Many different types of data will be collected in PAINSTORM, in order to try and better understand the complex nature of neuropathic pain. This research theme brings together the information generated in the rest of the project and analyses it together to look for correlations, causes, and effects. Using statistics, we can look for the factors that best explain neuropathic pain. We will also focus on tools that are easy to use in a clinical setting.
Taking data generated through Work Packages 1 to 6, we will use sophisticated statistical and computational approaches to model and quantify associations between a wide range of factors associated with neuropathic pain, testing their strength and validity. We will adopt two distinct approaches:
- Examining all available factors, with a view to understanding the pathophysiology of neuropathic pain;
- Focusing on factors and techniques with the greatest clinical utility, with a view to applying the findings in real-world medical practice.
Ultimately we will develop composite biomarker signatures for neuropathic pain, enabling greater understanding of neuropathic pain as well as individualised assessment, and therefore stratified treatment or prevention approaches.
Ranking clinical, biophysical, demographic, genetic and self-reported psychosocial measures
Exploratory hierarchical modelling will employ all available data to elucidate dependence and inter-dependence patterns among study variables. Features will be constructed by grouping original variables into larger mechanistic concepts (e.g. peripheral nerve fibre integrity, central processing). Hypothesis testing will determine which features are significantly different between patient subgroups and model agnostic feature selection, encapsulated in cross-validation, to determine the most powerful predictors in the context of predictive models.
Identifying patient subgroups sharing common traits to improve stratification
Based on high bivariate correlations, more complex multivariate factor and correspondence analyses will be investigated to uncover latent variables describing a common trait. After further validation, a structural equation model will be used to align the latent variables arising from each group into a single, unified model, describing input of various sources from stimulus perception along the nervous pathway to central processing. This methodology has been developed for immunology and we have started to apply it to multi-dimensional datasets relating to pain. In a separate analysis, we will use a fully data driven approach to identify interrelated factors arising from Work Packages 1 to 6, and common patterns. Here, we will use algorithmic clustering and machine learning.
A cross-validated predictive model for the patient subgroups
Finally, we will exploit the large, phenotypically harmonised datasets with longitudinal outcomes to build models that maximise predictive accuracy on the pseudo-independent validation set. Cross-validation will fine tune the parameters of an array of machine learning classification algorithms. We will include a burden/benefit ratio for included variables as identified with patient partners in Work Package 2 and the developed model will be informed by and cross-validated with the directed acyclic graph models developed in Work Package 3.
This report gives a brief overview of the activities and outputs of Work Package 7 within the PAINSTORM consortium, focusing on defining key research questions, discuss and align on data analysis and modelling methods and harmonise datasets across the consortium.
Core Research Questions
We held monthly meetings throughout 2024 – 2025 in close collaborations with WP3. Members of WP7 and WP3, people with lived experience, and representative WP leads across the consortium participated and addressed several overarching and work-package-specific questions. These included:
- WP3 - Psychosocial Impact: Determining the causal effect of factors like pain-related worrying, depression, sleep problems, and social support on neuropathic pain and disability.
- WP4 - Predictive Modelling of Pain and Neuropathy: Evaluating how baseline measures, including genetics, neurophysiology, demographics, clinical, molecular and biochemical data can predict a patient’s risk of developing chronic neuropathic pain.
- WP5 - Genetic Architecture: Investigating how combined genetic and non-genetic factors influence the odds of developing Neuropathic Pain in conditions such as Diabetic Peripheral Neuropathy.
- WP6 - Biomarker Validation: Examining the relationship between plasma markers (such as neurofilament light chain) and objective diagnostic measures like skin biopsy results or sensory abnormality.
Cross-WP integration: Building on these, the group also agreed over-arching research questions to connect findings across work packages – for example, incorporating biomedical and Quantitative Sensory Testing (QST) factors into WP3 Directed Acyclic Graphs (DAGs) and exploring how genetic and non-genetic factors combine to influence neuropathic pain outcomes.
Summary of Activities
Throughout the project, WP7 coordinated data analysis approaches, synthesising information from across the consortium.
Data Harmonisation: The team successfully produced a harmonised cross-centre dataset by combining and quality-checking individual REDCap databases from all PAINSTORM centres.
Methodological alignment: We held monthly data modelling meetings and co-organised a DAG workshop in London, focusing on introducing DAGs and discussing the development of the first DAG and a key methodological workshop in Ghent with WP3, focusing on both causal and predictive modelling techniques.
Key Outcomes
- Predictive Modelling: We developed predictive models using machine learning and conventional regression analysis to identify the most powerful predictors of neuropathic pain, response to treatmentand specific patient subgroups (clusters) sharing common traits.
- Perspectives on Pain Research: We published perspectives on causal modelling methods and data integration for the advancement of pain research.
- Digital Resources: We launched the Pain RNA-seq Hub (PRH), a public-facing tool for visualising pain-related genes and their network associations.
- Legal change in drug status: We investigated the impact of gabapentinoid reclassification in the UK in April 2019 on prescribing and death rates of gabapentin and pregabalin and of other anti-neuropathic pain medications in all four nations.
- This paper introduces the approach used in DOLORisk to build a risk model that aims to predict the onset and the resolution of chronic neuropathic pain: Development and external validation of multivariable risk models to predict incident and resolved neuropathic pain: a DOLORisk Dundee study
- This Nature article features an interview with Dr Jan Vollert and explains how the search for biomarkers, such as we are hoping to develop in PAINSTORM, could help improve pain treatment: Could biomarkers mean better pain treatment? Precision-medicine approaches for chronic pain could help people to receive effective treatment from the start.
- Zhao N, Bennett DLH, Baskozos G, Barry AM. Predicting “pain genes”: multi-modal data integration using probabilistic classifiers and interaction networks. Bioinformatics Advances (2024). doi: 10.1093/bioadv/vbae156
- Poppe, L., Steen, J., Loh, W.W., Crombez, G., De Block, F., Jacobs, N., Tennant, P..W.G., Van Cauwenberg, J. & De Paepe, A.L. How to develop causal directed acyclic graphs for observational health research: a scoping review. Health Psychology Review (2024). doi: 10.1080/17437199.2024.2402809
- Crombez G, Veirman E, Van Ryckeghem D, Scott W, De Paepe A. The effect of psychological factors on pain outcomes: lessons learned for the next generation of research. Pain Reports (2023). doi: 10.1097/PR9.0000000000001112
- Van Cauwenberg J, De Paepe A, Poppe L. Lost without a cause: time to embrace causal thinking using Directed Acyclic Graphs (DAGs). Int J Behav Nutr Phys Act (2023). doi: 10.1186/s12966-023-01545-8
- De Paepe, A.L., Gibby, A., Oporto Lisboa, L., Ehrhardt, B., Nunes, M., Fisher, E., Keogh, E., Eccleston, C., Woolley, C., McBeth, J. & Crombez, G. Building causal models in pain research: the case of executive functioning and transitions in pain states. PAIN (2025). doi: 10.1097/j.pain.0000000000003833
- Hebert HL, Pascal MMV, Smith BH, Wynick D, Bennett DLH. Big data, big consortia, and pain: UK Biobank, PAINSTORM, and DOLORisk. Pain Reports (2023). doi: 10.1097/PR9.0000000000001086
- Fundaun J, Thomas ET, Schmid AB, Baskozos G. The power of integrating data: advancing pain research using meta-analysis. Pain Reports (2022). doi: 10.1097/PR9.0000000000001038
- Baskozos G, Themistocleous AC, Hebert HL, Pascal MMV, John J, Callaghan BC, Laycock H, Granovsky Y, Crombez G, Yarnitsky D, Rice ASC, Smith BH, Bennett DLH. Classification of painful or painless diabetic peripheral neuropathy and identification of the most powerful predictors using machine learning models in large cross-sectional cohorts. BMC Med Inform Decis Mak (2022). doi: 10.1186/s12911-022-01890-x
Pain RNA-seq Hub (PRH): https://livedataoxford.shinyapps.io/drg-directory/
