Research Summary

The Quantified Patient in the The Doctor’s Office

PhD Research Project by Peter West, University of Southampton.

Supervised by Richard Giordano (University of Southampton), Max Van Kleek (University of Oxford), and Mark Weal (University of Southampton).

Apps and wearables are promising new data sources for tailoring healthcare towards individual patients. But how can doctors make use of these data? This research aims to identify the challenges of using data generated by consumer health technologies and how these challenge could be overcome by including doctors in the design of appropriate clinical information systems.

Fitbit and Apple Health are two popular consumer technologies amongst a growing plethora of wearables and smartphone apps for recording diverse aspects of health. From heart rate and physical activity, to sleep and mood, these devices collect data with potential to help clinicians diagnose disease, personalise treatments to individual patients, and avoid delivering unnecessary medical procedures.

While medical records only store information about small snapshots of patients' lives within clinical settings, consumer health technology could provide intimate details of patients' routines and habits outside of clinical settings.

To tailor health and care towards patients it is important to know what happens in between clinical visits. These technologies could help fill these gaps between clinical visits. Healthcare could therefore benefit from utilising these data to reduce costs associated with monitoring chronic illness. Realising this potential is vital as we enter an era of ageing population and rising healthcare costs.

The Challenges of Consumer Health Technology

There are three main challenges to using consumer health technology acknowledged in wider academic literature:

Doctors do not have sufficient time to use the data. General practitioners, for example, usually have around 15 minutes per patient, of which much of that time is spent talking with the patient. Interpreting and analysing data could take more time than a doctor has.

Consumer health devices are typically not clinically proven. They have not been the focus of any long-term studies to demonstrate their effectiveness or accuracy. We therefore cannot be sure if the data they generate are suitable for making decisions about a patient. If data are inaccurate, a decision based off the data could harm the patient.

It is difficult to identity which data are relevant. Consumer health devices collect a lot of data, including kinds of data which are unusual in a clinical context, such as caffeine consumption. It is not always obvious which kinds of data will be most relevant to the patient’s current condition.

These challenges have been faced before. Electronic medical records have had significant challenges to their adoption within clinics because they disrupted workflows. By working with clinicians in designing electronic medical records, many of these challenges have been overcome.

How can design address the challenges of consumer health technology?

Possibly the most important strategy to addressing the challenges of consumer health technology is to include those who will be using their data: clinicians. Clinicians are the people who meet the patients and who the patients will ultimately hand their data to. It may therefore be their role to interpret and analyse patient-generated data.

A lot was learned about electronic medical records when they were deployed in clinical settings because clinicians provided their feedback on how electronic medical records could be used, and the problems they encountered.

For my research, I conducted interviews and workshops with 13 clinicians with various roles, including cardiology, general practice, and mental health. In the workshops, we co-designed a software-based tool for using self-tracked data within the management of heart conditions.

The prototype addressed the main concerns of consumer health technology:

Concern: Doctors do not have sufficient time to use the data.

Design response: The tool visualises information in a visual and summarised form, allowing a clinician to quickly identify salient points. The tool allows ‘zooming in’ on specific areas to allow deeper analysis (Figure 2b).

Concern: Consumer health devices are typically not clinically proven.

Design response: An ‘audit’ tool is included in the prototype, which allows a clinician to quickly check if the device the patient used has been clinically evaluated and whether the device was used correctly by the patient (Figure 2a).

Concern: It is difficult to identity which data are relevant.

Design response: The prototype has an ‘investigation’ mode which uses an algorithm to identify which data are likely to be relevant to the patient’s current condition (Figure 2c).

The next stage of this research will be to study the use of the prototype within clinician-patient interactions to see how it may disrupt practice.

Publications

Manuscripts:

West P, Van Kleek M, Giordano R, Weal M, and Shadbolt N (2017) Information quality challenges of patient-generated data in clinical practice. Frontiers Public Health Vol. 5 p. 284.

West P, Van Kleek M, Giordano R, and Weal M (2018) Common barriers to the use of patient-generated data across clinical settings. The Conference on Human Factors in Computing Systems.

Talks:

West P (2016) King’s Fund

West P (2017a) Health self-tracking: How can doctors use your data? Presented at NExT++ Workshop: Artificial Intelligence Solutions to Information Rich Open Problems in The National Supercomputing Center, Wuxi, China on 26–27 Sep.

Posters:

West P, Giordano R, and Van Kleek M (2015) The Quantified Self and clinical decision making: understanding clinical decision bias and errors when using quantified self data. Presented at Health Services Research

Network Symposium in Nottingham, UK on 13–14 Jul. West P (2017b) Self-tracked data in clinical settings. Presented at WSTNet Web Science Summer School in Saint Petersburg, Russia on 1–7 Jul.

West P (2017a) Health self-tracking: How can doctors use your data? Presented at NExT++ Workshop: Artificial Intelligence Solutions to Information Rich Open Problems in The National Supercomputing Center, Wuxi, China on 26–27 Sep.