OP7
FEASIBILITY STUDY: ESTABLISHING A NATIONAL DATASET OF OCCUPATIONAL HEALTH ASSESSMENTS
McElvenny1, M. van Tongeren1, E. Demou2, C. Warhurst3, P. Elias3
1University of Manchester, UK
2University of Glasgow, UK
3University of Warwick, UK
Background
There is lack of good quality occupational health data in the UK that facilitates detailed investigations of the link between work and health. One untapped source of health data collected is by Occupational Health Professionals.
Aims
To determine what can be established in terms of accessible occupational health data set(s) in the UK.
Methods
Eighteen semi-structured interviews were carried out over 2025-2026 with occupational health practitioners and policymakers/professional bodies. Analysis used Framework Analysis – a standard applied qualitative method well‑suited to policy/practice questions – with inductive thematic coding within each framework domain. All analyses were undertaken using ChatGPT, with outputs verified by the interviewers.
Results
Participants described wide variability in data structures and coding. This variability supports the need for a “core minimum dataset” plus an optional “enhanced dataset”. High‑value use‑cases were: (a) benchmarking, (b) evaluation of policy interventions, (c) workforce planning/resourcing and (d) policy insight. Key barriers were: information governance, employer and commercial sensitivity, heterogeneity of IT systems and the resource burden of data extraction/cleaning—especially for longitudinal linkage. Enablers repeatedly mentioned: transparent data‑sharing agreements, strong independent oversight (ethics and information governance), trusted secure research environments/safe havens, aggregation thresholds for small numbers and delivering tangible benchmark outputs back to contributors.
Implications for Policy or Practice
The establishment of a national occupational health dataset was considered to be potentially achievable if pursued as a staged programme beginning with a small, well‑governed pilot focusing on a minimum common dataset, with early engagement of software vendors and large providers.
