We’re going to build OpenPrescribing for hospitals: here’s why

A sneak peek of our OpenPrescribing Hospitals prototyping… full online open version coming soon.

Medicines are the most common intervention in healthcare. We’re building a new service called OpenPrescribing Hospitals, to help health care professionals use open hospital medicines data to review medicines use quickly (because we know everyone is busy across the NHS but particularly on the frontline!) This is a screenshot of our beta version, which is still in development and will probably change quite a lot over the coming months.

We want users to be able to quickly see:

• Are they choosing the best evidence based medicines?
• Are the medicines they use as safe as possible?
• Are they issuing the best value medicines for patients?

In fact, you might recognise these reasons for building OpenPrescribing Hospitals as the principles of medicines optimisation! Similarly for those of you who have never heard of “medicines optimisation” but are keen to understand medicines use in your hospital, maybe you are a CEO, ward sister or patient representative, we want to make a tool that will make analysing hospital medicines easier and faster!

Why us?

We (at the Bennett Institute for Applied Data Science) are a mixed team of software developers, clinicians and academics who (amongst other things) build live, interactive data-driven tools and services for everyone to use. We already run a platform called OpenPrescribing that can be used to analyse all prescribing data in primary care in England. OpenPrescribing.net is very popular. It has 20,000 unique users a month, and 5,000 users subscribe to our newsletter. The OpenPrescribing platform is only possible because these data are openly available and have been since about 2010. Every month sees the release of about 17 million new rows of data. This has provided a rich dataset for those looking to carry out medicines optimisation activities in primary care.

OpenPrescribing makes medicines optimisation easier and faster in primary care

OpenPrescribing is one of a number of useful tools that has been widely used to support improvements in evidence based, safer and better value prescribing by GPs across England.

The OpenPrescribing platform has a number of features to support medicines optimisation, including bespoke analyses, pre-built measures, and tooling such as our improvement radar. We would like to build the same for hospital medicines data.

We have published some of our work using openprescribing.net and our approach to data driven medicines optimisation. Below are a handful of examples of the analyses that are possible using openly available data.

• Adherence and uptake of guidelines. We looked at adherence to NICE 2014 guidance on lipid modification which recommended the use of high-intensity statins for both primary and secondary prevention. We reported the proportion of low-/medium intensity statins of all statin therapy, between August 2010 and March 2019. In this study, our group looked at trends and geographical variation and developed tooling to identify sudden prescribing changes. (Curtis 2020)
• Shed light on unusual medicines use practice. We used our outlier detection tool within OpenPrescribing to identify unusual antipsychotic prescribing. We found extreme outliers in two small geographic regions. Without this tooling it is difficult to know how local practice compares with others. OpenPrescribing makes benchmarking easier. (MacKenna 2022)
• Identify and share best practice. By applying one of our existing OpenPrescribing analysis tools, we were able to rank NHS organisations by reduction in opioid prescribing rates. There is an opportunity to learn from well performing organisations about how they achieve and maintain lower rates of opioid prescribing. (Hopcroft 2024)

We think there are lots of opportunities within hospital medicines data to do similar analyses. For example, we wrote a short letter about variation in the speed of uptake of andexanet alfa, a new reversal agent for two common blood thinning agents. We saw marked variation in how this has been adopted between individual Trusts. We also expect that making secondary care data easy to navigate might enable other users to find new ways to help optimise hospitals’ use of medicines. If you think of them, share your ideas!

If you are interested in reading more about our own and others research using OpenPrescribing please have a look at our website.

What tools are currently available for secondary care medicines data?

As in primary care, hospital specialists do have some tools, but they are not openly available to wider users. This limits the ability to benchmark activity; explore measures of value, safety, and evidence based practice; and disseminate good practice to others.

We want to build tools for hospital medicines data that achieve the same benefits that OpenPrescribing has already demonstrated in primary care. We believe there are big opportunities to use hospital medicines data to optimise medicines use in hospitals, just as it has in GPs.

Why now?

We have been keen to work with hospital data for a long time and have openly advocated for it to be made available. As such, we were delighted when detailed data about what drugs hospitals use become accessible recently. The Secondary Care Medicines Dataset is now available on the NHS Business Services Authority Open Data Portal. All the heavy lifting is done by Rx-info who curate all the data coming in from many many individual Trusts, and data sources within those Trusts. You can check out more about all the work they do to make this dataset happen on their website. We believe the time is right to start building new tools to start using this dataset. We have a small amount of money from NHS England to further develop the idea,

What’s next?

We’ll have more to share with you about OpenPrescribing Hospitals in the coming months, including:

• What is in the Secondary Care Medicines Dataset, and how big is it?
• How should we describe the quantity of medicines issued?
• Trust data within the SCMD - including consistency and completeness of data
• How should we group medicines together?

Get in touch

We’ll post more updates here as we explore the data on some of the key insights, and test some of our ideas with a wider audience. If you would like to be involved in the development of OpenPrescribing Hospitals, please send an email our way: bennett@phc.ox.ac.uk. You can also keep up with updates to OpenPrescribing.net or OpenPrescribing Hospitals by signing up to our newsletter.