Tackling Antibiotic Resistance: Using the National Antibiotic Prescribing Dashboard to understand antibiotic prescribing in your region

Manchester icon Greater Manchester

Posted on the 27th February 2019

Researchers at the Greater Manchester Connected Health City (GM CHC) in collaboration with Public Health England have developed a National Antibiotic Prescribing Dashboard, which provides healthcare stakeholders and policy-makers with access to detailed analysis of antibiotic prescribing in primary care, and related results describing patient-level outcomes. The purpose is to feedback interesting and novel findings that will inform and educate, as well as ultimately drive behavioural change using a data driven approach.

Why is optimisation of antibiotic prescribing important?
Infections caused by antimicrobial resistance have increased by 35% since 2013, which on average contributes to more than 2,000 deaths each year in the UK. The link between overprescribing and the growing problem of antibiotic resistance has been highlighted by the World Health Organization and the NHS, along with the urgent need to address inappropriate prescribing.

In January 2019, the UK government unveiled a five-year strategy to tackle the threat of antibiotic resistance, including a target to reduce resistant infections by 10% by 2024. It is clear that significant measures need to be employed to achieve this goal.

GM CHC research has found large variability in antibiotic prescribing across the NHS, resulting in the ineffective targeting of patients who are at a high risk of developing infection-related complications requiring emergency hospital admissions and over-prescribing to those with low risk. This highlights the need to optimise antibiotic prescribing to ensure that the patients who receive antibiotics are the ones who really need them.

How is the National Antibiotic Prescribing Dashboard addressing this challenge?
The National Antibiotic Prescribing Dashboard aims to reduce the burden of antimicrobial resistance (AMR) through the optimisation of antibiotic use in primary care. Using data from more than 600 practices and over 40 million consultations, it generates novel data-driven insights into the use of antibiotics in the UK, providing information on the key drivers which influence prescribing in different areas.

Who will benefit from accessing the National Dashboard?
The National Antibiotic Prescribing Dashboard is a free resource which is designed for policymakers, researchers and healthcare professionals who want an overview of antibiotic prescribing in the UK and an insight into the factors impacting prescribing in their region. The purpose of the dashboard is to inform healthcare strategies relating to the usage of antibiotics, which are relevant to the specific drivers impacting prescribing in different regions.

How does the Dashboard address policy priorities?

  • It will assist to keep people healthy and support economic productivity and sustainable public services through the provision of evidence and monitoring for optimising the use of antibiotics in primary care by focusing on patients’ risks of hospital admissions, thus tailoring performance indicators to the challenges faced by a GP practice.
  • It provides the data infrastructure to evaluate the cost-effectiveness of innovations around antibiotic use, ensuring accountability of the health and care system to Parliament and the taxpayer.
  • It gives an overview of current prescribing and resistance trends throughout the country, highlighting regional differences and identifying where the most effective healthcare interventions can be made.

What useful information will the Dashboard provide?
The dashboard is divided into three broad areas, to include: benchmarking of practices prescribing, variability and appropriateness of antibiotic prescriptions, and patient risk scores associated with adverse outcomes. The analytics displayed help the user to answer important questions such as:

What factors drive antibiotic prescribing in primary care?

Comparing antibiotic prescribing rates between practices throughout the UK is undoubtedly complex, as many factors influence how much or how little a practice is prescribing. Using linear regression (specifically negative binomial regression) we can uncover the relationship between practice level prescribing rates and potential drivers of prescribing. A series of model comparisons aims to assess the relative contribution of each driver in explaining variation in prescribing rates, following the approach of Judd, McClelland, and Ryan (Data Analysis: A Model Comparison Approach, Second Edition, 2008).

Figure 1: Drivers of antibiotic prescribing in the UK together with their relative contribution. Relative importance is expressed as the percentage of explained variation of prescribing rates between practices.

The main drivers of antibiotic prescribing can be divided into four main categories with the relative importance of each factor expressed as a percentage. This plot demonstrates that differing practice characteristics can explain around 60% of the variation in antibiotic prescribing in the UK. For example, practice location accounts for 13.5% of the variation in antibiotic prescribing, whereas the number of consultations for lower respiratory tract infection explains 17.2%. The analysis shows that despite there being many immutable factors such as the makeup of a practice’s patient population, there are some aspects that could be changed to optimise antibiotic prescribing.

How much does primary care prescribing match the recommended guidelines?

Prescribing to guidelines is important to ensure that the most effective treatments are given to patients, as well as to maintain consistency of care throughout general practice. The dashboard summarises analysis using data from the last three years to show how antibiotic prescribing compares to guidelines, and highlight those conditions, which antibiotic prescriptions are deviating from the recommended guidelines.

Figure 2: The raw consultation counts (y-axis) and the proportion of consultations where the prescriptions were in agreement with the prescribing guidelines (blue) and the proportion of consultations where the prescriptions deviate from the prescribing guidelines (red), for each infectious condition based on the most recent first- and second- line recommended antibiotic treatment in national guidelines (NICE; PHE: GMMMG).

It can be seen that for upper respiratory tract infections there are a large proportion (almost 80%) of prescriptions that do not match the recommended guidelines, which is something that merits further investigation to uncover the explanation behind this disparity.

Additionally, outer ear (otitis externa) and sore throat infections also have a significant proportion of inappropriate prescriptions; this is in stark contrast with lower respiratory tract infections and urinary tract infections where prescriptions deviate from the guidelines by less than 1%.

Does a patient’s underlying risk level determine the chance of them getting an antibiotic?

Recent research indicates that a significant proportion of antibiotic prescriptions are unnecessary, potentially because the infection the patient is suffering from is viral, or the patient is likely to recover without treatment. This research corroborates what we find in our analysis when investigating how a patient’s risk level affects their chance of getting an antibiotic. The plot below shows the proportion of patients receiving an antibiotic split by their risk level (as assessed by a clinical risk prediction model) of developing further complications as a result of their infection.

Figure 3: The antibiotic prescribing rate (y-axis) against stratified risk group (x-axis). Patients are categorised based on their predictive risk (as assessed by a clinical risk prediction model) of experiencing a poor outcome if left untreated. The red line shows theoretical ideal prescribing, whereas the blue line shows the actual prescribing by risk in primary care.

This shows that the probability of getting an antibiotic is largely independent of a patient’s underlying risk level. It also highlights a clear way in which antibiotic prescribing can be more efficient: by getting GPs to consider important risk factors when prescribing, we can hopefully reduce superfluous usage and provide more targeted healthcare.

Accessing the National Antibiotic Prescribing Dashboard
The new dashboard is available to those working in a healthcare, research or government institution.

To access the dashboard, you should:
1) Register for an account here
2) Wait for an email from action@manchester.ac.uk to confirm your account’s approval and click on the link in the email to update your password.
3) Update your password & access the dashboard.

Register for a National Antibiotic Prescribing Dashboard account here

Read more posts...

Connected Health Cities North East and North Cumbria

Connected Health Cities in the North East and North Cumbria hosted a half day workshop to share some of the key lessons and outcomes of the three care pathway projects funded in the region. They...

Posted 08 Jul 2019

CHC proposes new Code of Practice for data use

Connected Health Cities has proposed a new Code of Practice for the conduct of research using data held in CHC Trustworthy Research Environments, which might be undertaken by public or private research organisations. The four...

Posted 25 Jun 2019