Changes to OpenPrescribing.net measures

Over the last 18 months we haven’t had the opportunity to make many changes to OpenPrescribing.net. This is due to a number of reasons including the need to focus on OpenSAFELY and the team spending more time working in the NHS to support the COVID response and vaccine rollout.

Recently we started reviewing a number of measures that needed the most urgent attention. As a result a number have been updated and a couple have been retired. These changes are all outlined below.

Updated Measures

Low/medium intensity statins

A user emailed us pointing out that some higher dose rosuvastatin were showing as a lower intensity statin, when it is high intensity. Following some investigation we discovered that a capsule formulation of rosuvastatin is now available on the UK market (significantly more expensive than the equivalent tablet). The measure is built in such a way that high intensity statins have to be explicitly excluded. As we weren’t aware of the new formulations, this hadn’t happened, and they were consequently erroneously included in the numerator. We have now fixed this. There were very few prescriptions of the capsules prescribed, and therefore users are unlikely to notice a difference.

Geeky aside: this is the sort of measure that would benefit from programmatically identifying the strength of a product using its dm+d entry, so that new formulations are automatically included (e.g. in this example

WHERE virtual_product_ingredient = 'rosuvastatin calcium'
	AND strength_value_numerator >=10
	AND form_route IN ('capsule.oral','tablet.oral'))

Low priority Lutein and antioxidants

Occasionally the NHS Business Services Authority changes the BNF code associated with a product. A recent change in designation meant that this measure has been showing zero prescribing in many cases. This was due to a change in BNF codes for the specific products listed in the measure. We have now updated the codes and should now be working correctly again.

Environmental impact of inhalers

When we originally created this measure we came across a slight issue: the NHS Dictionary of Medicines + Devices (dm+d) had defined “softmist” inhalers (such as Respimat devices) as pressurizedinhalation.inhalation and therefore would be classed as a pressurised MDI (and therefore in the numerator) in our measure. We therefore chose at the time to exclude all “softmist” inhalers from both the numerator and denominator until this was resolved.

Having discussed with the company that manufacture softmist devices, they requested a change in the dm+d classification, and it was subsequently re-classified as inhalationsolution.inhalation. We have now added this group into the measure denonminator.

If your organisation uses significant softmist devices (or uses very little), you may notice that your measure value and decile position has changed.

Retired Measures

We have also retired two measures from the site.

Nebivolol 2.5mg tablets

This measure was created in order to highlight the significant difference in price between nebivolol 5mg (£1.40 for 28) and 2.5mg tablets (£40.41 for 28) in 2016. Since then the price difference has reduced significantly, and therefore we have decided to retire the measure.

High Cost Angiotensin-Receptor Blockers

This measure was one of the first measures created on OpenPrescribing, and identified when ARBs cost more than 3 times the generic equivalent. In recent years there has been significant fluctuation in the cost of certain generic ARBs. In addition, as one of the earliest markers, the “high cost” drugs need manual curation, rather than the programmatic approach we have used in more recent measures. As a result the measure is less accurate that we would have liked it to be, and therefore we have retired it.

Most of these measure updates are as a result of user feedback, and we’re always delighted to hear from you. If you find something that is wrong, that you think could be better, have an idea for a new measure, or just want to tell us how great we are, you can contact us via email or on Twitter.

One other thing: at th eBennett Institute we are always looking for ways to improve on our processes. Given the way the site works and the measures are calculated we’ve always used a staging site to test our measures in the past. This is fine, but becomes an additional thing that has to be managed and kept up-to-date. Because of this Dave Evans has created a new process which allows us to deploy new or changed measures to a special part of the live site simply by typing a couple of commands into Slack. This is a awesome update to our processes which will enable us to check our changes much more quickly, and with the latest data. Thanks Dave!