In addition to the five research programmes, to optimally use the available methodological expertise, we have 3 methodlines that cover all programmes for optimal cross-fertilization.
1. Simulation modelling
The general channel includes the schedule of ClubMeth Simulation meetings, slides from previous presentations, and general documents. The five other content channels cover specific themes – individual-based modelling of life histories, individual-based modelling of populations, model calibration, programming for simulation modelling, and stakeholder involvement for model development – and include beginner and advanced level content materials for consultation
- Individual-based modelling of life histories
- In individual-based models (IBM) of life histories (also microsimulation models), persons are modelled one-by-one, and events in their lives are simulated. Usually, events like having an onset of a disease, natural progression of the diseases in various stages, getting an intervention (for example a screening test), diagnosis, treatment and death are modelled. Unobservable parameters are calibrated to observed data. These models are used to extrapolate the results of clinical trials to the population level, calculate the cost-effectiveness of interventions, and to evaluate unobservable events.
- Individual-based modelling of populations
- IBM of populations simulate dynamic populations as a whole and can capture interactions between simulated individuals. These models are commonly applied to study the spread and control of infectious diseases such as HIV, TB, Covid-19, and Neglected Tropical Diseases (NTDs). Simulated interactions concern contacts between individuals (e.g. through a dynamic sexual network) or capture vector-host dynamics (e.g. mosquito-borne infections). Models are calibrated against local data, and are used to inform policymakers on optimal prevention, control, and treatment strategies.
- Model calibration
- Simulation models are composed of parameters that drive the model outputs. To allow extrapolation of the results of clinical trials and project disease epidemiology, models first need to be tuned so that their outputs reproduce observed data. This channel provides guidelines for model calibration: the process of finding the right combination of model parameters to produce outputs consistent with observed data. However, model calibration is broader than just parameter fitting. Therefore, this channel also includes information on making choices on constrictions on the range of parameter values, which components of the observed data to reproduce, and trade-offs in which components of the observed data to prioritise in reproducing.
- Programming for simulation modelling
- We have experience with individual-based simulation (or microsimulation) which goes back for some 40 years. Currently, we mostly use Python, Java and R as programming languages. We are continuously working to strive for professional quality of our projects based on simulation modelling. This means explicit attention is given to systematic unit testing of programming code, as well as a consistent setup and coding style between projects. Also, we aim for documentation driven development, thereby ensuring that code and project documentation gets the attention that it deserves.
- Stakeholder involvement for model development
- Stakeholder involvement helps ensure that models reflect the realities of disease, healthcare delivery, and public health programmes. By working closely with experts, healthcare professionals, policymakers, and affected communities, we identify key processes, assumptions, and uncertainties that shape model outcomes. This improves model quality, transparency, and relevance while strengthening trust and support among stakeholders. Current methodological developments include group model building and structured expert elicitation to support more systematic model development.
2. Health Interventions
- Implementation and evaluation
- Public and patient involvement
- Co-design
- Qualitative research
3. Quantitative Data Analysis
- Prediction modeling
- Prediction models are statistical tools to estimate diagnoses, health risks, and future outcomes using existing data. These models can support clinical decisions, guide treatment strategies, identify individuals who may benefit from preventive interventions, and inform public health policy. This theme focuses on developing, validating, and implementing models that are accurate, generalisable, and relevant in practice. It also emphasises interpretability, stakeholder involvement, and the need to ensure prediction models contribute to equitable health outcomes.
- Causal inference
- Causal inference develops methods to determine whether observed relationships reflect true cause-and-effect mechanisms. It draws on evidence from both experimental and observational studies, using approaches such as causal diagrams, target trial emulation, quasi-experimental designs, mediation analysis, and g-methods. By moving beyond simple associations, causal inference provides robust evidence on the effectiveness of interventions, supporting public health policy, healthcare decision making, and efforts to reduce health inequalities.
- Cost-effectiveness analysis
- Cost-effectiveness analysis evaluates the costs and health outcomes of alternative interventions, treatments, and policies. It includes methods such as cost-effectiveness, cost-utility, cost-minimisation, and budget impact analyses, using data from clinical studies and decision-analytic models. The theme also covers cost and cost-of-illness analyses that quantify the economic burden associated with diseases and treatments. By assessing value for money and the efficient use of limited resources, cost-effectiveness analysis supports evidence-informed decisions in healthcare and public health.
- General statistics
- General statistics provides the methodological foundation for quantitative research. The theme covers statistical methods used throughout the research process, from study design and data exploration to modelling, inference, and reporting. Topics include descriptive statistics, regression methods, multivariable analysis, missing data, longitudinal and clustered data analysis, and Bayesian approaches. By supporting the appropriate application and interpretation of statistical methods, general statistics helps ensure that research findings are robust, reliable, and scientifically meaningful.
- Programming for data analysis
- Programming for data analysis focuses on the computational tools and workflows needed to manage, analyse, and interpret research data. The theme includes statistical programming in R, Python, and SPSS, reproducible research practices, version control with Git, data visualisation, and interactive applications. By promoting efficient, transparent, and reproducible analyses, programming supports the implementation of advanced quantitative methods and the generation of reliable evidence across a wide range of public health and healthcare research.