Professor Sue Mallett

• Honorary Professor in Diagnostic and Prognostic Medical Statistics (2020-
• Professor in Diagnostic and Prognostic Medical Statistics (2019-2020)
• Senior Lecturer in Medical Statistics, University of Birmingham (2015-2019)
• University Research Lectureship, University of Oxford (2011-2014)
• Fellow Royal Statistical Society 2010
• Diploma in Statistics, Open University, 2004
• DPhil Molecular Immunology, University of Oxford,  1991
• BA Hons Biochemistry, University of Oxford, 1988

Dr Sue Mallett is an Honorary Professor in diagnostic and prognostic medical statistics based in the Test Evaluation Research Group (TERG) led by Professor Jon Deeks at University of Birmingham. Her current work is both as a methodologist and clinical trial statistician, specialising in diagnosis and prognosis. Sue has dual skills training as a medical statistician and in laboratory research as a biochemist and molecular immunologist. 

Sue Mallett qualified with a B.A. Hons in biochemistry at Oxford University and a DPhil in molecular immunology on the CD4 T cell activation marker OX40. She worked for 6 years in immunology and virology, including 2 years at University of California San Francisco. 

Sue re-trained as a statistician through the Open University UK, whilst working in a range of jobs to build up her experience and portfolio in statistics and clinical trial design. She worked with Professor Sir Roy Anderson and Professor Neil Ferguson at the start up company Oxford Biologica Ltd on modelling of drug resistance in HIV and influenza. She joined the UK Cochrane Centre for a short project working with Professor Mike Clarke and Dr Sally Hopewell on understanding a typical Cochrane review and publication bias. She worked on cancer pathways in the NHS with the Oxford Cancer Intelligence Unit.

In 2002 she joined the Centre for Statistics in Medicine at Oxford University, working with Professor Doug Altman and Professor Jon Deeks on the methods, design and analysis of diagnostic accuracy studies and diagnostic accuracy systematic reviews.  She extended her research to include prognostic studies through collaborations with Professor Patrick Royston on a review of methods and on evidence underlying the REMARK reporting guidelines for tumour marker studies with Professor Willi Sauerbrei.

She built up long term collaborations with leading researchers in radiology through Professor Steve Halligan (UCL), Professor Stuart Taylor (UCL) and Professor Vicky Goh (Kings College London), including methodological and radiological research, and NIHR funded diagnostic accuracy studies.

Her collaborations have included a very successful NIHR programme grant led by Professor Steve Halligan included: methodological work on issues with ROC AUC in diagnostic imaging studies, the first studies using eye tracking in 3D medical imaging using CT colonography, support for FDA approval of computer assisted technology in CT colonography and decision choice experiments with patients and health professionals on diagnostic testing.

In 2011 Sue moved to the Department of Primary Care Health Sciences in Oxford where she taught on the MSc in Evidence Based Health Care, and developed collaborations including a systematic review of studies predicting risk of hospitalisation in at risk children with influenza.

In 2015 Sue moved to Test Evaluation Research Group (TERG) at University of Birmingham to work with Professor Jon Deeks. Her work includes leading statistical analysis on five NIHR funded diagnostic accuracy or prediction study clinical trials (STREAMLINE lung & STREAMLINE colon for cancer metastasis, METRIC for Crohn’s disease) and two NIHR prognostic studies (PROSPECT for colon cancer recurrence, ARCHIE: antibiotics for at risk children with influenza). Sue joins the statistical team on diagnostic trials ongoing in the Birmingham Clinical Trials Unit.

Sue has over 90 peer reviewed research publications, most in diagnosis and prognosis, including design, analysis and methodology research. She is a member of the Cochrane Diagnostic Test Accuracy Editorial team, regularly reviews Cochrane protocols and is on the Steering Committee for QUADAS-2 and QUADAS-2C (risk of bias assessment tool for diagnostic accuracy studies) and PROBAST (risk of bias assessment tool for prediction studies). 

Sue is a member of the steering committee member of the NIHR statistics group (since its foundation in 2011), a national group which represents and enables best practice for statistical researchers working on NIHR funded projects  (http://www.statistics-group.nihr.ac.uk/). Sue co-leads the NIHR Statistics Group section on Imaging Studies (2014- present) and the Diagnostic and Prognostic section (since June 2018).

• Public Health MPH/Diploma/Certificate, University of Birmingham (2015-)
• Medicine and Surgery MBChB, University of Birmingham (2015-)
• MSc Evidence Based Medicine, University of Oxford (2011-2015)

Sue is interested in supervising doctoral research students in the following areas

• Methods for evaluation of medical tests and biomarkers
• Risk of bias in diagnostic studies or prediction studies
• Non-randomised clinical trial design
• Methods for systematic reviews

If you are interested in studying any of these subject areas please contact Sue directly, or 
for any general doctoral research enquiries, please email mds-gradschool@contacts.bham.ac.uk

For a full list of available Doctoral Research opportunities, please visit our Doctoral Research programme listings.

Overview of current research

Sue Mallett's research is to design and analyse diagnostic and prognostic studies, including ongoing clinical trials and methodological work. She works closely with clinical teams to improve diagnostic pathways in the NHS, by designing studies and identifying best methods and evidence to inform NHS practice and future NICE guidelines.

In Sue's research she brings expertise from two disciplinary areas - medical statistics and biochemistry/ immunology/virology. This gives her an important understanding of clinical researcher goals, medical terminology, laboratory methods and underlying disease mechanisms.

Sue has developed new diagnostic clinical trial designs to extend beyond diagnostic accuracy and to include patient management decisions based on the diagnostic tests results. This is a new area enabling the linkage between diagnostic accuracy and health economics to be based, where feasible, on trial based evidence of the diagnostic test impact on patient management decisions. This contrasts to current methods where health economic models are based on assumptions informed by expert opinion on patient management resulting from diagnostic test results. Trials in Crohn’s disease, colorectal cancer and lung cancer using this design have produced valuable insights into both roles of diagnostic tests in patient management and also on trial design. 

In another important innovation in this area, Sue has worked with clinical colleagues to design and analyse diagnostic accuracy studies of whole pathways within the NHS. This has allowed novel outcomes from diagnostic accuracy studies to be obtained, such as the time to complete a diagnostic pathway and the number of tests needed to reach a diagnostic decision. 

Sue's objective in design of diagnostic trials is to compare two or more tests or test pathways against a reference standard, so that new tests and test combinations are compared against standard clinical practice.

Recent research projects:

• NIHR funded diagnostic accuracy study METRIC on the use of ultrasound and magnetic resonance imaging in the assessment of Crohn’s disease with Prof Stuart Taylor (UCL). This study provides evidence on the accuracy of non-ionising imaging for diagnosis and monitoring of Crohn’s disease. This study included novel research on patient management and consensus panel methods used as a reference standard.  This project includes cohort follow up to investigate baseline imaging features as a predictor of severe Crohn’s disease.
• PROBAST: A tool to assess risk of bias and applicability of prediction model studies. Tool and Explanation and Elaboration paper. Sue has been a leading member of the International Steering group for this is a new risk of bias tool for prediction model studies. This domain based tool is designed for use in systematic reviews but it can be used more generally in critical appraisal of prediction model studies. Potential users include organisations supporting decision making, researchers and clinicians with an interest in evidence-based medicine or involved in guideline development as well as journal editors and manuscript reviewers.
• STREAMLINE COLON and LUNG are two diagnostic accuracy trials of metastases following cancer diagnosis. These trials compare a pathway using whole body MRI to the standard NICE recommended pathway based on CT-PET. This study includes analysis of patient management decisions, time to diagnosis, number of tests in each pathway and patient costing. 

Other ongoing research projects include:

• Systematic reviews on prediction factors for development of severe Crohn’s disease in collaboration with Darren Boone and Steve Halligan at UCL.
• Systematic reviews of biomarkers for ovarian cancer, in collaboration with Sudha Sundar and Jon Deeks at University of Birmingham.
• Collaborative project with Professor Seena Fazel on clinical prediction rules for persons diagnosed with a first episode of severe mental illness for risk of violent crime and suicide
• Hernia research including methodological systematic reviews and analysis of cohort data, in collaboration with Sam Parker, Steve Halligan and Al Windsor at UCL.
• Improving the prediction of metastatic disease in primary colorectal cancer (PROSPECT) in collaboration with Vicky Goh at King’s College London.
• Impact of multiparametric MRI on staging and management decisions in women with ovarian cancer (MROC study) in collaboration with Andrea Rockall at Imperial College London.

Previous diagnostic research:

• NIHR programme of research focussed on CT colonography with Steve Halligan (UCL) and Doug Altman (Oxford) leading to over 20 publications.  This research has been influential in clinical guidelines on the diagnosis of patients with colorectal cancer, and is used as evidence to inform NICE guideline CG131 on the diagnosis and management of colorectal cancer, as well as the second ESGAR consensus statement on CT colonography. The research on a computer assisted detection device in radiology was key evidence in a Federal Drug Administration (FDA) Device approval. The group also collaborated with vision perception researchers in an innovative programme of work on eye tracking radiologists interpreting 3D videos of colonography, identifying key aspects influencing diagnostic accuracy in this increasingly important area of imaging.
• Post-mortem imaging as an alternative to autopsy in the diagnosis of adult deaths with Prof Ian Roberts (University of Oxford). This has been used as evidence to inform the Royal College of Radiologists/Royal College of Pathologist statement on standards for medico-legal post-mortem cross sectional imaging in adults. 

Methodological research in diagnosis has included:  

• Systematic reviews of diagnostic tests in cancer: Review of methods and reporting.
• Methodological work on measurements of diagnostic accuracy including the disadvantages of ROC AUC and longitudinal missing data methods
• Systematic review on bias in imaging studies - the effect of manipulating clinical context, recall bias and reporting intensity
• Systematic review of multi-reader multi-case studies using the area under the receiver operator curve as a measure of diagnostic accuracy 

 Prognostic research includes:

• Systematic review of prognostic studies including IPD MA: Children at risk of influenza-related complications in primary and ambulatory care: a systematic review of published and unpublished data
• Collaboration with Professor Carl Moons on checklist for data extraction for systematic reviews of prognostic studies
• Collaboration with Prof Willi Sauerbrei  on REMARK reporting guidelines 
• Methodology reviews of prediction modelling studies in cancer with Prof Doug Altman and Prof Patrick Royston.

Research groups and centres

• NIHR statistics group section on Imaging group and Diagnosis and Prognosis. Co-lead of sections. 
• Member of International Steering Committee for a risk of bias tool for prognostic modelling studies  (PROBAST: Prediction study Risk Of Bias Assessment Tool)
• Member of International Steering Committee for QUADAS2 and QUADAS-2C, risk of bias assessment tools for diagnostic accuracy studies.  QUADAS2 is used for NICE, HTA, The Cochrane Collaboration Diagnostic Test Accuracy reviews as well as other international Health Technology organisations developing health guidance on behalf of governments.  
• Member of Cochrane Diagnostic Test Accuracy Editorial Team (since 2015)
• Member, Cochrane Diagnostic Tests Accuracy Working Group Grant funding panel member for the NIHR Programme Grants for Applied Research grant review sub-panel (PGfAR) (since 2016)

• NIHR statistics group section on Imaging group and Diagnosis and Prognosis. Co-lead of sections. 
• Member of International Steering Committee for a risk of bias tool for prognostic modelling studies  (PROBAST: Prediction studyRisk Of BiasAssessment Tool)
• Member of International Steering Committee for QUADAS2 and QUADAS-2C, risk of bias assessment tools for diagnostic accuracy studies.  QUADAS2 is used for NICE, HTA, The Cochrane Collaboration Diagnostic Test Accuracy reviews as well as other international Health Technology organisations developing health guidance on behalf of governments.  
• Member of Cochrane Diagnostic Test Accuracy Editorial Team (since 2015)
• Member, Cochrane Diagnostic Tests Accuracy Working Group
• Grant funding panel member for the NIHR Programme Grants for Applied Research grant review sub-panel (PGfAR) (since 2016)