We develop decision analytical models in R. We have experience with all sorts of models, ranging from simple decision trees and Markov models, to sophisticated individual patient simulations. We are also happy to help you figure out what kind of model best fits your decision problem, and/or implement more advanced analytical methods, such as value of information, or calibration. (Why R?)
Do you have an old, clunky decision model in MS Excel that could do with a turbo boost? Or do you want to upgrade a model and run computationally expensive analyses (like EVPPI)? We 'translate' your model into R, make it faster, easier to use, with a clear documentation, that ensures that it can be easily maintained and updated in the future.
Our particular area of expertise is in the creation of bespoke interactive user-interfaces for health economic models constructed in R and data dashboards. We work in short sprints and in close collaboration with you to develop interfaces that are intuitive and easy to use for your target audience.
The biggest barrier to adoption of R is the lack of familiarity of health economists, who have been working in Excel for years.
We provide short courses and workshops to get you up to speed with health economic modelling in R and to set up a modern, collaborative working environment. Previous courses included:
We adapt the sessions to meet the requirements of the group, and tailor the case studies to your interests - please get in touch to discuss your training needs.
Take a fresh, interactive approach to communicating the results of your decision model - R Shiny allows stakeholders to directly interact with the model and let them try their own scenarios.
sadm-mk2 is a simple prototype of a clean web-based user interface for a time-dependant Markov model. It highlights some of the advantages that R/shiny can offer.
The sadm-mk2 model is built in R and C++. It runs 1,000 PSA iterations in less than 4 seconds. This allows us to explore many different scenarios 'on the fly', e.g. during a meeting. It also enables running advanced value of information analyses - even value of sample information analyses (estimating the expected benefit from additional trials) are feasible.
Traditionally, decision modelling involved multiple applications: Stata for the survival analysis, Excel with VBA makros for the model, Word and Powerpoint for the reporting. With R, we can now facilitate an integrated analysis pipeline all the way through: figures in your report will be automatically updated if something in the survival model at the beginning is changed.
Web-based user-interfaces allow stakeholders to interact with decision models, even if they are not familiar with R or programming in general. They can tweak different parameters, input their assumptions, and explore their own scenarios. This helps to communicate the results of decision models transparently and more effectively than in a report.
"I am not sure I can think of anything that we could have done better in this project – to be honest it went remarkably smoothly from the outset and you were very responsive to all our queries. Given the overall outcome for [the client], the project was a barnstorming success – so thank you for your support!".
"I very much enjoyed the course and really appreciated the efforts of the leaders to look into questions/add in extra material in between sessions to make the course as tailored and relevant to us as possible - you definitely exceeded expectations in this respect. Thank you!"
"[The client] was very impressed and liked it ! There are a couple of things arising, one being that they liked it so much they were wondering about converting the other model(s) into a similar format. [...]"
"They were super impressed !"
"Really accessible, loved the format of the website with practice questions. This made it far less daunting for a complete beginner like myself".
"The examples were very good to work through. Clear explanation. Lots of chances to ask questions. Really very well run".
Paul is a health scientist based in Bochum, Germany. Trained as a medical doctor and epidemiologist, he combines a deep domain knowledge with strong technical skills in data science, programming, and decision modelling. His main research interest is currently in the valuation of health benefits for economic evaluations and the interpersonal comparability of preferences. He is a keen advocate of open science email@example.com
Rob is a health economist based in Sheffield, UK. His research focuses on the methods used to estimate the costs and benefits of public health interventions, with a specific interest in microsimulation modelling in R. He is a expert advisor in Public Health Economics & Decision Science to the WHO-HEAT project. He is currently working at the Joint Biosecurity Centre to help inform the UK government response to the firstname.lastname@example.org
Sarah is a health economist with a background in health psychology based in Sheffield, UK . Her research focuses on the inclusion of psychological indicators within complex health economic models. She has experience working with a wide range of stakeholders and building complex microsimulation models in R.email@example.com
Shang joined Dark Peak Analytics in 2022. She has a MSc in Health Economics from the University of York, and has previously worked as a Health Economics Consultant for Amaris Consulting in Shanghai, MSD in Budapest and The University of Sheffield. Shang is currently building an Agent Based Model to better understand the role of social networks on alcohol drinking firstname.lastname@example.org
Wael is a health economist with a background in pharmacy and public health. In addition, he has experience in decision-analytic modelling, econometrics, and data science. He is currently pursuing a PhD in public health, economics and decision science at the University of Sheffield.email@example.com