The HMRI Clinical Research Design and Statistical Services offer all medical researchers and PhD students free initial consultations to discuss your grant proposal.
Services available to researchers:
- Study design including use of directed acyclic graphs to help identify sources of confounding;
- Sample size and power calculations
- Statistical analysis plans
- Advice on the treatment of missing data
- Use of statistical software including Stata, SPSS, SAS, JMP, GraphPad, R, and WinBUGS.
- Estimation of budget line items for analysis
Who can the Research Design and Statistics Unit help?
The support services are accessible to all researchers, registrars, and staff of the University of Newcastle, Hunter New England Local Health District and HMRI, with each HMRI affiliated researcher provided with an initial consultation of 1 hour free of charge. In addition, support for grant applications is also provided free on the understanding that a budget for statistical analysis will be included in the grant, and if successful, the work will be brought to the unit. Further work is charged at rates described below.
The unit will also undertake epidemiological and statistical support for biomedical researchers external to the University of Newcastle, the Hunter New England Local Health District or HMRI, although this is at a different fee structure.
Research Design and Statistics aim to provide comprehensive epidemiological and statistical support to biomedical researchers.
We encourage researchers to contact us at the initial stage of grant applications for help with:
- Study design
- Choice of instruments
- IT needs, e.g. web-based randomisation and data capture, web-based educational tools etc
- Statistical analysis plans
- Sample size calculation
PhD project planning
We encourage students to get advice on design and analysis at the planning stage of their project. Students are generally expected to run their own analyses, but CReDITSS provides a package for guidance and advice (see fee structure below).
We also provide data analysis services under a flexible model. We can provide advice and guidance if staff and researchers want to perform the analysis themselves with the oversight of a statistician, or we can perform the analysis ourselves according to a statistical analysis plan (derived in consultation with the researchers) and provide a full report. This also includes helping with the drafting of the manuscript and responding to reviewers.
Data Storage and Integrity
In the event that we perform the analysis we also provide a repository of data in case of future queries, i.e. the original dataset, cleaned dataset, syntax and command files, are all saved in case any questions of data integrity or version control are raised in the future. All files are on a password protected server in HMRI with regular backup.
Our IT unit provides a web-based randomisation service with the option of web-based data entry that can be customised to suit the needs of each project. This can handle varying block sizes, stratification, and minimisation.
The IT Team can help you customise and manage REDCap software to suit your research needs.
REDCap features include:
- Build online surveys and databases quickly and securely- Create and design your project rapidly using secure web authentication from your browser. No extra software is required.
- Fast and flexible - Conception to production-level survey/database for small projects in less than one day, and for larger or more complex databases in a few days.
- Export data to common data analysis packages - Export your data to Microsoft Excel, PDF, SAS, Stata, R, or SPSS for analysis.
- Ad Hoc Reporting - Create custom queries for generating reports to view or download.
- Scheduling - Utilize a built-in project calendar and scheduling module for organizing your events and appointments.
- Easily manage a contact list of survey respondents or create a simple survey link - Build a list of email contacts, create custom email invitations, and track who responds, or you may also create a single survey link to email out or post on a website.
- Advanced features - Auto-validation, calculated fields, file uploading, branching/skip logic, and survey stop actions.
- REDCap API - Have external applications connect to REDCap remotely in a programmatic or automated fashion.
- Data Queries - Document the process of resolving data issues using the Data Resolution Workflow module.
- Piping - Inject previously collected data values into question labels, survey invitation emails, etc. to provide a more customized experience.
- Multi-site access - REDCap databases/surveys can be used by researchers from multiple sites and institutions
The last few years has seen many methodological innovations in epidemiological and statistical methods. The CReDITSS Unit offers help in the following areas:
- Directed acyclic graphs to help identify sources of confounding; this includes both cross-sectional designs and longitudinal designs with time varying confounders.
- Various statistical techniques:
- T-tests /chi square tests/ANOVA (or non parametric equivalent)
- Linear regression
- Logistic regression/ discriminant analysis
- Poisson/negative binomial regression
- Survival analysis
- Exploratory factor analysis
- Dependent/correlated data such as longitudinal or clustered (modelled with mixed models or GEEs)
- Competing risk survival analysis
- Predictive modelling with internal validation
- Structural equation modelling
- Latent class analysis/regression
- Linked data
- Bayesian analysis
- Mediation analysis, calculating direct and indirect effects, using marginal structural models
- Treatment of missing data, e.g. multiple imputation, pattern mixture models
- Genome-wide association studies, including imputation to HapMap and 1000 Genomes, and principal components analysis for control of population stratification
Making an appointment
The unit’s business manager, Michelle Gillam, is the first point of contact. She will book a time for the initial consultation. Please contact Michelle using the contact form below.
It is a good idea to come prepared with a clear idea of your aims, hypotheses, primary outcomes, and budget.
The researchers meet with the senior epidemiologist, or one of the senior statisticians, or both, to discuss the project, the main research questions and to put together the outline of the statistical analysis plan.
Based on that plan, a written quote is put together with the main analyses discussed and an estimate of the amount of time anticipated to deliver each analysis.
This quote is sent to the researcher for verification and if confirmed, the project is assigned to a statistician in the unit.
Depending on the workload of the unit there may be varying timelines for the delivery of the results and this will be discussed with the researchers.
Statistical analyses are written up in a report that includes the methods, the results, and an interpretation. We will routinely meet with researchers to take them through the results face-to-face and answer any questions.
There is some leeway to modify and clarify various analyses but new directions and new analyses will need to be costed and quoted separately.
The quote includes time to provide feedback on the manuscript, assist in writing up the statistical methods, results and discussion, and respond to reviewers.
We subscribe to the authorship guide suggested by the International Committee of Medical Journal Editors (ICMJE). The ICMJE recommends that authorship be based on the following 4 criteria:
Substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work; AND
Drafting the work or revising it critically for important intellectual content; AND
Final approval of the version to be published; AND
Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Where these criteria are met we expect that unit staff will be included as authors on any publications resulting from their analyses and input
Statistical Packages Used
Statistical software used in the unit includes: Stata, SAS, JMP, R, and WinBUGS.
The unit does not routinely use SPSS and GraphPad so support for these packages is limited.