Big-Data Techniques to Understand the Relationship Between Weather and Pain

  • Supervisors: Prof. David Schultz
    Prof. Will Dixon
    Dr. Thomas House

  • External Supervisors: Prof. Paul Roebber (University of Wisconsin—Milwaukee, United States)

  • Contact:

    Prof. David Schultz, david.schultz@manchester.ac.uk

  • CASE Partner: N/A

Application deadline: 30 May 2018

Introduction:

Can people with chronic pain predict the weather?  The Cloudy With a Chance of Pain project (http://www.cloudywithachanceofpain.com) was designed to collect health and weather data using smartphones from people suffering from chronic pain to determine if there is a relationship between changes in the weather and people's reporting of pain.  Over 13,000 joined the study, producing over 5 million reports.  Study participants entered ten different personal measures of pain, activity, sleep behavior, and activity for as long as 14 months.  Coupled with the GPS data from their smartphones, each entered report can be associated with corresponding weather data.  With the combined health and weather data, we can determine the relationship between weather and pain in more detail and with more confidence than has been previously done in the past, answering this age-old question.

Project Summary:

The overall aim of the project is to analyze these datasets to understand the factors that affect individual's sensitivity to the weather, if any.  As the data begins to be analyzed for this relationship, new techniques in data analysis and data mining will be needed to remove confounding effects from this large, multivariate dataset.  Preliminary results suggest that the outcome varies by different health conditions, but more sophisticated data-mining, machine-learning, and artificial-intelligence approaches would benefit this research. For example, modern data science approaches allow complex trends to be learned from data within a framework that flexibly incorporates expert meteorological and medical knowledge, rather than imposing, for instance, a simple linear relationship a priori.  Such modern approaches would therefore benefit such a large and diverse dataset as Cloudy With a Chance of Pain.  We will build a flexible mathematical representation of meteorological knowledge that would allow us to extract more from the data.

With those problems solved, we plan to produce an individualized pain forecast product.  Ideally, a person would enter their age, gender, medical condition, types of medication, etc., into an app, and, based on the forecast for the next week, the app would reveal the best days for physical activity or days to rest when the pain is expected to be most severe.

The project would be well suited to a student with a quantitative background in a physical science, mathematics, computer science, or engineering discipline. Under the supervision of the expert project team, the student working on this cross-disciplinary project will gain a wide breath of training in meteorology, health data, data-mining, machine-learning, and artificial-intelligence approaches.  Work in this topic will prepare the student for work in the exciting discipline of health informatics, the financial sector, weather and climate forecasting, or other big-data disciplines that would contribute to the data-driven economy.

References:

Beukenhorst, A. L., D. M. Schultz, J. McBeth, R. Lakshminarayana, J. C. Sergeant, and W. G. Dixon, 2017: Using smartphones for research outside clinical settings: How operating systems, app developers and user determine geolocation data quality in mHealth studies. MEDINFO 2017: Precision Healthcare through Informatics, A. V. Gundlapalli et al., Eds., IOS Press, 10–14, doi: 10.3233/978-1-61499-830-3-10.

Druce, K. L., J. McBeth, S. N. van der Veer, D. A. Selby, B. Vidgen, K. Georgatzis, B. Hellman, R. Lakshminarayana, A. Chowdhury, D. M. Schultz, C. Sanders, J. C. Sergeant, and W. G.  Dixon, 2017: Recruitment and ongoing engagement in a UK smartphone study examining the association between weather and pain: Cohort study.  J. Med. Internet Res. mHealth uHealth, 5(11): e168, doi:10.2196/mhealth.8162.

Reade, S., K. Spencer, J. C. Sergeant, M. Sperrin, D. M. Schultz, J. Ainsworth, R. Lakshminarayana, B. Hellman, B. James, J. McBeth, C. Sanders, and W. G. Dixon, 2017: Cloudy with a Chance of Pain:  Engagement and subsequent attrition of daily data entry in a smartphone pilot study tracking weather, disease severity, and physical activity in patients with rheumatoid arthritis. J. Med. Internet Res. mHealth uHealth, 5(3), e37, doi: 10.2196/mhealth.6496.

Apply Now