Authors: Alec Davies*, , Mark Green, University of Liverpool , Alex Singleton, University of Liverpool
Topics: Medical and Health Geography, Quantitative Methods
Keywords: Self-medication, Data Mining, Health Geography, Geographic Data Science
Session Type: Paper
Start / End Time: 3:20 PM / 5:00 PM
Room: Bayside A, Sheraton, 4th Floor
Geodemographic information allows effective exploration of at risk populations. Traditionally national level data comes via decennial census. Although detailed, census data are an outdated snapshot of time. New forms of data allow highly detailed data over a broad time frame aiding rational health planning. Self-medication involves an individual self-prescribing medication purchased over-the-counter to treat an ailment. Self-medicating patients assume greater responsibility for the management of minor ailments as they self-diagnose and select products without prescription. Part of the self-care movement, this hybridised process can involve advice from health care professionals or services such as WebMD. Accurate self-medication can relieve the burden on health practitioners and reduce health care costs whilst increasing patient knowledge and ability. This study uses transaction level loyalty card data from a high street retailer to investigate factors influencing self-medication. Exploratory analysis investigates various products groups within the new form of data. Binary logistic regression is used to explore factors that may influence self-medication. As the new form of data is resultant of big data, machine learning techniques are used. Extreme gradient boosting, a method using greedy approximation is applied. Results are compared to investigate inference. The data brings innovative application to an important area in the evolution of medication. Accurate self-medictaion can significantly decrease the burden on the NHS within the UK, with potential to dramatically reduce prescription costs. The data bring original application to investigate the phenomena, with results showing geodemographic influencers of self-medication products.