Emily is Research and Trials Lead at the Global Safety Evidence Centre. She leads on primary evidence collection and research methodology, with a particular focus on quasi-/experimental approaches.
Areas of Research Interest
This page is approximately a 6 minute read
Running trials to build the global safety evidence base
The Global Safety Evidence Centre aims to build a strong and accessible evidence base for what works to improve safety worldwide. Trials are a key part of that ambition. This Areas of Research Interest (ARI) page sets out what we mean by a trial and why we want to run more of them, some example lines of enquiry we are interested in, and how industry partners and international organisations can work with us to design and deliver this work.
What is a trial, and why do we want to run them?
A trial is a prospective, controlled study that tests whether deliberately changing something — an intervention — causes a change in an outcome of interest. Unlike observational research, which looks for patterns and associations in data that already exist, a trial involves an active, planned change and measures its effect going forward. In its most rigorous form, a randomised controlled trial (RCT), the units of study — which might be individuals, teams, sites or whole organisations — are randomly allocated to receive an intervention, or not at a given time. Randomisation can help us say with confidence, that any difference in outcome was caused by the intervention itself, rather than by pre-existing differences between the groups being compared*.
Much safety practice today rests on experience and expert judgement, and often for good reason: safety systems are broad, complex, cross-disciplinary and rarely offer one-size-fits-all solutions. Importantly, the Global Safety Evidence Centre’s Stakeholder Research function aims to unpick and document this tacit knowledge further (Advisory Panel, what practitioners need). However, reliance on expert judgement also means that many long-standing assumptions about what improves safety have not been tested against a genuine counterfactual — what would have happened anyway, without the intervention. The Global Safety Evidence Centre’s evidence reviews, including our review of reviews of occupational safety and health (OSH) interventions, have found variable evidence quality, inconsistent outcome measures and a shortage of rigorous intervention studies across the safety field.
Running trials can help us move from describing safety problems to testing solutions — building an accessible, evidence-based toolkit of activities and interventions that are shown to improve safety outcomes, rather than assumed to. Trials complement systems thinking by helping to unpick which specific activities within a complex system are actually doing the work, and they support more robust statistical conclusions by better satisfying the assumptions that underpin sound inference. Alongside evidence synthesis, trials can provide strong evidence for the effectiveness of different safety activities and processes.
* There are contexts where randomisation trials are not required to show effectiveness, in which case this call out needs not apply. For example, highly precise and controlled mechanical testing in engineering.
Example trial ideas
We want to grow a portfolio of trials across the Global Safety Evidence Centre’s Safe Work and Safety Science programmes as well as the Foundation’s priority areas - safer maritime systems, skilled people for safer engineering, safer sustainable infrastructure - and particularly relating to existing and emerging global challenges i.e., new and emerging technology, the changing global workforce, and adapting to climate change (Lloyd’s Register Foundation Strategy 2024-2029). The examples below illustrate the kind of work we are interested in supporting but are not exclusive.
Testing interventions to reduce workplace harm
A key line of enquiry applies trial methods directly to occupational safety and health (OSH) and wider workplace harm interventions — testing whether specific programmes, tools or changes to ways of working actually reduce injury, fatality, psycho-social wellbeing concerns or unsafe behaviour. Examples relevant to our priority areas include:
- Safer maritime systems: cluster-randomised or split-plot trials that allocate ships to different safety-intervention conditions and crew members within each ship to sub-conditions, to establish which onboard harm-reduction programmes genuinely reduce incidents at sea.
- Skilled people for safer engineering: a) trials of apprenticeship or vocational training curricula, testing whether specific safety-skills modules reduce harm rates among new entrants to high-risk trades, compared with standard training; b) trials to establish which ways of learning are most effective and cost appropriate for different safety contexts; c) trials to establish the effectiveness of new technologies (e.g. VR) for learning in different safety contexts.
- Safer sustainable infrastructure: stepped-wedge (roll-out) trials looking at safety effects of engineering controls or physical safety redesign on construction sites, where every site eventually receives the intervention. Interactions between engineering controls and human factors are of particular interest and could be explored with a factorial design.
- Emerging technology: A/B trials of digital risk-assessment tools or safety apps, testing whether a redesigned interface changes reporting behaviour, downstream safety outcomes and general safety messaging effectiveness.
- Changing global workforce: trials of harm-reduction interventions tailored to multilingual/cultural teams, different age groups, rapidly developing industrial environments, more mobile or more casualised workforce, such as adapted inductions for contract or gig workers.
Adapting to climate change: trials exploring heat-stress interventions and extreme-weather protection measures for outdoor and exposed workers, as a changing climate increases exposure to environmental hazards. For example, exploring the effectiveness of changing working patterns, implementing monitoring technology and installing heat-resistant infrastructure on worker psycho-social wellbeing and physical safety. Stepped-wedge trials with a sustainability programme would be particularly valuable.
Timing: importance for safety outcomes
A second, less explored interest explores timing itself as something to be tested, rather than a fixed backdrop. Leading indicators — the early, predictive measures organisations use to anticipate safety outcomes before something goes wrong — are typically collected at whatever point is administratively convenient. We are interested in trials that randomise the timing of measurement, interventions or leadership behaviours, for example, to establish when these things matter most, not just whether they matter at all. Illustrative examples include:
- Randomising the point in a project’s lifecycle (mobilisation, peak activity, demobilisation) at which a leading indicator survey is administered, to test whether the same measure predicts outcomes differently depending on when it is taken. This could be particularly relevant to major maritime and infrastructure projects with distinct phases.
- Randomising when safety-leadership or supervisory-skills training is delivered — early versus mid-project — to test whether front-loading training produces better outcomes than just-in-time delivery, relevant to skills development pipelines.
- Trials comparing different leadership styles (for example, transformational versus transactional safety leadership) overall and at different project stages, to test whether the ‘right’ style is time-dependent rather than fixed — applicable across maritime crews, infrastructure teams, and increasingly dispersed or digitally managed workforces.
- Trials examining whether the timing of climate-risk briefings, relative to seasonal or project milestones, changes their effectiveness for organisations managing infrastructure exposed to changing climate hazards.
- Trials testing whether digital or remote monitoring of leading indicators produces different results depending on project stage, relevant to organisations managing a changing workforce that is more distributed and less office-based than before.
A call to industry partners and global organisations
The Global Safety Evidence Centre wants to partner and collaborate in this area. We are inviting industry partners, employers, practitioner bodies, and international and national organisations, interested in safety effectiveness, and working in different safety fields and contexts, to reach out to discuss trial ideas. We are particularly interested in partners who can offer:
- Access to real-world settings — industrial sites, projects or workforces — where an intervention could be tested and, where appropriate, randomised.
- Existing programmes or interventions whose effectiveness has not yet been rigorously tested.
- Data infrastructure or leading (and lagging) indicator measurement systems that could support a trial.
- Skills and interest in pursuing this line of work.
The Global Safety Evidence Centre’s Research and Trials function can provide methodological support — including trial design and review— to help shape an idea into a deliverable study, analysis or programme.