Prescribing percentiles explained

The prescribing percentage and the percentile are different

A measure first defines a numerator and a denominator: for example, selected prescription items divided by all the eligible items in that measure. If an illustrative practice has 12 selected items out of 200 eligible items, the proportion is 6%. These are made-up numbers to explain the calculation.

The percentile then places that proportion among eligible practices for the same measure and month. A percentile of 80/100 does not mean 80% of prescriptions are affected, 80 patients need a review or the practice is 80% safe.

Read the measure definition and its direction before interpreting a higher or lower value. A public aggregate cannot tell you whether an individual prescription was appropriate.

How AegisRX calculates its independent comparisons

When the prescribing check shows independently calculated NHSBSA comparisons, it uses a selected subset of published OpenPrescribing measure definitions. It does not reproduce the entire OpenPrescribing service. Each result links to its definition and shows the data month, source and comparison details.

The eligible group comprises practices in the dated NHS England practice directory with prescribing data for that month and a positive denominator for that measure. A practice with no eligible items is excluded from that comparison; a genuine zero numerator with a positive denominator is retained.

For this independent calculation, the percentile is 100 multiplied by the number of eligible practices with a strictly lower proportion, divided by the total eligible practice count minus one. Equal values share the lowest rank in their tied group. The comparison requires more than one eligible practice.

Illustrative example: if 20 of 101 eligible practices have a strictly lower proportion, the percentile is 100 × 20 ÷ 100 = 20. If several practices share a value, they share its rank. The result is a position within this group, not a clinical target.

Check the source and month on the result

NHSBSA’s English Prescribing Data contains information about prescriptions issued in England and dispensed in England and several other UK and island jurisdictions. It is published monthly. It is aggregate prescribing data, not a live view of the patient record. Read the NHSBSA dataset description.

AegisRX may display published OpenPrescribing results when available, or its clearly labelled independent NHSBSA calculation. The independent cohort and use of monthly presentation codes can differ from OpenPrescribing’s methodology. Values from the two sources should not be treated as interchangeable.

The prescribing month, retrieval or calculation date and page date describe different things. For a like-for-like comparison, check the actual prescribing month and measure, rather than assuming two recently loaded pages contain the same period.

Why a high or low percentile needs context

• Small denominators: a few items can move a proportion substantially.
• Different populations: practices serve patients with different needs and clinical circumstances.
• Changing peers: eligibility and the distribution of prescribing can change between months.
• Recorded activity: aggregate data does not establish the indication, individual reasoning or monitoring completed.
• Coverage: an unavailable measure is not the same as a zero value or a reassuring result.

Use the figure to choose a question for a pharmacist or clinical team to investigate. Do not change a patient’s treatment solely because of a percentile.

A practical way to review a result

1. Confirm the practice, prescribing month and named data source.
2. Read the exact measure definition and check the numerator and denominator.
3. Look at the eligible peer count and whether a small number of items could explain the result.
4. Ask the responsible clinical team to examine the relevant local records and context.
5. Record what was found and, if action is appropriate, agree a review method using consistent definitions and periods.

The free prescribing check is a starting point for that discussion. A practice’s detailed medicines monitoring and recall work needs its own clinical records, agreed scope and professional review.