Involving all patients in decision-making. How to improve ‘option talk’ and risk communication in general practice for patients with limited health literacy?

We included 28 relevant scientific publications from which we distracted the following insights:

• Only describing the probabilities on positive and negative outcomes in a verbal manner, e.g. ‘your risk on hospitalisation is small’, is confusing as each person gives another meaning to the term ‘small’. It is preferable to describe the probabilities on positive and negative outcomes with numbers such as ‘2 out of 100 patients like you will be hospitalised’.
• People who have difficulty in understanding numbers due to limited numeracy may benefit from illustrations with visuals and drawings.
• It is important to important to present the information on the pros and cons as neutral as possible. For example; ‘2 out of 100 patients like you will be hospitalised. This also means that 98 out of 100 patients will not to be hospitalised’. This is also referred to as neutral framing.

To answer the question we have also analysed how options and risks are communicated in the patient decision aids that were available for the Dutch patients in 2021. We included 199 patient decision aids. Most patient decision aids report the probabilities on positive and negative outcomes with both words and numbers. Often the information was not framed in a neutral manner. Many patient decision aids visualise the numeric probabilities, e.g. with a population diagram (icon arrays) of 10 by 10 puppets to visualise the number of 100. Most patient decision aids use too many and too difficult words. Patient decision aids usually use table formats to list the treatment options and the positive and negative outcomes. Few patient decision aids, less than 2% of the total, were developed together with patients with limited health literacy. These patient decision aids used illustrations to describe the treatment options. The more detailed information on pros and cons was presented stepwise in small doses per option. The table format was not used in these decision aids.

The lessons that we have learned have been discussed in focus group interviews with general practitioners, experts and patients with limited health literacy. It was concluded that the doctor needs to be well informed of the patient’s context in order to be able to share decision-making. The patients expressed to really want to be informed about all available options. The preference to be informed with details about the probabilities on positive and negative outcomes differs between patients.

This project has resulted in a list of recommendations on how doctors can inform patients with limited health literacy about probabilities of medical options. These recommendations are especially important for health care professionals and developers of patient decision aids.

Recommendations for clinicians and developers of patient decision aids for SDM with patients with limited heath literacy

It must be considered that these recommendations have different levels of evidence. For some recommendations tailoring to the individual person is highly relevant as the group of “people with limited health literacy” is diverse. It may e.g. relate to functional illiteracy (difficulties in writing and reading skills), to numeracy or graph illiteracy, or to critical illiteracy (analyze information critically).

General aspects of shared decision making

• People with limited health literacy are not easy to identify. Try to sense the level of health literacy of the patient by using tools such as the teach back-method. Do not purposefully hand papers upside down to the patient. Use the information from the patient file.
• SDM is an iterative process, and the patient context should be well considered. There is always a gap between receiver (patient) and transmitter (doctor). It is the task of the doctor to see where the gaps are and to try to find a common level for the communication process. Take time to get to know the context of the patient. Clarify their knowledge about the diagnosis and what the patient needs to make a decision.
• Different risk communication strategies result in different risk perceptions influenced by various literacy skills. Be aware that people with limited health literacy are in general less likely to understand risk information correctly.

The option talk within shared decision making

• Present all options to the patient.
• Consider information overload. Use a layered approach when giving information. Opt for next consultation for further deliberation or deciding if the situation allows to do so.
• Consider that risk communication can fan fear in people with limited health literacy. Explore how much detail on risk and benefits the patient wants to know.
• Suggest patient to bring a relative or friend to the consultation.
• Offer patient to make notes or to audiotape the consultation.

Use of information material

• Consider computer skills of patient, also offer printouts.
• Walk through the information material together with the patient.
• Use illustrations, small amount of text, short sentences, clear layout, large letter types.
• Consider that the matrix format of tables to present options and their risks and benefits (fact sheet/keuzekaart) might be too difficult for subgroups of people with limited health literacy and needs to be further explained. The matrix format might be too difficult for people who have difficulties in writing and reading skills (functional illiteracy).

Risk communication

• People with limited numeracy tend to assign higher probabilities to low-frequency labels, and significantly lower probabilities for high-frequency labels compared to people with higher numeracy. Avoid use of verbal risk description only as it can lead to misinterpretation.
• Risk information is perceived as less risky when positively framed and as riskier when negatively framed. People with limited numeracy seem more susceptible to framing than people with high numeracy. Visual formats reduce or eliminate the framing effect. Use neutral framing.
• People with limited numeracy tend to show denominator neglect leading to inaccurate estimates. Equal denominators sizes can support better understanding. Use natural frequencies rather than percentages. Use same size of denominator, such as x-in-100 or x-in-1000. Avoid 1-in-x- format such as 1 in 5.
• Icon arrays can improve accuracy of absolute and relative risk reduction. But consider the numeracy and graph literacy of the person. For people with limited graph literacy visual risk information is not always useful. Visual risk communication strategies should be as simple as possible.
• Combine numerical with visual risk communication.
• Icon arrays seem to work best with smaller denominators (<100). Whereas bar charts might perform better for medium (100–499) and large (500–999) numerators. Icon arrays helped to eliminate denominator neglect. Use icon arrays to give a gist and consider the denominator size. (Gist representation = Getting an idea of the underlying message without exactly understanding numerical details of the risk message.)
• Risk ladders seem difficult to understand for people with limited health literacy. Do not use contextualization e.g., comparison with a daily risk as they lead to confusion.
• Adding verbal explanations of how to interpret visual information might help participants with low numeracy to use risk information.