Provenance framework for mHealth

Abstract

Abstract-Mobile health technologies allow patients to collect their health information outside the hospital and share this information with others. But how can data consumers know whether to trust the sensor-collected and human-entered data they receive? Data consumers might be able to verify the accuracy and authenticity of the data if they have information about its origin and about changes made to it, i.e., the provenance of the data. We propose a provenance framework for mHealth devices, to collect and share provenance metadata and help the data consumer verify whether certain provenance properties are satisfied by the data they receive. This paper describes the programming model for this framework, which describes the rules to be implemented for providing provenance-collecting capabilities to an mHealth application. I. PROVENANCE IN MHEALTHConsider Jane, who is using one or more mobile health (mHealth) devices, continuously or periodically collecting her health-related information into her mobile phone. The phone periodically uploads this information, along with other healthrelated information that Jane manually inputs to her phone, to her electronic health record (EHR). Jane can then share her health information with her health providers, family and friends, peers, employer, insurer, and researchers. But how can these data consumers know whether to trust the sensorcollected and human-entered data they receive? What confidence do they have that it is accurate and authentic? Suppose Jane falls sick, and her husband Jack (her caregiver) collects her health information on her behalf. Jack has no experience using the devices. Can the system help him understand when he is not collecting accurate information?Since personal and home-use devices are not maintained regularly, like hospital devices, users like Jane and Jack might not realize when the devices are malfunctioning or uncalibrated. If other people in Jane's household use similar devices, they might accidentally confuse their own device for Jane's. mHealth devices, being mobile, can be stolen or misplaced -and then used by another person. Jane, or her caregivers, might deliberately fake or hide data collected using mHealth devices, when there is an incentive to do so. In all these scenarios, the data collected and shared by the devices might be inaccurate, or about the wrong person. If so, its use could prove damaging or even fatal, especially if the data is used by health providers for diagnosis or treatment.Data consumers might be able to verify the accuracy and authenticity of the data if they have information about its origin and about changes made to it, i.e., the provenance of the data. Previous research has looked at provenance of electronic health records and our work complements this research [1], [2]. In the case of health data that is collected using mobile sensors, it is necessary to collect information about the data's origin, so that consumers can determine the accuracy and authenticity of the data; our framework provides these capabilities, ...

Privacy and Security in Mobile Health: A Research Agenda

“…For a blood-pressure reading, for example, it is important to know whether the subject applied the cuff correctly to her arm, rested her arm on a flat surface, and remained still throughout the reading. Aarathi Prasad and her colleagues proposed one approach to the specification and collection of contextual evidence for mHealth sensor data, 14 but much more needs to be done to recognize the many factors that affect the quality of such data. 15

Research challenge: develop extensible methods for collecting, storing, and presenting contextual information along with health-related data collected by mHealth devices and apps to help data consumers verify and interpret the health data.

…”

Addressing biosignal data sharing security issues with robust watermarking

“…The framework provides mechanisms based on Attribute Based Encrypyion (ABE) for scalable key management, flexible access for multiple classes of data consumers, and efficient user revocation. In [14], the authors propose a detailed data provenance framework to collect and share provenance metadata for patient health records, to help data consumers verify the accuracy and authenticity of the data and track its origins and changes made to it. However, the framework does not define and address a threat model.…”

Provenance Data Management in Health Information Systems: A Systematic Literature Review

“…All 17 studies were found in the SLR process together with the snowball technique, which are [ 90 , 91 , 92 , 93 , 94 , 95 , 96 , 97 , 98 , 99 , 100 , 101 , 102 , 103 , 104 , 105 , 106 ] and are focused on managing provenance data in HISs. In Table 10 , the similarities of the 17 studies are observed in the following characteristics that contribute to the management of provenance data in HISs: use of models from the W3C PROV family; use of different models from the W3C PROV family; use of provenance techniques with blockchain; and use of provenance techniques with middleware.…”

“…It is important to emphasize that the types of HISs shown in Figure 5 have different characteristics, as detailed in this study. Moreover, based on the information in Figure 5 , the authors who mentioned the types of HISs in their studies in relation to the management of provenance data in the selected studies are as follows: EHRs [ 93 , 94 , 95 , 96 , 103 , 106 ]; PHRs [ 90 , 97 , 99 , 100 , 101 , 102 , 104 ]; the LHS [ 98 ]; HMSs [ 92 ]; the CRIS [ 91 ]; and the HIS [ 105 ].…”