Incident counts don't measure safety. Improvement does.
What makes a hospital safer is what happens after each report: staff who hear back, lessons that become lasting fixes, and proof that practice changed.
the problem
Cluster incidents. See your system. Strengthen resilience.
Related incidents across wards, shifts and standards surface as one trend that no single report can show. Your teams redesign the system, instead of patching each incident.
Safe by design, not by vigilance.
A hospital that records, or a hospital that learns?
| A hospital that records | A hospital that learns, with Root | |
|---|---|---|
| After an incident | Filed, and closed when the form is complete | Turned into a lesson, with an owner and a deadline |
| Corrective actions | Too often "retrain staff" | Stronger system fixes alongside training, each rated for strength |
| Lessons | Stay in the case file | Shared with the teams who need them, and confirmed in practice by audits |
| Staff | Stop reporting when nothing changes | Hear back, and see what changed |
| What leadership sees | Incident counts | Whether the hospital learned |
how it works
Listen. Engage. Learn. Redesign. Prove. Measure.
- 01Report in seconds, named or anonymous.
- 02Every reporter hears back; a ward huddle while it's fresh.
- 03Why it made sense at the time, across five elements.
- 04Fix the system, each fix rated for strength.
- 05Audits confirm practice changed.
- 06Leadership sees whether the hospital learned.
engage
Every reporter hears back. Root makes it easy.
Root drafts a short reply and suggests how much can be shared. The manager approves it and sends it in their own name. Anonymous reporters follow along with a private code.
Reply to reporter · INC-2041
Drafted by Root
Thank you for reporting the near-miss on Ward 6. The two look-alike vials are now stored apart, and pharmacy has changed the labels. Your report made this happen.
one incident, through root
A tenfold overdose, and what the hospital learned.
In a crowded emergency department, a nurse gives a tenfold overdose of heparin. A senior doctor catches it 15 minutes later, and the patient is stabilised. The nurse reports it herself.
The five elements are the SEIPS model (Systems Engineering Initiative for Patient Safety). "Work as imagined versus work as done" comes from Safety-II thinking (Hollnagel).
Reporting staff can trust.
Anonymous, and still heard
Follow what happened with a private code.
Your words, unchanged
Kept exactly as you reported them.
Witnesses see one question
Never the whole investigation.
Systems, not scapegoats
Fixes target how work is done.
by role
A view for everyone who learns from incidents.
Frontline staff
Report, hear back, see what changed.
Managers
Replies drafted, reminders on time.
Quality and risk
Stronger actions, proof they stuck.
Educators
Training drawn from real incidents.
Leadership
Whether the hospital is learning.
the evidence behind the approach
- Reviews of incident reporting conclude the feedback loop must be closed with timely action, and that feedback encourages future reporting (Benn et al., Quality and Safety in Health Care, 2009).
- Healthcare has built a strong culture of reporting but a poor culture of learning (Macrae, BMJ Quality & Safety, 2016).
- Widely used safety guidance rates training on its own as a weaker action, and asks for stronger system fixes alongside it (US Veterans Affairs National Center for Patient Safety; Institute for Healthcare Improvement).
details
The details, when you want them.
Reporting and triage +
Voice or text; named or anonymous; facility and unit tagged automatically. Seriousness rated as the report arrives; urgent hazards alert the right people at once.
Standards +
Incidents matched to the standards you already audit against, and one incident to several standards at once.
Investigation +
A prompt to hold a ward huddle while memories are fresh; analysis across five elements (person, task, tools and technology, environment, organisation); work as imagined compared with work as done; case ownership, notes and files; witnesses asked one question, without seeing the case.
Actions +
Owners, deadlines, strength ratings, and a recheck date to confirm each fix worked.
Feedback +
AI-drafted replies approved by a person; nudges at key moments; anonymous reference codes.
Learning +
Training topics and content suggested in one click; lessons shared with teams; practice confirmed by audits.
Leadership dashboard +
Learning measures, not just incident counts.
Security +
See IT and security.
resources
Take it to your team
faq
Questions quality and risk teams ask
Will AI reply to staff automatically? +
No. Root drafts the reply; a manager approves it and sends it in their own name.
What if the details can't be shared? +
Root suggests what can be said. Even a short acknowledgement answers the question every reporter asks: did anyone care?
How do anonymous reporters hear back? +
With a private reference code. They can follow progress without revealing who they are.
Isn't training the obvious fix? +
Sometimes. But training on its own is one of the weaker fixes. Root suggests stronger system changes alongside it, and audits check that whatever was chosen actually worked.
What does leadership see? +
Whether the hospital is learning: how quickly staff hear back, lessons shared, actions completed, and practice changes confirmed by audits.
Does it work with our standards? +
Incidents are matched to the standards you already audit against in SafetyOS, and one incident can count against several standards at once.
How is it priced? +
As an annual subscription. We'll give you an exact quote for your hospital.
Some answers depend on your hospital. We'll cover them with you.
Become a hospital that learns from every incident. Let's start there.
Book a walkthrough- →Your price, based on your hospital
- →How your teams reply to reporters today, and how Root changes it
- →Setting up anonymous reporting and reference codes
- →Turning lessons into training with your educators
- →The learning measures for your leadership dashboard
- →Connecting incidents to your audits in SafetyOS