I came to the Univesity of Aberdeen in 2015 to join the team at the Centre for Health Data Science.
My first projects here were on the Aberdeen Children of the 1950s. The Children of the 1950s were the last generation assigned to secondary school based on an exam score from age 11. We collected 20 years of hospital admissions for 5000 study members to see if selective schooling effected long-term health. You can see the registered report of our regression discontinuity analysis here.
My current research uses health care records to study the diagnosis and treatment of fibromyalgia as part of the PACFIND project. We are linking 15 years of patients' clinic visits, hospital admissions, and prescribing data to detailed personal surveys. We are also trying to identify undiagnosed fibromyalgia patients from GP records. The goal of this work is to map the complicated health care journeys fibromyalgia patients have to find ways to improve treatment.
I’m originally from rural Pennsylvania in the United States. I have a PhD from the University of Massachusetts and a BSc from Brown University, with a background in genomics and bioinformatics. You can read my papers on electron transfer in bacteria here. Before I came to Aberdeen I spent a few years living in Oxford and Barcelona with a philosopher.
I'm very interested in open and reproducible research practice and in public engagement. I'm the local lead for the UK Reproducibility Network here in Aberdeen, and run the Open Research Working Group. All students and staff are welcome to join - just drop me an email.
A manifesto for reproducible science - good introduction to open science practice
Preregistration templates for studies
Registered Reports - a rigorous way to publish (get a paper accepted before you see the data)
FAIR Guiding Principles for scientific data management - basics of transparency
Good enough practices in scientific computing - project organisation for reproducibility
Blind analysis - to prevent bias
The Turing Way - guide to reproducible science (strong on code testing and review)
BiorXiv - preprint server for manuscripts
Stats and Errors
Statistical tests, P values, confidence intervals, and power - a guide to misinterpretations
Communicating results about treatment effects - very useful guide on reporting significance
Common statistical tests are linear models - elegant way of thinking about stats tests
STRATOS - guidance in the design and analysis of observational studies
Data Skills for Reproducible Science - course on data wrangling, iteration, simulation, workflows
Stats545 - another good course on data skills for reproducible science
R for Data Science - very useful textbook
R best practices - basic project setup, links to additional resources
Course Syllabi for Open and Reproducible Methods - excellent source for additional references
Reproducibility Bibliography - extensive references focused on guidelines and examples
ReproducibiliTea journal clubs around the world
Fibromyalgia is a common condition causing widespread chronic pain, associated with profound fatigue and marked sleep disturbance. Diagnosis and management are complex. People can wait up to ten years for a diagnosis, involving many general practice consultations and referral to a number to different specialists. While there is considerable evidence on effective treatments, most individuals with fibromyalgia are not receiving timely diagnosis or access to effective treatments. There is almost no evidence on how to organise health services to deliver recommended therapies. Patients feel dissatisfied with current services believing that no-one is willing to take responsibility for their care.
As part of the PACFiND project on Patient-centred Care of Fibromyalgia, we want to provide better services for people with fibromyalgia which ensure:
- Timely diagnosis
- Equitable access to effective treatments
- Higher levels of patient satisfaction
- Reduced costs to the NHS
Our work aims to provide an in-depth analysis of the patient journey and identify patterns of healthcare use using data linkage to map the patients’ healthcare journey. We aim to understand what happens to people who have fibromyalgia in terms of their consultations, referrals, time to diagnosis, and treatments received.
Aberdeen Children of the 1950s is an established population-based birth cohort that includes everyone born in Aberdeen (1950-1955) who went to primary school there; 12,150 people in total.
ACONF pioneered linkage to health and administrative data, starting in 1962 with linkages to the Maternal and Neonatal Databank, education, and census databases. ACONF now permanently holds a uniquely rich set of early-life data for people over 60, including:
- Health data from pregnancy, birth, and primary school records
- Cognitive testing at 7, 9, 11 and personality assessments
- Socioeconomic data for participants, parents and grandparents, including home value, crowding, amenities and home culture
- Social network data for friendship and families
- Qualitative interviews from 1960s and 1980s and recent in-depth life histories
Since 2015, we have successfully demonstrated comprehensive data linkages for ACONF, creating 20 years of virtual follow-up through project-specific linkages to:
- Hospital admissions
- Mental health admissions
- Maternity admissions
- Cancer registries
- Community prescribing
- In-patient and out-patient biochemistry
In the past fifteen years, the information collected on the Aberdeen Children of the 1950s has been used in many scientific studies with active collaborators across the UK, Hong Kong, Brazil, India, and Australia. The rise and decline of fishing and oil industries occurred during ACONF’s lives and provide important comparators for global development. Several studies have analysed the effects of being born into poverty and demonstrated the impact this has on our health throughout life. Other studies have looked into causes of common illnesses like heart disease and diabetes, and taught us how important health during pregnancy and early life are for our wellbeing in adulthood. Since 2015, we have worked with cohort participants, the NHS, and the local authority to prioritise research directions.
Our research on the Aberdeen Birth Cohorts is done through the Administrative Data Research Centre - Scotland.
We're also part of Health Data Research UK.
Our research on chronic pain is supported by a £1,200,000 grant from Arthritis Research UK
Our research on the Aberdeen Birth Cohorts is part of a £7,500,000 grant from the ESRC devoted to the analysis of large administrative data sets.
Our public engagement work is supported by the Gordon and Ena Baxter Foundation.
- Further Info
I'm the local lead for the UK Reproducibility Network at the University of Aberdeen.
I'm on the Athena SWAN self-assessment team, working for gender equality at the university.
Open Science Resources
8 Easy Steps to Open Science: An Annotated Reading List
ReproducibiliTea journal clubs around the world
Getting started with R
Stats545 - Exploratory Data Analysis
Data Visualization textbook by Kieran Healy