What is it about?
There’s lots of evidence to suggest that people who don’t regularly use their asthma medication get more recurring symptoms and asthma attacks than people who use it regularly, like their doctor told them. Researchers often want to be able to estimate how often people are taking their asthma medicine by using data collected routinely by doctors, hospital staff, pharmacists and others. You can do this by looking for gaps between prescriptions being given to patients and the medicine being picked up from the pharmacy, as well as time after a medicine should have been used up before another is requested. It’s not a fool-proof way to measure medicine taking, but it doesn’t cost much to get the data because it’s already collected for other reasons. Using data that has both the date that the medicine was prescribed by the GP and dispensed at the pharmacy is most useful. If you only have the GPs records, you can’t be sure that the patient collected the medicine. On the other hand, if you only have the pharmacy records, you don’t know if a patient is still being prescribed their medicine. Unfortunately, in the UK, at the moment the GP records and pharmacy records are not linked together automatically, and it’s no simple task as prescriptions don’t have a unique ID that’s also recorded by the pharmacist. We’ve written a set of instructions to find the dispensing record for a prescribed medication as accurately as possible. The data we used to write the instructions, and also to assess their accuracy, was from the Salford Lung Study – a UK clinical trial that assessed the effectiveness of a new medicine for asthma (or COPD) compared with the usual care. Participants only had to visit a clinic at the start and end of the study year, but they could ask for extra doctor’s appointments whenever they wanted, and they picked up their medicine from a pharmacy like normal. Our instructions to find the pharmacy record that matched the doctor’s record converted information into a common format (such as 500 micrograms is equal to 0.5 milligrams) so that the computer could tell if two records meant the same, even if they looked different. We also used knowledge about medicines and pharmacy processes to tell the computer of other differences that were allowed – like if a generic (unbranded) medication was given out by the pharmacist instead of a branded one of the same type. We compared the dataset generated from these instructions to the dataset generated when the computer only looked for identical information from the GP and pharmacy. Our instructions for matching the records found dispensing information for 70% of the asthma prescriptions. This lines up pretty well with other studies which have looked at how often people collect prescribed asthma medications, if a little bit lower than expected. There were some records of medications being dispensed which were not matched to a prescription (17%) – this is probably down to a problem with our matching instructions. When we looked closer at these cases, a lot of them had gaps in the pharmacy records so we couldn’t be confident they were the same prescription.
Photo by Aleksandr Barsukov on Unsplash
Why is it important?
We’ve written a detailed set of computer instructions to find pharmacy dispensing records for prescribed medications, which take into account different ways that the same information might be recorded. This will allow researchers to find pharmacy records and only use collected medicines to estimate more accurately how much people are taking their asthma medicine.
Read the Original
This page is a summary of: Linkage of primary care prescribing records and pharmacy dispensing Records in the Salford Lung Study: application in asthma, BMC Medical Research Methodology, December 2020, Springer Science + Business Media,
You can read the full text:
The following have contributed to this page