The Centre for Health Informatics is seeking a research associate in health data science to drive innovation at the interface between biostatistics, machine learning, and software engineering. Working with clinical, IT and healthcare data experts you will drive the modelling of health data using state-of-the-art methods from biostatistics and machine learning.
Ideally, you will have previous relevant experience in academia or industry encompassing a strong mathematical element and significant statistical knowledge. Ideally, you will be educated to PhD level in Statistics, Machine Learning, or related fields; have experience in the analysis of complex healthcare problems and datasets; and have significant statistical modelling and programming experience. If you do not have a PhD you will be considered if you have significant demonstrable research and/or industry experience and are in possession of a relevant higher degree. Domain knowledge of statistical and machine learning analyses of health, biological or social datasets; epidemiological models and complex causal inference; and multi-level regression are highly sought.
Safe Prescribing of Medication
Certain medications may be contra-indicated in given populations, or may require high levels of patient monitoring when used. Examples of such hazards: prescribing beta blockers to patients with asthma, or failure to test thyroid function regularly in patients receiving amiodarone. Using linked primary and secondary data, we have developed a tool to identify such medication safety hazards. We would like to evaluate the potential impact of such hazards: i.e. if the rate of such hazards were reduced, how many medication-related hospital admissions could be prevented? We are also developing and evaluating interventions (such as dashboards) aiming to reduce the prevalence of these hazards.
Dynamic and Longitudinal Approaches to Predictive Modelling
Clinical predictive models are used across healthcare to aid in decision making, planning and audit; these are often based on simple logistic regression models using information about a patient at a fixed point in time. We are developing methodology to exploit the richer sources of data that are now available to us. For example:
- Utilising longitudinal biomarker and risk factor information: can a model be improved by using the full history of risk factor changes over time?
- Responding to emerging data in an on-line fashion: as health data becomes more ‘connected’ can our models respond dynamically to emerging trends in outcome rates, policy changes or secular trends?
- Multiple outcomes and comorbidities: can we build joint, integrated models that consider multiple outcomes simultaneously (e.g. stroke, heart attack, onset of diabetes)?
The School of Health Sciences is committed to promoting equality and diversity, including the Athena SWAN charter for promoting women’s careers in STEMM subjects (science, technology, engineering, mathematics and medicine) in higher education. The [School/Institute] received a Bronze Award in 2013 for their commitment to the representation of women in the workplace and we particularly welcome applications from women for this post. Appointment will always be made on merit. For further information, please visit http://www.mhs.manchester.ac.uk/about-us/athena/.
Please note that we are unable to respond to enquiries, accept CV's or applications from Recruitment Agencies
Enquiries about the vacancy, shortlisting and interviews:
Matthew Sperrin, Lecturer in Health Data Science
Tel: 0161 306 7629
Tel: 0161 275 4499
Tel: 01565 818 234
This vacancy will close for applications at midnight on the closing date.