The Quantified Self and Clinical Decision Making

This project builds on the findings of Mr Peter West's ECS Web Science DTC MSc Project, Finding Meaning in Personal Data: The Dangers of Data without Context which was supervised by Giordano and Van Kleek.

An increasing number of people are using pervasive and embedded technologies, such as mobile phones and wearable/implanted sensors, to record data about their daily lives, and store such data in the cloud. This may be referred to as the Quantified Self. Data collected by such technologies may be useful to healthcare practitioners for providing information about both individual patient and population health, and for helping healthcare practitioners—such as GPs—make clinical decisions by providing insight into behaviours that may be a contributing cause of morbidity. Although the potential for such devices and the data they provide to aid clinical decision making is large, there are real dangers. Clinical decisionmaking in GP settings, for example, is often based on heuristics and pattern recognition owing to the large flow of patients and pressures on available time. Research in both cognitive science and behavioural economics strongly suggest that heuristic decision making is plagued by bias and error. This can be especially problematic in health care settings where health care practitioners have no training or experience using Quantified Self Data or Open Data to make decisions, and where errors in diagnoses lead to increased morbidities and avoidable mortality. The quality of decision making in such settings may therefore deteriorate with the introduction of Quantified Self Data.

The MSc project—exploratory in nature—explored incorrect or biased inference in data collected from embedded devices; the effect on decisions due to bias; and the effect of providing context with the datasets by providing relevant and/or irrelevant data. The aims of this project are:

• Build a body of evidence at a larger and more diverse scale with input from primary care providers in a range of settings and contexts;
• Enquire from health care, organizational, web science, and sociological perspectives;
• Gather evidence on the education/development needs of health care practitioners regarding decision making using quantified self data;
• Link with industrial partners who can inform new products and services;
• Develop a set of data-based hypotheses that can be tested in subsequent research.

The Quantified Self Paper has received and Honourable Mention Award following its submission to the ACM CHI 2016 Conference. Out of 2325 paper submissions, it is one of 92 that has received an Honourable Mention Award. These papers were selected by the CHI2016 programme committee as well as a special Best Paper Committee and represent some of the best work at CHI.