Request

I would like to have access to antibiotic prescribing split by sex and also by age. Ideally, I would like the antibiotic type prescribed by GP practice split by age (in 5year age groupings but with small children separate e.g. something like 0-1, 2-5, 6-10, 11-15, 16-20 etc) and sex for the last 10 years (so 2014 onwards?). I would like the data by month. I am an Associate Professor at the London School of Hygiene and Tropical Medicine and I am asking for this data as I have an MRC CDA fellowship to explore how antimicrobial resistance varies by age and sex and am working on these patterns in the UK. With this data I would be able to better parameterise my models to explore the drivers of antimicrobial resistance and hence better target interventions. I have been in touch with Data Serveice Support for the NHS Business Service Authority who said it should be available and have guided my request.

Your request was received on 19 January 2024 and has been dealt with under the terms of the Freedom of Information Act 2000.

On 8 February 2024, you clarified the following:

For antibiotic type I mean either by individual antibiotic drug name (i.e. gentamicin) or if this is not available by antibiotic family (e.g. aminoglycoside).

On 14 February 2024, you clarified the following:

I was told that it would be available at the GP practice level. I think if grouped by month and the 5yr age bands and sex it should be a relatively large number so should be OK re: statistical disclosure? But not sure what the limits are? Is there a granular level below national but above GP practice level that would be available?

Ideally the number of patients of each age group and sex that received any antibiotic prescription. As the key information for me is the age / sex distribution, I believe that would be number of prescription items given to how many patients in each age and sex group

On 16 February 2024, you clarified the following:

Would it be gender or sex that it would be split by?

Redacting any data points between 1-4 would be fine.

Yes, please still at the GP level and with the parent levels please. For some of the analysis, I’d like to look at correlation with other indicators that are at the GP practice level but would also aggregate for higher level ecological analysis.

Does the data have any associated metadata on these prescriptions? E.g. reason for prescription?

On 19 February 2024, you added the following:

Having thought about this – would it be possible to get at the GP level, and then also at the aggregated ICB level in order for me to check how far off the redacted points make the estimates at the GP level?

Response

The data is too large to send via email. Please visit the open data portal on the below link to receive a copy of this information.

https://opendata.nhsbsa.net/dataset/foi-01671

2014 Data

I am writing to advise you that following a search of our paper and electronic records, I have established that the information you requested is not held by the NHS Business Services Authority (NHSBSA). We only hold patient information from April 2015 onwards.

April 2015 to December 2023

Fewer than Five

Please be aware that I have decided not to release the full details where the total number of patients falls below five. This is because the patients could be identified when combined with other information that may be in the public domain or reasonably available.

This information falls under the exemption in section 40 subsections 2 and 3A (a) of the Freedom of Information Act.

This is because it would breach the first data protection principle as: a) it is not fair to disclose personal details to the world and is likely to cause damage or distress. these details are not of sufficient interest to the public to warrant an intrusion into the privacy of the patients. Please click the web link to see the exemption in full:

https://www.legislation.gov.uk/ukpga/2000/36/section/40

Breach of Patient Confidentiality

Please note that the identification of patients is also a breach of the common law duty of confidence. A patient who has been identified could make a claim against the NHSBSA or yourself for the disclosure of the confidential information.

The information requested is therefore being withheld as it falls under the exemption in section 41(1) ‘Information provided in confidence’ of the Freedom of Information Act.

Please click the below weblink to see the exemption in full.

https://www.legislation.gov.uk/ukpga/2000/36/section/41

The data is too large to send via email. Please visit the open data portal on the below link to receive a copy of this information.

https://opendata.nhsbsa.net/dataset/foi-01671

Data Source:

NHSBSA Information Services Data Warehouse. NHSBSA Prescription Services process prescriptions and information is then used to make payments to pharmacists and appliance contractors (in England) for prescriptions dispensed in primary care settings. This involves processing approximately one billion prescription items and payments totalling over £9 billion each year. The information gathered from this process is then used to provide information on costs and trends in prescribing in England and Wales to over 25,000 registered NHS and Department of Health and Social Care (DHSC) users.

Time Period:

The data covers the period April 2015 to December 2023 inclusive. Each month's data is provided in a separate CSV file.

Contents:

The data shows the number of unique patients and prescription items for prescribing of antibiotic medicines.

Prescribing Organisation:

Data is shown for prescribing in GP practices in England only; the data is presented at practice level and at Integrated Care Board (ICB) level, separately. To allow mapping to parent organisation levels, a lookup file has been provided which shows the sub ICB location, ICB and NHS region for all practices included in the data, based on the current NHS organisation structure. Where a practice has closed in a period prior to the formation of the current structure, it may be mapped to a closed or since-merged Clinical Commissioning Group (CCG), area team or region record.

Dispensing Organisation:

No restrictions have been placed on dispensing location.

Medicines:

Data is shown for prescribing of antibiotic medicines, as defined as those medicines contained in British National Formulary (BNF) section 5.1. The data is presented at chemical substance level. BNF chemical substance code is included in each data file, with the mapping from code to description included for all chemical substances in the lookup file.

Patient Data:

Unique patient counts are calculated for the other attributes shown in each row of the data, and do not include unidentified patients (item counts do include unidentified prescribing). For context, 96.7% of all prescription items included in this analysis can be attributed to an identifiable patient. The figures for the number of identifiable patients should not be combined and reported at any other level than provided, as this may result in the double-counting of patients.

Age Bands:

Patient age bands are based upon the age captured on the prescription form during payment processing. While age is not captured on all forms, in this data 97.4% of all items can be associated with a patient with a known age. Please note that column E, 'AGE_BAND', gives ages grouped as 0-1, 2-5, 6-10, 11-15, 16-20, etc. Certain groups have been recognised as dates and autocorrected: ages 2-5 displays as ‘02-May’, 6-10 displays ‘06-Oct’, and 11-15 displays ‘15-Nov’. The data is unaffected, however if you wish to correct the ages, this can be done using the Find & Replace tool (Ctrl-H) after opening the file: 1. In the ‘Find what:’ box, enter the date ‘02/05/2024’ 2. In the ‘Replace with:’ box, enter the range ‘2 to 5’ 3. Press ‘Replace All’ 4. Repeat steps one to three replacing ‘06/10/2024’ with ‘6 to 10’, and ‘01/11/2015’ with ’11 to 15’ The below link will guide you how to use the find/replace function. https://support.microsoft.com/en-gb/office/find-or-replace-text-and-numbers-on-a-worksheet-0e304ca5-ecef-4808-b90f-fdb42f892e90 Gender: Patient gender has been reported using the latest patient gender information held by the NHSBSA Information Services data warehouse at the time that the data was extracted. This uses information from either the most recent Electronic Prescription Service (EPS) message or from the last time that NHSBSA received data about the patient's gender from NHS Personal Demographics Service. Gender is either: Male, Female, Not Known (where no information is held) or Indeterminate (where information is held which does not confirm either Male or Female).

Redaction of Small Numbers:

Data points relating to fewer than five unique patients or items have been suppressed. Suppressed cells have been replaced with an asterisk ('*') as a placeholder. Where a zero is reported in a cell, this should be interpreted as showing that we have no identified data relating to that category. A placeholder should be interpreted as 'between 1 and 4' or 'a small number' of the fact in question.

Exclusions:

The data excludes - • Items not dispensed, disallowed and those returned to the contractor for further clarification.

• Prescriptions prescribed and dispensed in Prisons, Hospitals and Private prescriptions.

• Items prescribed but not presented for dispensing or not submitted to NHS Prescription Services by the dispenser.

Accessing CSVs with more than one million rows

CSVs with more than one million rows can be viewed using add-ons to existing software, such as the Microsoft PowerPivot add-on for Excel, to handle larger data sets. The Microsoft PowerPivot add-on for Excel is available using the link in the 'Related Links' section below. Once PowerPivot has been installed, to load the large files, please follow the instructions below. Note that it may take at least 20 to 30 minutes to load one monthly file.

1. Start Excel as normal
2. Click on the PowerPivot tab
3. Click on the PowerPivot Window icon (top left)
4. In the PowerPivot Window, click on the "From Other Sources" icon
5. In the Table Import Wizard e.g. scroll to the bottom and select Text File
6. Browse to the file you want to open and choose the file extension you require e.g. CSV

Once the data has been imported you can view it in a spreadsheet.

Please note that this request and our response is published on our Freedom of Information disclosure log at:

https://opendata.nhsbsa.net/dataset/foi-01671