Understanding the Secondary Care Medicines Dataset

At the Bennett Institute we are building OpenPrescribing Hospitals. We want to enable any interested user to conduct analyses on the Secondary Care Medicines Dataset (SCMD). We are excited to be working on this project and are grateful to the teams at Rx-info for their heroic work curating this dataset and also to the NHSBSA for hosting the data. The SCMD is critical to the development of the platform so let’s start with a brief description of what the dataset contains, why it is useful and how it can be used.

What is the SCMD dataset?

The SCMD dataset is processed pharmacy stock control data in a dictionary of medicines and devices (dm+d) format (we’ll describe the dm+d format below). This represents the distribution of medicines within hospitals, such as the issuing of stock to a ward. So whilst it isn’t a direct measure of medicine use, it is a generally reliable surrogate for actual usage (though it does come with some quirks, which we’ll discuss in the next blog).

This means that when we write about what we find in the data, we use phrases such as ‘More doxycycline was issued in winter months’. We don’t use phrases such as prescribed because issued products may not have been prescribed to an individual. This is an important distinction from the English prescribing data (the source of primary care data used in OpenPrescribing) which is derived directly from individual prescriptions that have been dispensed for an individual patient in primary care.

Where does the data come from?

The data within the SCMD originate from pharmacy stock control systems within NHS trusts. This data is collected, aggregated and curated from all NHS Acute, Teaching, Specialist, Mental Health and Community Trusts in England by Rx-Info.

The dataset is published by the National Health Service Business Services Authority (NHSBSA) on their Open Data Portal.

How much data is there?

The SCMD was first published in August 2020 with monthly data dating back to January 2019. Since then, monthly data containing about 300 thousand rows has been published, so the SCMD contains a little over 20 million rows of data as of late 2024.

This is a lot of data but it is small in comparison to the English prescribing data, which contains approximately 1.6 billion rows recorded since September 2010, rising at about 17 million rows/month.

What data does the SCMD contain?

The dataset includes all medicines and devices issued by NHS trusts in England which have a dm+d code. The dm+d is the standard dictionary for medicines and devices used across the NHS and contains standardised codes, descriptions and metadata for individual items. The dm+d has five components which we’ll discuss in depth in a future blog, but you can also find more detail in our blog What is the dm+d?

All of the data within the SCMD is reported at the same level of the dm+d: Virtual Medicinal Product (VMP). VMPs describe a general class of medicines or device that may be available as an actual product. An example of a VMP, as shown in the diagram below, is Paracetamol 500mg tablets. For each VMP, the dm+d indicates the ingredients in the product and their strength as well as the form the medication takes (e.g. tablet) and its routes of administration (e.g. oral). Unlike Actual Medicinal Products (AMPs), VMPs do not indicate the manufacturer of the medicine.

Diagram adapted from the NHS Business Services Authority

There are over 10 thousand VMPs issued within NHS trusts in England, but trusts only submit data for medicines and devices they use. Most big acute hospitals typically only report issuing around 3 thousand VMPs in a given month with specialist and community trusts using far fewer.

Each trust included in the SCMD submits a quantity for each VMP they issue each month. Alongside each quantity, is the assicuated unit of measure (there are lots of different units of measure, but that’s a topic for a future blog!) and an indicative cost (we’ll also describe what this means in a future blog, but in short, this is most often not what the NHS pays as actual prices are confidential).

Why is this dataset useful?

The SCMD allows anyone to analyse how hospitals use medicines. Hospital use of medicines is a large and rapidly growing part of NHS spending (over £10bn in 2024 and rising at over 8%/year). Knowing which medicines are used helps understand whether the most cost-effective choices are being made. But hospital teams also need to know whether they are using medicines safely; the pattern of use of specific medicines can help identify issues for hospital pharmacists, and when new drugs are launched it is useful to monitor their uptake.

We will continue to write about the work we are doing to build OpenPrescribing Hospitals. If you have any comments or ideas for the platform. Please get in touch!