Routinely-collected hospital data offers statisticians an enormous and tempting resource, but its structure works against many of the assumptions we bring to analysis. Because these data are collected, stored, and optimised for mandatory reporting and billing rather than research, they shape and often distort how we define study populations, comparison groups, outcomes, and timelines.
This presentation examines what episode-level, fragmented, incomplete and skewed data means for analysis: how administrative coding drives outcome ascertainment, how unstable definitions and shifting reporting requirements confound long-term trends, and where these features quietly threaten the validity of otherwise standard approaches. I will also make the case that sound analysis of these data depends as much on relationships with the people who enter and manage them as on statistical expertise.
Presenter

Dr Taya Collyer
Latest Resources
-
-
ACTA STInG Webinar: The target trial framework for causal inference: a primer
-
Adaptive Allies: The Evolving Roles of DSMBs and Investigators in Bayesian Platform Trials – Part 3
-
The challenges of stopping trials for benefit: Lessons learnt from the immediate Kangaroo Mother Care (iKMC) trial – Part 2
-
STInG Chat – What makes this trial great?
-
Platform trials: key features, when to use them and methodological challenges
-
Interim Data Sharing by Data Safety Monitoring Board: What to share, with whom, and why – Part 1
-
Data Integrity Tool for assessing the integrity of randomised trials
-
The value of getting involved with ACTA STInG
-
Winner of the ACTA STInG Excellence in Trial Statistics award shares how ACTA STInG provided valuable networking opportunities during the pandemic
-
Data and Safety Monitoring Board (DSMB) Mentoring Program
-
ACTA Webinar – Developing a Best Practice Guide for Registry Based Trials
Latest Events
-
The Perils and Benefits of Subgroup Analyses in Trials
-
ViCBiostat Summer School
-
ACTA webinar: Tackling dichotomania (p<0.05) and other statistical diseases in pursuit of better science
-
ACTA Webinar: Charting the content of data monitoring committee charters for clinical trials
-
ACTA 2026 Trial of the Year Awards Ceremony
-
ACTA STInG Webinar: The target trial framework for causal inference: a primer
-
STInG Chat – What makes this trial great?
-
ACTA 2025 Clinical Trials and Registries Symposium
-
ACTA Webinar – Data Integrity Tool for assessing the integrity of randomised trials
-
ACTA Webinar – Developing a Best Practice Guide for Registry Based Trials
-
STInG Chat – How to build a career as a clinical trials biostatistician
-
Good Statistical Practice – GCP training for statisticians
-
ACTA – Historical and non-concurrent controls in clinical trials
-
STInG Chat: ‘Securing Funding: Tips and Tricks for Investigator Grants’