Process of collecting data against a set of standards, in an attempt to improve compliance with those standards. But often ineffective – especially when no changes made (“closing audit loop”) to improve performance. And slow – by the time you’ve collected the data and summarized it, you’ve wasted time that could be been used to improve things you’ve already seen going wrong.

Often just turns into criticism, with intervention being no more than “perform better”. And then just induces resistance.

Cochrane review of audits in general found median 4.3% improvement with compliance, which isn’t much, but potentially more with incremental gains and repeated audit. And potentially scalable. And about a quarter achieve nothing.

Audit chain only as strong as weakest link – awareness of standards, reliability of processes, feedback.  All aspects of programmed should be designed with a focus on desired change in behaviour, and barriers should be anticipated.

Checklist for doing audit well –

• Can you recommend actions consistent with established goals and priorities
• Actions that are under audience’s control?
• Actions that are specific
• Can you provide multiple data points as feedback ASAP – and as often as frequency allows
• Individual as well as general feedback if possible
• Can you provide comparators that reinforce desired behaviour change
• Format of feedback – link visuals with summary message, multiple methods, minimize distractions.
• Actionable plan along with feedback
• Address barriers
• Short, actionable message with optional detail
• Realistic goals.
• Address credibility. 
• Anticipate defensive reactions
• Construct feedback through social interaction

Feedback to clinicians, who all think they’re great, requires careful thought.  Trying to improve already high performance may be a waste of effort, there is a ceiling for most things where organisation close to max capacity.

Patients and the public often surprised by the extent of variation.  They express frustration at difficulties in routinely measuring less technical aspects of care, such as consultation skills and patient centredness.  Patients are an untapped force for change which audit could learn to harness.

[ Revitalising audit, BMJ 2020;368:m213]

See Bawa-Garba case.

In a complex system, some errors are unavoidable. Their incidence can be reduced by better system design but they cannot be eliminated.

An error may have been the result of a decision, that could have been made differently, but when people are trying to work in an under-resourced and overstretched system, errors may be difficult to avoid.

Deterrence should therefore not involve criminalisation, but those who can influence the system eg managers and service directors.

Many errors are minor and inconsequential, but James Reason’s Swiss Cheese model highlights how a number of errors can align to enable more serious harm to occur.

In cases of “gross negligence”, the prosecution must establish beyond reasonable doubt that the failures caused the death. Unfortunately, a not guilty verdict may suggest that care was adequate. Prosecutions focus on the individual, not on the wider team or the healthcare system.

Good Practice

• Families should receive open disclosure and an apology
• If possible, the harm should be treated as a priority
• When relevant, compensation should be paid
• Appropriate mechanisms should be in place to hold to account those responsible for delivery of care
• Punishment may be appropriate but should be proportionate to the moral culpability of the behaviour, not the outcome of complex clinical problems
• Responses to problems should be timely – complex systems need repeated and rapid adjustment. Not served well by lengthy investigations.
• Motivated staff should be afforded the safety of a “just culture”, rather than “no blame” or “who’s the one to blame”.
• Culture of safety from health minister to most junior

Legal basis

According to the Bolam judgement, if a doctor has acted according to proper and accepted practice (standards of one’s peers), s/he is not guilty of medical negligence (even if some disagree with that practice); whereas the Bolitho judgement adds that practice must however be logically defensible – it’s not enough just that others do it!

Montgomery vs NHS Lanarkshire –

Bellshill shoulder dystopia and CP – Supreme Court overruled previous judgment. “Whether a risk is material should not be reduced to percentages”, nor should it be simply for the doctor to decide. “Would a reasonable person in the patient’s position be likely to attach significance to a risk?” Or does the doctor have reason to think the particular patient might feel it was significant?  That is the test of “materiality”. 

Patient does not have to ask specific question. Indeed, responsibility for concerns to be explored. 

Does not mean patients must be bombarded with information, must be comprehensible. 

Official verdict is as follows:

• Doctor is under a duty to take reasonable care to ensure that her patient is aware of any material risks involved in any recommended treatment, and of any reasonable alternative or variant treatments.
• The test of materiality is whether, in the circumstances of the particular case, a reasonable person in the patient’s position would be likely to attach significance to the risk, or the doctor is or should reasonably be aware that the particular patient would be likely to attach significance to it.
• The significance of a given risk is likely to reflect a variety of factors besides its magnitude: for example, the nature of the risk, the effect which its occurrence would have upon the life of the patient, the importance to the patient of the benefits sought to be achieved by the treatment, the alternatives available, and the risks involved in those alternatives. The assessment is therefore fact sensitive, and sensitive also to the characteristics of the patient.
• What amounts to a material risk and the skill and judgement required in explaining risks to the patient is for the Court, and not the medical profession, to judge.
• To prove a breach of duty to advise and warn, a pursuer does not require to prove that no doctor of ordinary skill would have failed to have given her advice, if acting with ordinary care, as supported by medical opinion.
• A patient may decide that she does not wish to know what her risks and options are.
• The “therapeutic exception”, which allows a doctor to withhold information from a patient only applies if its disclosure would be seriously detrimental to the patient’s health, or in circumstances of necessity, such as where the patient is unconscious or unable to decide.
• Causation continues to be based on a subjective test: what the pursuer was likely to have done, had she been warned and advised properly. To prevent hindsight bias, this will require to be tested by other evidence.

Interesting that language of consumers and choice was used in judgement, as if these decisions are simply a matter of providing information and allowing rational judgement – when lots of evidence to say that isn’t how people behave in practice!