L8 Early Professional Experience Quality and safety in clinical practice

Overview

An Early Professional Experience lecture by an anaesthetist on unintended consequences in healthcare: how often patients are harmed, why errors happen, what makes medicine especially error-prone, what other industries do better, and how to respond personally and professionally when something goes wrong. It moves from incidence figures, through the psychology of slips and lapses and the Swiss cheese model of accident causation, to safety culture and teamwork, and closes with a case of a near-fatal obstetric collapse that went right because of a large coordinated team.

Framing and starting position

• The lecturer is an anaesthetist; anaesthetists are described as obsessed by mistakes.
• Undergraduate input on this topic is minimal, and the lecturer notes doctors make up 7% of the workforce referred to on the slide.
• Key framing statements: you will make mistakes; “First do Harm” (an inversion of the usual aphorism); and, importantly, mostly things go well.
• The slide asking “Hands up if you want to help people?” shows a normal distribution curve with an arrow reading “You are here” on the right-hand descending side near the top, positioning students on a distribution of motivation to help people relative to the peak.
• The lecture’s structure was given as: what goes wrong, why it happens, getting through it, and all of it true.

Warning

One slide is image-only (a person in scrubs in a theatre surrounded by blood-stained drapes, instruments and drapes on the floor). [slide carries no text or caption, so its intended meaning is not stated]

Incidence: how bad is it?

Key figures to know:

• Adverse event rate: 12.9%
• 15% of these are associated with permanent disability or death (sentinel events)
• One third are preventable
• USA: 40,000 deaths
• Social and economic cost: millions of dollars

Scale analogy given on the slide: this is the equivalent of a Boeing 747 crashing one a day, every day in the USA, and 3.5 a year in New Zealand.

Why healthcare makes errors

It is not about motivation

Bored workers doing boring work and highly skilled motivated employees have about the same error rate, or less error with better design. Motivation is therefore not the lever; design is. The question posed is why healthcare mucks up despite motivated staff.

Errors vs mistakes (the lecturer’s distinction)

• Mistake: a slip or lapse, when your brain isn’t able to stop you.
• Error: when you get it wrong. Two forms:
  • because you don’t know (you might at least be hesitant)
  • because you are confident but wrong, which is the worst case because you will ignore warning signs

Branching decisions multiply the chance of error

A single presenting symptom generates a wide differential and a long chain of decisions, each an opportunity for error. The worked example:

1. “I have a cough” gives a differential of laryngitis, tracheitis, pneumonia, ACE inhibitor, smoker, lung disease, PE, TB and more.
2. Narrowing to pneumonia still leaves bacterial, fungal, viral or parasitic causes. Diagnosis involves exam, CXR or CT, sputum and bloods.
3. Treatment branches again: antibiotics, physio, admission to hospital versus bed rest at home, isolation or not, and which antibiotic.

Why our brains produce slips and lapses

Our brains are energy efficient, so we pattern-match instead of reading. The triggers listed are:

• similar words
• similar colours
• what you saw last week
• what you usually see

The illustration is a red octagonal sign, shaped exactly like a STOP sign, printed with the word “GO”: the expected pattern overrides the actual content.

States that increase error: HALT

Against the belief that “it won’t happen to me”, the mnemonic HALT:

• Hungry
• Angry
• Late
• Tired
• and Sick

“Stress” increases errors. Related point: there are no studies showing that yelling helps, so help stop it.

Worked examples of error

• Look-alike ampoules: a row of six near-identical glass ampoules, five labelled Ephedrine Sulfate Injection and the fourth from the left instead Heparin Sodium (porcine mucous) 5,000 I.U. in 0.2 mL. The point: people try in medicine but it is hard to reduce errors, and making a serious drug error should not be this easy (EZDrugID.org campaign for safer medication packaging).
• Wrong-sided knee replacement, left prosthesis put into the right knee, locally. Contributing system factors: the department used to hold 2 hip sets and 1 knee set, and now holds 9 hip sets and 5 knees; 2 orthopaedic theatres run simultaneously; and there are 3 universal versus 2 sided sets.
• Emergency Department errors from notes and results mix-ups in a busy department with open shared notes storage: an 80 year old man sent for a pregnancy ultrasound, a renal patient apparently “cured”, anaemia suddenly appearing. Solutions offered: pre-printed lab forms, notes kept on the bed, and ultimately an electronic record.
• Potassium down a central line: an anaesthetic registrar injects potassium too quickly down a central line and the patient arrests.

The Swiss cheese model of accident causation

Harm reaches the patient only when weaknesses in several successive layers line up. In the potassium case the slices (left to right) and their holes were:

Layer (slice)	Hole that lined up
Experience	OD patient
Nurses	ICU
Me	Monday
KCL down CVL	Too fast

Each weakness alone would have been insufficient; a trajectory through all four aligned holes let the error reach the patient.

Outcome bias and blame

Using the potassium patient: how you feel about, or judge, the clinician differs depending on whether the patient lived or died. The error is the same, but punishment tends to vary. A system view and system change prevent future error; blame will not.

Who is good at safety, and what they do

Important

The consistent message is that engineered solutions and team systems outperform policies, posters and exhortation.

• Engineers: they assume people are idiots, so they engineer solutions rather than relying on policies and posters. The slide cites a figure of 180,000 and uses a front-loading washing machine as the example of a designed-in safeguard.
• Pilots: motivated by extreme self interest, and they use checklists.
• Recent successes in healthcare:
  • the red syringe (“Red is for Relaxant”, for muscle relaxants such as rocuronium/Norcuron)
  • the safety cannula
  • pre-filled syringes
  • the CVL (central venous line) bundle

Safety culture and teamwork

• The safety culture slogan, repeated in the summary: See it. Say it. Fix it.
• Avoiding error is a team activity. People will throw you a life raft, including patients: Korero Mai.
• $\text{Performance}=\text{skill}\times \text{collaboration}$
• Crew resource management is named as the model imported from aviation.

Medical culture: the wrong heroes

Medicine idolises the wrong archetype (the “lone genius”, illustrated with Dr Gregory House). The flawed features listed:

• the man alone
• professional secrets
• no weakness, e.g. “I can do a tracheostomy”
• no encouragement, e.g. the orthopaedic registrar

Most of us are perfectionists, but error is inevitable and serious bad outcomes will occur. When it happens to you it will hurt, which raises the question of why we do not talk about it.

When things go wrong

1. When it is not you

• Withhold judgement until you have answered the questions.
• Could it happen to anyone else? That is the test of system error versus personal knowledge deficit.
• Investigate it: is this usual (wrong) practice?
• Fix it.

2. When it is you

• Admit it. This is really hard.
• Tell the patient and family. Also hard.
• Get “me” (a senior colleague) to come along.

3. Living with it

• Beware retrospective smarts, that is, beating yourself up with hindsight.
• Have a self-care plan.
• TALK. Do not drink or take drugs.

Summary points from the lecture

• Error happens and no one is immune.
• A broad view and system change will help prevent future error; blame will not.
• “Every error is a treasure” (Don Berwick), because most go unnoticed.
• Tell patients the truth.
• See it, say it, fix it.
• Be “safe”.

Case: obstetric collapse, when things go right

Framing: for things to go right, first they have to go wrong; you have to get used to bad stuff; the case is shared with the patient’s permission.

Presentation. Called to chaos on obstetrics, with no idea what was happening, extreme technical difficulties, and a decision to go to theatre. The woman was desperately unwell: uncooperative, swollen, mottled, with a petechial rash, and two senior colleagues (Jochen and Barbara) were both worried.

Plan. Get the baby out, and try to work out what is going on. Differential considered:

• eclampsia
• seizure
• abruption
• sepsis

Course. General anaesthetic, baby delivered. Syntocinon and antibiotics given, with concern about the baby. Then oozing started, progressing to bleeding and then to “hosing”. A large-bore line (“big pipe”) was placed in the neck. Adrenaline and hysterectomy followed; the lecturer can honestly say it was life saving (decided by a 4-way vote).

Team involved. Jochen, Paul, Steve and Chris, Nick and Nicole, nurses, surgeons, orderlies, lab staff, Jim Faed, midwives, and NICU/paediatrics.

Outcome. Discharged on 4th January after 14 days in hospital. Rea, John and Dylan: Dylan is a normal 6 year old now.

Warning

The intraoperative anaesthetic record shown lists drugs including ketamine, rocuronium, suxamethonium, oxytocin, oxytocin/ergometrine, cefazolin and what appears to be carboprost, plus calcium chloride 10 mL, NaHCO3 8.4% 60 mL, adrenaline, fentanyl, phenylephrine and noradrenaline infusions, and multiple blood products (RBC, FFP, cryoprecipitate, platelets). [handwriting largely illegible: exact doses, units and timing could not be reliably transcribed]

Warning

The nursing note shown is dated 28.12.18 and records observations stable, afebrile, temperature 37.4 to 36.8. [handwritten note is partially cut off at the bottom of the slide and portions are illegible]

Self-test

1. State the adverse event rate quoted, and the proportion of those events associated with permanent disability or death.
2. What proportion of adverse events are said to be preventable, and how many deaths per year does the lecture attribute to them in the USA?
3. Explain the aviation analogy used to convey the scale of adverse-event deaths in the USA and in New Zealand.
4. Distinguish a “mistake” from an “error” as the lecturer defines them, and explain which form of error is the worst case and why.
5. Explain why the comparison between bored factory workers and highly skilled motivated employees matters for patient safety.
6. Using the cough example, describe how a single presenting symptom generates opportunities for error through to treatment.
7. List the four brain-related triggers for slips and lapses given, and explain what the STOP/GO sign illustrates.
8. Expand the mnemonic HALT, including the additional letter, and state what else is said to increase errors.
9. Describe the look-alike ampoule example and the safety problem it demonstrates.
10. List the system factors offered as context for the wrong-sided knee replacement.
11. Give the three Emergency Department error examples and the three solutions proposed.
12. Draw out the Swiss cheese model for the potassium case: name the four layers and the hole that lined up in each.
13. Predict what the Swiss cheese model implies about preventing the next potassium incident, and contrast that with blaming the registrar.
14. Explain what the “did the patient live or die?” question about the potassium patient is designed to demonstrate.
15. Explain why engineers are held up as good at safety, and what they avoid using.
16. Why are pilots cited, and what tool do they bring?
17. List the four recent healthcare safety successes named.
18. State the three-step safety culture slogan and the performance equation given.
19. List the four features of the “wrong heroes” medical culture.
20. When something goes wrong and it was not you, what four steps should you take?
21. When the error is yours, what three things does the lecture say to do?
22. Describe the advice on living with an error, including what to do and what to avoid.
23. Quote the Don Berwick line used in the summary and explain the reason given for it.
24. In the obstetric collapse case, list the four diagnoses considered in the initial plan.
25. Describe the sequence of events in theatre in the obstetric case from induction to the life-saving intervention, and state the outcome.
26. Integrative: take the wrong-sided knee replacement and explain it using the system-versus-person framework, the Swiss cheese model, and the engineering approach to solutions.

Answers

Reveal answers

1. Adverse event rate 12.9%; 15% of these are associated with permanent disability or death (sentinel events).
2. One third are preventable; 40,000 deaths in the USA.
3. The death toll is equivalent to a Boeing 747 crashing one a day, every day, in the USA, and 3.5 such crashes a year in New Zealand.
4. A mistake is a slip or lapse, when your brain isn’t able to stop you. An error is when you get it wrong, either because you don’t know (in which case you might be hesitant) or because you are confident but wrong. The latter is the worst case because you will ignore warning signs.
5. Their error rates are about the same, or lower with better design, so error is not explained by lack of skill or motivation. Better design, not more motivation, reduces error.
6. “I have a cough” gives a differential of laryngitis, tracheitis, pneumonia, ACE inhibitor, smoker, lung disease, PE and TB among others. Narrowing to pneumonia still leaves bacterial, fungal, viral and parasitic causes, worked up with exam, CXR or CT, sputum and bloods. Treatment branches again into antibiotics, physio, admission versus bed rest at home, isolation or not, and which antibiotic. Each branch point is an opportunity for error.
7. Similar words, similar colours, what you saw last week, and what you usually see, because our brains are energy efficient. The red octagonal sign printed “GO” shows that the expected pattern (red octagon means STOP) overrides what is actually written.
8. Hungry, Angry, Late, Tired, and Sick. “Stress” also increases errors.
9. Six visually near-identical ampoules: five labelled Ephedrine Sulfate Injection and the fourth from the left labelled Heparin Sodium (porcine mucous) 5,000 I.U. in 0.2 mL. Packaging that looks alike makes a serious drug error easy despite people trying; hence the campaign for safer medication packaging.
10. Stock has grown from 2 hip sets and 1 knee set to 9 hip sets and 5 knees; 2 orthopaedic theatres run at once; and there are 3 universal versus 2 sided sets.
11. Errors: an 80 year old man sent for a pregnancy ultrasound, a renal patient apparently cured, and anaemia suddenly appearing. Solutions: pre-printed lab forms, notes kept on the bed, and an electronic record.
12. Layers and holes: Experience / OD patient; Nurses / ICU; Me / Monday; KCL down CVL / too fast. The error reached the patient only because the holes in all four layers aligned.
13. Prevention means closing or misaligning layers, that is, changing the system conditions (who gives it, where, how fast, what is stocked) so the holes no longer line up. Blaming the registrar removes none of the other layers’ holes and will not prevent recurrence, whereas a broad view and system change will.
14. Outcome bias: the error is identical whether the patient lived or died, but judgement and punishment vary with the outcome.
15. Engineers assume people are idiots and therefore engineer the solution into the device, rather than relying on policies and posters.
16. Pilots are cited because they have extreme self interest in safety, and they bring checklists.
17. The red syringe, the safety cannula, pre-filled syringes, and the CVL bundle.
18. See it, say it, fix it. Performance equals skill multiplied by collaboration.
19. The man alone; professional secrets; no weakness (“I can do a tracheostomy”); and no encouragement (the orthopaedic registrar example).
20. Withhold judgement until you have answered; ask whether it could happen to anyone else, which distinguishes system error from personal knowledge deficit; investigate whether this is usual (wrong) practice; and fix it.
21. Admit it, which is really hard; tell the patient and family, also hard; and get a senior colleague to come along with you.
22. Beware retrospective smarts, which is beating yourself up with hindsight. Have a self-care plan. Talk about it, and do not drink or take drugs.
23. “Every error is a treasure”, because most errors go unnoticed, so the ones that surface are the opportunities to learn and change the system.
24. Eclampsia, seizure, abruption, and sepsis.
25. General anaesthetic and the baby was delivered; syntocinon and antibiotics were given with concern about the baby; then oozing began, progressing to bleeding and then hosing; a large-bore line was placed in the neck; adrenaline was given and a hysterectomy performed, judged life saving by a 4-way vote. She was discharged on 4th January after 14 days in hospital, and the baby, Dylan, is a normal 6 year old.
26. System versus person: the test is whether it could happen to anyone else, and the growth from 2 hip sets and 1 knee to 9 hip sets and 5 knees, two concurrent orthopaedic theatres, and 3 universal versus 2 sided sets says it could, so this is system error rather than an individual knowledge deficit. Swiss cheese: those stock, staffing and theatre-throughput conditions are holes in successive layers that aligned to let a left prosthesis reach a right knee. Engineering approach: rather than a policy or a poster telling staff to check the side, design the system so the wrong set cannot be selected or opened, in the way the red relaxant syringe and the CVL bundle designed out earlier errors.