High reliability organisations and systems thinking
- Concept
Aircraft carriers and air traffic control run dangerous work with very few disasters. Hospitals can learn how they do it.
What it is
A high reliability organisation (HRO) works in a complex, high-risk setting yet has far fewer serious failures than expected. It does so through a shared mindset: it hunts for small failures, distrusts simple explanations, watches the front line, recovers quickly when things go wrong and lets the person with the most expertise decide, whatever their rank. Systems thinking explains why: a process is only as reliable as the chain of steps it depends on.
Key facts
- Preoccupation with failure: small errors and near misses are reported and studied as signs of a weak system.
- Reluctance to simplify: «human error» is not accepted as the final explanation; the team looks for the conditions behind it.
- Sensitivity to operations: leaders know what is happening at the bedside now (safety huddles, rounds, real-time data).
- Commitment to resilience: errors will happen, so the team trains to detect them early and contain them.
- Deference to expertise: decisions move to the person with the knowledge, for example a nurse who stops a procedure.
- Complex systems: many interacting parts, tight coupling and time pressure mean small failures can spread fast.
Worked example
- A made-up ward gives a high-risk medicine through 5 steps: prescribe, verify, prepare, give, monitor. Each step works 95 % of the time.
- Reliability of the chain = 0.95 × 0.95 × 0.95 × 0.95 × 0.95 = 0.95⁵ ≈ 0.774, so about 77 % of doses pass every step without a failure.
- Out of 1 000 doses, about 1 000 × (1 − 0.774) ≈ 226 have at least one failed step somewhere.
- Option A: make every step 99 % reliable. 0.99⁵ ≈ 0.951, so failures fall to about 49 per 1 000.
- Option B: remove 2 steps by having pharmacy prepare the dose ready to give. 0.95³ ≈ 0.857, so about 143 failures per 1 000.
- Doing both (3 steps at 99 %): 0.99³ ≈ 0.970, about 30 failures per 1 000.
- Reading: shortening the chain and strengthening each link both help, and together they cut failures by about 87 % (226 to 30).
High reliability vs quality improvement vs incident analysis
| Feature | High reliability | Quality improvement | Incident analysis |
|---|---|---|---|
| What it is | A culture and mindset across the organisation | A method to test and spread changes | A review of one event or a cluster |
| Time frame | Continuous, every shift | Weeks to months per project | After the event |
| Key question | Where is the next failure coming from? | Did this change make things better? | What lined up to cause this harm? |
| Typical tools | Safety huddles, stop-the-line, near-miss reporting | PDSA, run charts | Root cause analysis, barrier ranking |
Markers
- HRO
- Preoccupation with failure
- Deference to expertise
- Series reliability
In practice
- Speak up when something looks wrong: an HRO expects anyone to stop the line, and leaders thank them for it.
- Use a daily safety huddle: what went wrong yesterday, what could go wrong today, who needs help.
- Report near misses: they show the same weak links as harm, at no cost to a patient.
Common traps
- An HRO is an organisation that makes no errors: it expects errors and is built to catch and contain them early
- Adding one more check always makes a process safer: each step is another link that can fail; a check helps only if it is independent and catches what others miss
- Average reliability tells you overall reliability: in a chain the reliabilities multiply, so 20 steps at 99 % succeed together only about 82 % of the time
- Deference to expertise means deference to seniority: authority moves to whoever knows most about the problem at that moment
Test yourself
Think of the answer, then tap the question to see it.
A discharge process has 8 steps, each 98 % reliable. Roughly what share of discharges pass every step, and what does this suggest?
0.98⁸ ≈ 0.85, so about 85 %; about 1 in 7 discharges has a failure. Fewer or more reliable steps would help more than urging staff to be careful.
A senior surgeon wants to continue, but the anaesthesia nurse sees a dangerous trend on the monitor and asks to pause. Which HRO principle applies?
Deference to expertise: the person with the relevant knowledge at that moment should be able to stop the work, regardless of rank.
A team closes an event review with ‘human error, staff reminded’. Which HRO principle has been ignored, and why does it matter?
Reluctance to simplify: stopping at the person leaves the conditions that made the error likely, so it will recur with someone else.
Why does an HRO welcome a rise in near-miss reports after a new reporting campaign?
Preoccupation with failure: more reports usually mean more willingness to speak, not more danger, and each near miss reveals a weak link before a patient is harmed.
Connections
Educational content on epidemiological and statistical methods; the examples use made-up numbers to teach the calculation. Not a clinical or legal guide: current Colombian rules are in the official sources.
Sources: Patient safety curriculum guide: multi-professional edition (World Health Organization, 2011); Patient safety (World Health Organization, 2023); Global Patient Safety Action Plan 2021–2030 (World Health Organization, 2021).
Spotted a mistake?We are always improving for the community. Tell us what is wrong and our team will review it.Report a mistake
Representative model, for educational purposes only.
