Novel metrics to quantify patterns of medication use in asthma

asthma

This project explores patterns of medication usage and adherence and their relationship with clinical outcomes in asthma.

Program Type: Masters/PhD

Research Group: Respiratory Data Analytics Team, Airway Physiology and Imaging Group 

Supervisors: Dr Cindy Thamrin, Prof Helen Reddel, Ms Jacqueline Huvanandana

Synopsis: Patterns of medication use are important contributors to asthma management, symptom control, and risk of exacerbations. Current methods to describe these patterns are limited to simple metrics such as averaged usage, which do not take into account the wide range of behaviour seen over time. Describing these patterns in more detail may be more reflective of important clinical outcomes. This project will involve the development and application of novel and established data analysis techniques to describe patterns in medication use in asthma patients, and relate them to patient outcomes. It represents a great opportunity to apply advanced data analytics approaches to answer clinically-relevant questions.

Research Plan: Using a range of medium to large datasets obtained from past asthma clinical trials, we will develop and apply novel metrics to describe adherence patterns, obtained from daily electronic monitoring of medication usage. Methods based on time series analyses, frequency domain methods, entropy and other feature extraction techniques will be explored. Data reduction and machine learning techniques will be applied to determine the presence of any distinct patterns of medication usage. Relationships with patient outcomes will then be explored using standard regression techniques as well as machine learning methods. 

Significance: The project findings will aid understanding of the complex relationship between medication usage and subsequent risk/outcomes in asthma. They will also help provide insight whether there are optimal patterns of medication usage, and equip healthcare providers in their conversations with patients in order to improve their adherence to medication.

Candidate: The ideal candidate will have a strong background in engineering/biomedical engineering/computer science, with an interest in making a difference in health and medical research. Applicants must already have been awarded a first class Honours degree in Engineering, Computer Science, or Advanced Science, or hold equivalent qualifications or relevant and substantial research experience in a relevant field. Experience in programming in Matlab/Python/R, and/or in machine learning methods, is an advantage.

Contact: Dr Cindy Thamrin cindy.thamrin@sydney.edu.au