Design, conduct and interpret epidemiological studies in large routine healthcare databases
This module introduces the particular challenges and opportunities of epidemiology in large routine health care databases, such as the Clinical Practice Research Datalink. Lectures and practical sessions will cover protocols and approvals, data management and longitudinal analysis, and interpretation of findings. Students will explore and analyse data using Stata software. Special topics include sample size considerations in very large databases, a view from the protocol reviewers’ perspective and an introduction to specialist study designs. Participants will apply their new knowledge and skills to a large database research project and produce a study protocol.
Tutors include Professor Richard Stevens, Chair of one of the CPRD's Research and Data Governance (RDG) Expert Review Committees (ERC) and member of the Central Advisory Committee (CAC), and Dr Clare Bankhead, Head of the Nuffield Department of Primary Care Health Sciences CPRD Group, who host the University’s access to CPRD. Module coordinators Dr Emily McFadden and Dr Margaret Smith both have extensive experience in the use of routine healthcare data for research. Emily is a Department Lecturer and senior statistical epidemiologist within the Centre for Evidence-Based Medicine. Margaret is a senior statistician and epidemiologist within Oxford’s CPRD research group. Both are members of (different!) CPRD RDG Expert Review Committees.
The last date for receipt of complete applications is 5pm Friday 16th December 2022. Regrettably, late applications cannot be accepted.
Aims of the module
This module aims to provide competence in the design, execution and interpretation of analyses of large databases of routinely collected data, such as the Clinical Practice Research Datalink (CPRD).
Previous statistical experience including the following is a pre-requisite for this module:
- Statistical concepts – p-values, confidence intervals, standard error
- Multivariate linear, logistic and Cox regression
- Competence in statistical software including Stata
Students who have taken the module Essential Medical Statistics will meet these requirements.
Key learning outcomes:
- Gain knowledge of examples of large routine healthcare databases in the UK that are available for epidemiological research.
- Understand the types of epidemiological questions and study designs where routinely collected data may be of use.
- Understand and gain experience, using Stata software, of key practical components involved in designing research in a database, with focus on the CPRD:
- Identify different populations for analysis and understand which are appropriate to address specific research questions
- Identify recorded demographic, clinical history and prescribing data and understand their potential limitations
- Gain specific skills in Stata for data extraction and analysis
- Gain familiarity of potential data linkages, and their advantages and disadvantages. Gain awareness of data security issues.
- Understand relevant methodological issues and possible solutions in routinely collected healthcare data, including: measurement error, missing data, addressing confounding, multiple testing and sample size calculations.
- Gain awareness of specialist design and analytical methods relevant to routinely collected data.
- Understand the strengths and weaknesses of routinely collected data and use this knowledge to interpret and critically appraise study findings.
- Gain the ability to plan a research protocol to answer questions in evidence based healthcare using routinely collected data.