[MI] Medicines and Pregnancy Registry - Antiepileptic use in females aged 0 to 54 in England: April 2018 to September 2023

Medicines dispensed in Primary Care refers to all prescriptions that have been prescribed in England and dispensed in the UK, typically by community pharmacies and dispensing doctors. For September 2020, over 95% of prescriptions had their patient details verified by NHS England’s Master Person Service (MPS) providing a solid basis for a patient register based on the medicines they have been prescribed.

The current analysis is restricted to a reporting period of 66 months of data (April 2018 to September 2023). The aim is to improve this over time as new data becomes available and earlier prescribing data is added back to April 2015.

Linking to both maternity and HES datasets provide two significant benefits, getting early sight of pregnancies through the maternity dataset whilst also using HES data to validate birth outcomes for the maternity episodes. In addition, any pregnancies that only appear in one of the two datasets are captured. HES data is however, provisional for the latest financial year and therefore not all pregnancies may have been captured for the latest 6 months (April 2023 to September 2023). In addition, births that occurred post September 2023 in the HES data and are not captured in the maternity dataset may result in a slight undercount of pregnancies for the latest 9 months of the reporting period.

Patients have been identified and linked across datasets if they have passed the validation criteria set by NHS England’s Master Person Service for each individual dataset. A female may therefore have been prescribed valproate and have had a pregnancy but if any inaccuracies occur in any of the three data sources their data will not have been verified and identified in the analysis. Only the accurate and verified data will have been included.

Not all conceptions captured in the data from the maternity dataset resulted in a birth. For example, a female could have had a miscarriage early in her pregnancy and then conceived very soon afterwards. Previously both conceptions would be treated as a full term pregnancy in the analysis resulting in a very slight over count in some of the figures. Further work to de-duplicate any overlapping pregnancies attributed to the same female (owing to conflicting source data) has been undertaken in the latest publication. Pregnancy records are subject to additional validation procedures to de-duplicate wherever possible.

Early miscarriages that occur prior to a booked appointment within the maternity dataset are not captured in the data.

Prescribing data as defined earlier in the report is currently only captured by month. Any prescribing identified to have taken place during the month of conception or the month of delivery may in fact have taken place prior to conception or after delivery. 

Due to the combination of the limitations stated above with regards time frames, coverage and quality of the administrative data, the number of females and pregnancies in the report may be slightly undercounted.