Discrete choice experiments in AMS

In this section, we will describe a series of examples of DCE’s that have been used to explore aspects of antimicrobial stewardship (AMS).

1. UK public preferences for using online pharmacies to obtain an antibiotic prescription

This study, available here, explored the preferences of the UK general public towards using online pharmacies to obtain antibiotic prescriptions compared to consulting with their primary care practitioner (their GP). The work stemmed from a concern that online providers might have a lower threshold for prescribing antibiotics than GPs working in the NHS, and that a shift to online prescribing could negatively affect antimicrobial resistance (AMR).

The goal of this study was to understand public preferences and what would make online consultations more or less appealing.

The study asked members of the public to:

The following infographic shows an example of the choice questions – 5 attributes covering aspects of the consultation experience, and a choice of two types of consultation that have been labelled to help identify preference.

A screen-reader version of this infographic is available as a PDF here.

The findings from this study revealed a strong preference for consultations via an individual’s local medical centre rather than an online provider, with an estimated willingness to pay £11 for that type of consultation, as well as a preference for consultations to be face-to-face.

However, a latent class analysis identified a subgroup with a lower strength of preference for traditional consultations, but a stronger preference for convenience. This group may be more open to consulting with online providers. Which means, if these providers do indeed have more permissive AMS requirements than the NHS, this may provide a mechanism for increased antibiotic use. The study identified sociodemographic characteristics of this respondent group, which could be used to develop targeted public health interventions to discourage inappropriate use of antibiotics.

2. Delayed prescription of antibiotics

Delayed prescription is an approach to reducing antibiotic consumption in primary care. With this approach, a GP gives a prescription for antibiotics but asks patients to wait and only take them if the condition doesn’t improve or gets worse. Delayed prescription has been shown to reduce antibiotic consumption in primary care without a negative impact on patient outcomes.

This study, available here, aimed to understand in what situations the public would be more likely to accept a delayed prescription, with a similar study ran for GPs.

A screen-reader compatible PDF of this infographic is available here.

The above figure shows an example question from the public study. The instructions to respondents indicated that they had a respiratory tract infection, and reminded them of the option of delayed antibiotic prescribing.

There are seven attributes, each of which could take 3 or 4 levels. Note that this has a slightly different structure, with one profile and respondents choosing from two options. The analysis, however, is very similar, using a straightforward logistic regression model.

This study found that the acceptance of delayed prescription was driven by symptom severity and – particularly among parents considering treatment for a child – duration of illness. Females, older adults, and people with good understanding of antibiotics were more open to delayed prescription. This information could help GPs in deciding how to present this option.

These findings were able to suggest topics for educational interventions and potential population sub-groups to target with these interventions. Based on the results, we estimate that delayed prescription could reduce antibiotic consumption for sore throats by around half a million prescriptions per year in the UK.

3. Stopping antibiotics

The study, available here, considers the ‘review and revise’ process for antibiotics given in emergency care, and the factors that make clinicians more or less likely to stop antibiotics.

Respondents were given a scenario, which was described as follows:

The structure of the choice question was similar to example 2 – one profile and two alternative actions. This can be seen in the following infographic.

Click here for a screen-reader compatible PDF.

The study found that prescribers were more likely to continue antibiotics if discontinuation is in conflict with local guidelines. The risk due to discontinuing was found to be more important than the risk due to continuing, particularly if the symptoms clearly indicate antibiotics, or if the patient is frail. Free-text comments indicated that the response to treatment plays an important role in the decision whether to discontinue, and this feature should be included in future studies on discontinuation.

4. Preferences for diagnostics in managing antimicrobial resistance

This study, available here, looked at preferences for diagnostics that would support the choice of whether or not to use an antibiotic. It was conducted among ‘patients’ – that is, recent users of antibiotics – and compares responses across 7 countries. Respondents were asked:

This is a simpler study for respondents than the other examples, as shown in the infographic below, with only 3 attributes in each choice question: speed, convenience and confidence.

This is also available as a screen-reader compatible PDF.

The study found that convenience was the most important attribute when data were combined across all countries. Speed was the least important attribute in all countries except the UK. The most important attribute – convenience or confidence – varied by country. Given that no test is going to be optimised in all attributes, the authors concluded that there is no optimal ‘one size fits all’ test that would be well accepted across western Europe.

Whenever you are ready, click ‘next’ to proceed to a summary of the DCE steps.