging costs. The rich sensor-generated data provides foundation for medical informatics that improves healthcare by acquiring and transmitting knowledge for application. Many sensor data mining techniques are applied in healthcare to gain business intelligence for further decision making.