Update May 2021 Based on feedback received from users, the NHS dictionary of medicines and devices has decided to cease maintenance of the “hospital only” value at AMP level (it remains at AMPP level). As a result this means that we are no longer able to produce our “Hospital Only measure” and we have retired it.

Just prior to the COVID-19 emergency we launched a new measure on OpenPrescribing, primary care prescribing of medicines defined by the NHS dictionary of medicines (dm+d) and devices as “hospital only”. In this blog we set out what exactly this means, describe some of our early investigations and interesting (but nerdy) details, and ask for your help in improving and maintaining this dm+d field.

Introduction and Background

We have written previously describing the [NHS dm+d]/blog/2019/08/what-is-the-dm-d-the-nhs-dictionary-of-medicines-and-devices/). Briefly, dm+d is the standard dictionary for the medicines and devices used across the NHS. It contains standardised codes, descriptions, and metadata (such as price and pack size) for every entry, and its use is mandated across the NHS in electronic systems dealing with medicines. One of the dm+d fields, “Restrictions on availability”, can denote a medicine or device as “hospital only” in any system that uses dm+d.

What is “hospital only” prescribing?

The dm+d editorial policy states “this is a medicinal product where the manufacturer has stated that the product should only be used in hospitals” and provides no more detail. We know that many OpenPrescribing users will be very familiar with local lists or formularies of “hospital only” medicines but our measure only pertains to those designated as “hospital only” by the dm+d.

Why is this important?

We were very interested to see a single list, maintained nationally, of “hospital only” medicines that we could use to examine primary care prescribing, especially after our own investigations around potentially unsafe prescribing of zuclopenthixol and Calum Polwart’s analysis of high dose dexamethasone prescribing in primary care. Although the dm+d is not a decision support tool, its use is mandated across the NHS in software systems including the ones GPs use to write prescriptions. The “hospital only” designation can be used to inform what is displayed to GPs and help avoid situations where “hospital only” products are prescribed accidentally in primary care, which could be unsafe or costly.

What is the pattern of prescribing “hospital only” prescribing in primary care?

You can see prescribing of “hospital only” medicines per 1000 patients for England here (and Figure 1) and as always we have built a measure for every single practice, primary care network, clinical commissioning group, sustainability and transformation partnership, and region. Over the last decade, prescribing of “hospital only” medicines has been steadily increasing and in the last 12 months, 150,513 items have been dispensed at a cost of £3.5million.

Figure 1. The number of items that have been dispensed in primary care and currently designated “hospital only” by NHS dm+d (live data on OpenPrescribing).

What types of medicines are “hospital only”?

You can interrogate the full list of items which are designated “hospital only” using our advanced dm+d browser search facility and all those prescribed on our All England measure. Based on our own experience and feedback from OpenPrescribing users, we have grouped items into four broad groups for discussion.

Definitely “hospital only”

A lot of medicines that are “hospital only” are what we think most OpenPrescribing users would expect. Although the editorial policy is limited on criteria it seems manufacturers have designated products as “hospital only” for reasons of safety or funding allocation. For example in the medicines license for sunitinib it explicitly mentions that a specialist cancer doctor must be involved in prescribing and in England these are generally found in hospitals.

Data entry errors: Probably not “hospital only”

We found a small list of products where we think there might have been a data entry error and a medicine should not be designated “hospital only” in the dm+d; for example Moonia — a branded version of desogestrel, a common contraceptive. This is a very small list and in a database as large as dm+d some inaccuracies might be expected; we have sent the list to the dm+d team for further investigation. If you spot any other examples please contact the dm+d team at nhsbsa.dmdenquiries@nhs.net and let us know so we can ensure our measure is updated.

Probably “hospital only” in the past but over time this has changed

We found some medicines which we think may have originally been “hospital only” but we think it may no longer be appropriate to designate them as “hospital only”. This could be due to doctors’ increased experience with the product and services being established in the community to give care closer to home, or perhaps the reasons for a medicines use have changed. We saw some other medicines like injectable antibiotics and palliative care medicines that we thought may be ok to prescribe in primary care with appropriate advice and supervision but this has created some debate on twitter already so it may not always be clear cut. If you think there are products which should no longer be designated “hospital only” please contact the dm+d team at nhsbsa.dmdenquiries@nhs.net

Missing “hospital only” designation

We think there are products in dm+d that should be designated “hospital only” but aren’t and for these, the dm+d team need your suggestions. Our own investigations have resulted in the manufacturers requesting that zuclopenthixol acetate be designated as “hospital only”. Callum Polwart has also written about high dose dexamethasone and restricting it to hospital use only. If you think there are products which should be designated “hospital only” please contact the dm+d team at nhsbsa.dmdenquiries@nhs.net or the manufacturer.

Anything else?

There is one other important nerdy detail for understanding our measure. The “hospital only” field in dm+d is specified at the Actual Medicinal Product (AMP) level which means they are real1 products and not the generic Virtual Medicinal Product. The NHS Business Services Authority manages the dm+d standard, and has responsibility for its accuracy and the overwhelming majority of prescriptions they process are compliant with this. However the BSA publishes prescribing data in a less granular form using BNF codes. Multiple AMPs may link to a single BNF code and it is possible that each of these AMPs has different “hospital only” designations. Sometimes this is warranted; for example with colistimethate injection where there are slight differences in licenses between products or perhaps a company has made a product that they have chosen to issue only to a hospital, while their competitors will supply primary care. It is also possible that errors exist here or manufacturers have not updated their dm+d listing as the place of use has changed over time. If you spot any of these please let us know and we can remove from our measure.

This is an experimental measure and as always, we welcome your feedback, and would be delighted to hear if you think a measure of “hospital only” prescribing may be useful. To get in touch, drop us an email on bennett@phc.ox.ac.uk or visit us on Twitter at @openprescribing. Additionally the dm+d team and manufacturers may need your suggestions on what dm+d listings may need updating! Get in touch with the dm+d team at nhsbsa.dmdenquiries@nhs.net. Contact details can generally be found for manufacturers by searching https://www.medicines.org.uk/emc.


  1. It turns out some AMPs are not real products. We have discussed this on our GitHub repo and will write more about it when we have time. ↩︎