- Probabilistic Sensitivity Analysis in Health Economics. Get a working paper (joint work with A. Philip Dawid). Published in Statistical Methods for Medical Research
- A decision-theoretic framework for the application of cost-effectiveness analysis in regulatory processes. Get a working paper (joint work with Pierluigi Russo - download from UCL research reports page). Published in Pharmacoeconomics
- Bayesian hierarchical model for the prediction of football results. Get a working paper (presented at the IWSM 2008 in Utrecht, Netherlands. Joint work with Marta Blangiardo). Published in the Journal of Applied Statistics
- Health economic evaluation of the treatment of osteoporosis in five Italian regions. Joint work with Gian Luca di Tanna, Fabio Pammolli and Davide Integlia. Get a poster presented at the European ISPOR 2009
- Regression discontinuity design in Epidemiology. Joint work in progress with Sara Geneletti, Irene Petersen, Irwin Nazareth, Richard Morris, Nick Freemantle, Linda Sharples, Sylvia Richardson and Philip Dawid. Get a presentation (of a very draft version!) at the London School of Hygiene and Tropical Medicine
- Health economic evaluation of HPV vaccination. Joint work with Francesco Mennini, Alessandro Capone, Andrea Marcellusi, Bengt Jonsson, Mike Drummond and Peter Zweifel. Get a presentation at the 8th World Congress in Health Economics (iHEA 2011, Toronto). Published ahead-of-print in Medical Care, and also see a post in the blog
- Bayesian hierarchical models to investigate the impact of telomere length in pre-implantation genetic screening in IVF. With Anastasia Mania and Siobhan Sen Gupta. Modelling using Integrated Nested Laplace Approximation. Get a presentation at the UCL Biostatistics Network Seminars and a post in the blog
- Is the UK really discriminated against in the Eurovision Song Contest? Bayesian hierarchical modelling of the votes. Joint work with Marta Blangiardo (see some rather confused and very preliminary discussion here and here and a working paper here). After the publication in the Journal of Applied Statistics, just before the 2014 Eurovision contest, we got media coverage, for example here, here, here, here, here and here.
- Spatial and spatio-temporal models with R-INLA. Review paper (with some new insights) - joint work with Marta Blangiardo, Michela Cameletti and Håvard Rue. Get a working paper and R scripts from here. Published in Spatial and Spatio-temporal Epidemiology (and available here)
- Bayesian models for cost-effectiveness analysis in the presence of structural zero costs. This work will extend the framework of hurdle models commonly recommended to tackle the issue of individual patients with observed zero costs, to include a full cost-effectiveness model, accounting for correlation between costs and a suitable measure of clinical effectiveness (eg QALYs). A working paper is available from the arxiv, while a final version has been published in Statistics in Medicine (it is available here, through OpenAccess). I have also written an R package implementing the framework presented in the paper, which is available here.
- Design and sample size calculations for trials based on the stepped wedge design. This is joint work with Rumana Omar, Gareth Ambler, Emma Beard, Andrew Copas and Michael King - it will be part of a special issue in the journal Trials (involving researchers at UCL and the London School of Hygiene and Tropical Medicine). We investigate the main design issues and the implications in terms of sample size calculations; we are also working on developing a R package to perform power analysis and sample size calculations (based on analytical formulae as well as simulations).
- Health economic evaluation of Human Papillomavirus vaccination. The project investigates statistical methods to model the population dynamics of vaccinations and their impact on the cost-effectiveness of different vaccination strategies against HPV. More information is available here.
- Interpreting lower urinary tract symptoms using Bayesian methods: improved diagnostic inference, faster delivery, less investigation and lower costs. This work investigates the use of Bayesian non parametric methods in model selection problems for risk/diagnosis prediction. The idea is to apply a "spike & slab" prior to model a set of potential factors that "explain" a given diagnosis (eg of lower urinary tract infection). The main methodological results have been published in Statistics in Medicine.
- Health economic evaluation of polycystic ovary syndrome (PCOS). This work will assess the prevalence of PCOS in the UK, modelling data from the relevant (and scarce) literature, as well as using an administrative database of the UK clinical practice. She will also complement her work by including an economic evaluation of potential interventions to reduce the impact of PCOS.
- Full Bayesian methods to model utility measures using mixture of distributions. This work will propose novel models to quantify the measures of utilities currently used in economic studies. In addition, it will address the issue of mapping across different utility scales, using Bayesian hierarchical models.
- A modelling framework for estimation of benefit using single arm clinical trials. This work will investigate statistical issues arising from single arm studies and their implications in terms of adoption in the pharmaceutical market (eg by the Food and Drug Administration, in the US, or the European Medicine Agency, in the EU). A paper describing the first results of this research has been published in BMJ Open.
- Bayesian computations for the Expected Value of Partial Information for health economic evaluation using Gaussian processes and INLA. This methodological work will investigate innovative methods for the quantification of the value of obtaining new evidence on some of the quantities involved in a decision model. This has the potential of impacting research prioritisation. A paper presenting our model is available from the arxiv (and has been published to Statistics in Medicine).
- Full Bayesian methods to handle missing data in health economic evaluations. In this work we will explore modelling missingness in health economic data. This is a slightly more complex issue, since the outcome is multivariate (comprising of a measure of cost and a measure of clinical benefit) and missing can occur in either or both variables. In addition to correlation between them, the missingness mechanisms may also be correlated, which implies a more complex modelling required.