RISK: What are my chances, doc?

Communicating risk is essential in obtaining and being able to demonstrate valid consent.

  • Date: 10 October 2014

MOST doctors understand the concept of relative risk, but what about patients? The majority will simply grow frustrated or tune-out if a discussion aimed at joint decision-making becomes a lecture in statistics. Yet communicating risk is essential in obtaining and being able to demonstrate valid consent.

This requires meaningful dialogue with the patient, which includes a discussion about the chance or probability of things going wrong. GMC guidance for doctors, Consent: patients and doctors making decisions together, states that when sharing information and discussing treatment options “you must give patients the information they want or need about, amongst other things, the potential benefits, risks and burdens, and the likelihood of success for each option”.

Entering a discussion about risk probabilities I am always reminded of Mark Twain’s quote about "lies, damned lies, and statistics”.

A very public example of patients being misled about risk probability occurred in 1995 when the UK’s Committee on Safety of Medicines decided to warn doctors that a new, third-generation oral contraceptive pill doubled the risk of thrombosis. This was seized upon by a frenzied media and resulted in thousands of women stopping their contraceptive pill, even though the actual risk had merely increased from a one-in-7,000 chance of getting the disease to a two-in-7,000 chance.

Are doctors confused by statistics? A new book by one prominent statistician says they are – and this makes it hard for patients to come to informed decisions about treatment.

Gerd Gigerenzer is director of the Harding Center for Risk Literacy in Berlin and in his book Risk Savvy he takes aim at health professionals for not giving patients the information they need in a way in which they can understand in order to make valid choices about their care and treatment.

Gigerenzer describes how in a series of workshops in 2006 and 2007 he posed the same statistical problem to over 1,000 gynaecologists relating to the results of a positive mammography screening. The doctors in the workshops were provided with additional relevant clinical information to base their answers on and in a typical session only around 21 per cent provided the correct answer. Apparently this is a worse result than if the doctors had been answering at random!

The problem then may be two-fold. It’s not only being unable to produce relevant statistics for patients for every treatment option; it’s also about being unable to make sense of those statistics when placed in front of you.

Part of the difficulty here may be in setting out risk probabilities as percentages, which apparently a lot of us struggle to understand. Possible alternatives are the use of simple fact boxes and tools such as option grids, which set out frequently asked questions concerning a test or procedure and then off er likely outcomes for both having and not having the test done.

Another alternative way of expressing the relative risk uses numbers of people instead, and where possible with the aid of diagrams. In Gigerenzer’s example of a positive mammogram, the reality looks visually clearer if set out on a flow chart format: http://www.bbc.co.uk/news/magazine-28166019

Other visual aids used to communicate risk probabilities include diagrams representing percentages out of 100 stick figures. These can off er a handy short-hand of risk which can be utilised as part of a range of complimentary data formats together providing enough fl exibility to address the needs of a variety of patients.

There are other factors to consider in communicating relative risk to patients:

  • Guard against over simplification of language: terms such as ‘common’ or ‘rare’ can assume a shared perspective, when in fact patients may judge risk by a different order of magnitude.
  • Patients may best understand absolute risk expressed in natural frequencies, i.e. 1:200 patients suff er a postoperative complication.
  • Presenting absolute risk figures alone has also been shown to lead to either an overweighting of low probabilities or an underweighting of high probabilities.

One particular study looking at probabilities of harm found that the term ‘frequent’ was interpreted on average as equivalent to around 70 per cent. However, the range of answers provided by participants was from 30 per cent through to 90 per cent.

What is the law, and what do the regulators say about all of this?

The landmark medico-legal case Chester v Afshar confirmed a duty to warn patients about risk. In the case, Ms Chester was left paralysed following surgery for a lumbar disc protrusion. The court ruled that Dr Afshar was negligent in failing to warn her of the 1-2 per cent risk of the procedure going wrong. It’s interesting to note that the court’s chosen method of communication here was in percentages, rather than 1:100 or 2:100 cases.

GMC guidance on consent is heavy on what is required and expected from doctors and what they must ensure has been conveyed, but silent on how the actual risk probability and impact is communicated. No two people are alike in the ability to comprehend risk so it is up to the individual healthcare professional to judge if a patient truly understands.

Alan Frame is a risk adviser at MDDUS.

This page was correct at the time of publication. Any guidance is intended as general guidance for members only. If you are a member and need specific advice relating to your own circumstances, please contact one of our advisers.

Read more from this issue of Insight

Insight (formerly Summons) is published quarterly and distributed to all MDDUS members throughout the UK. It provides a mix of articles on risk, medico-legal and regulatory matters as well as general features and profiles of interest to our members. Browse all current and back issues below.
In this issue

Related Content

Coroner's inquests

Medico-legal principles

Risk: Can reflective practice be “incriminating”?

For registration, or any login issues, please visit our login page.