Getting more from the Secondary Care Medicines Data using the dictionary of medicines and devices

In the previous blog we highlighted that the varying units and quantity scales within the Secondary Care Medicines Data (SCMD) present challenges when comparing the quantities of different products and suggested that additional information about individual products is needed to make sensible comparisons. Below we describe how we can use the dictionary of medicines and devices (dm+d) to find this information. This will contain a lot of detail, so if you’re not already familiar with the dm+d, we recommend first reading our dm+d explainer and our post describing how it relates to the SCMD. We’ve added a glossary at the end to help reference the technical terms and acronyms used throughout this post.

We need to know more about the products in the SCMD

There are lots of different products issued within secondary care (over 12 thousand!). We’ve previously explained that products are reported at the level of Virtual Medicinal Products (VMPs), with the raw data looking like this:

By itself, this does not allow convenient grouping of individual products, but as the VMP codes represent identifiers within the dm+d, they can be used to identify additional information about a product, like what group of products it belongs to, the product form, its chemical content or its route of administration.

The structure of the dm+d

The dm+d is made up of 5 main classes of information, one of which is VMPs. The other classes and how they are associated is shown in the figure below. The most important when working with the SCMD are Virtual Therapeutic Moiety (VTM) and VMPs.

Diagram adapted from the NHS Business Services Authority

Each of the 5 main dm+d classes has its own set of attributes (such as a unique identifier and name), and is associated with subclasses which themselves have attributes to further define and categorise medicinal products.

The figure below shows the subclasses associated with VMPs, which we’ll go through individually below. If you want to look at the structure of the dm+d in more depth and see all of the information it contains, then we recommend looking at the data model description provided by the NHSBSA.

Virtual Medicinal Product (VMP)

Let’s start by looking at the attributes a VMP can have. The table below shows some of the attributes which we will be using for the VMP, Paracetamol 500mg soluble tablets (you can see all of the attributes for this VMP using the OpenPrescribing dm+d browser). Together these attributes make up the unit dose form information for the product.

A unit dose form is a single measured quantity of a medicine, packaged and ready for administration to a patient in a single unit. For example, a tablet or a vial to be injected. Not all products have a unit dose; the Dose form field indicates if it does. A unit dose form is either Discrete (countable units like tablets or capsules), Continuous (measurable quantities like liquids or creams) or Not applicable (products where the concept of a dose doesn’t make sense, such as colostomy bags or catheters - these are medical devices used for collection or access rather than delivering measured amounts of medicine).

The other fields are only filled if the unit dose form is discrete. These include the unit of measure for the unit dose and an indication of the size of the unit dose. The size of the unit dose also has a unit of measure, which may not always be the same as the unit of measure of the unit dose itself. For example, a liquid medicine can have a unit dose measured in spoonfuls, and these spoonfuls can have a size of 5ml.

Virtual Therapeutic Moiety (VTM)

The highest level of dm+d is the Virtual Therapeutic Moiety (VTM). VTMs are an abstract representation of medical substances or devices, without any indication of strength or form. VTMs are associated with VMPs in a hierarchical relationship; each VMP is associated with one VTM, but a single VTM can be associated with many individual VMPs. For example, the VTM Paracetamol is associated with 38 different VMPs.

This is the easiest way to group products within the SCMD but there are a couple of details to be aware of:

• VMPs within the same VTM can have different forms and different units of measure, so aggregation of quantity is not always straightforward.
• VTMs can be a single substance (e.g. paracetamol), a combination of substances (e.g. Budesonide + Formoterol), or different VTMs for different salts (e.g. Perindopril erbumine and Perindopril arginine).

Virtual Product Ingredient (VPI)

A VMP can have associated ingredients (remembering that hospital stock data includes products such as devices which aren’t expected to have an ingredient). This information is contained within the VPI class. VPIs may have an associated strength, which can be given as a single value (the ingredient strength numerator), or two values (the ingredient strength numerator and denominator) to represent strength ‘per’ for ingredients such as solutions where there may be e.g. 10mg dissolved in 1ml of solution. Each strength value has an associated unit of measure.

The table below shows the VPI for Paracetamol 500mg soluble tablets. This product contains a single ingredient, paracetamol, which has a strength represented by its numerator alone.

An example of a VPI with both a numerator and denominator is shown below.

Route and form

The route of administration and form of VMPs is captured by 3 classes within the dm+d: Ontology Form and Route, Dose form and Drug Route. An example of these is shown in the table below.

Let’s go through these one at a time.

Ontology Drug Form & Route

Each VMP is associated with one or more codes with a textual description that represent the form and route of administration. For example, the form and route for Paracetamol 500mg soluble tablets is solution.oral. This is called the ontology form and route because this is specific to the dm+d ontology and doesn’t use existing codes from coding systems such as SNOMED CT.

Dose Form and Drug Route

There are two additional fields separately indicating form and route(s) for each VMP. Unlike the above field, these are SNOMED CT codes. The form here represents the dispensing form rather than the administration form, so although similar to the field above, will not be exactly the same. Our example product is dispensed as tablets but these tablets are dissolved in liquid and administered as an oral solution.

A VMP always has a single form but can have multiple routes (e.g. an injection can have both the intramuscular and intravenous route). Since the SCMD represents stock control data, there is no way to verify if the reported issued quantity has actually been used. Even when a medicine is administered, the SCMD does not contain route information. For products with a single route listed in the dm+d, this is likely the route of administration used, though exceptions exist - for example, the contents of an injection vial might occasionally be given orally. For products with multiple routes listed in the dm+d, we cannot reasonably assign a specific route.

Linking the dm+d with SCMD

Above, we have shown that having the SCMD in dm+d format makes the SCMD much richer than it initially appears. This extra information is available by linking the SCMD data with the dm+d using VMP codes. Once this data is linked, you can start to think about different ways of measuring the quantity of medicines issued beyond what is available in the raw data. For example, you can determine the quantity of individual ingredients within products (this is what we’re calling Ingredient Quantity, which we’ll discuss in a later post).

When we’re building OpenPrescribing Hospitals, we want to allow sensible analysis and comparison of as much of the data as possible. Understanding the dm+d and how it can be used is critical for this.

In a future series of blog posts, we will explore the different ways to measure quantity. We’ve tackled this problem before with OpenPrescribing but the SCMD has its own set of challenges. If you know more about this please get in touch!

Glossary