Medicines and exposure to other medical and lifestyle factors immediately before and during pregnancy or when mothers are breastfeeding can have an impact on their babies. These impacts can be both beneficial (for e.g. babies of vaccinated mothers being protected against infections) and harmful (e.g. developmental disorders in children as an adverse effect of maternal treatment). The recently published Cumberlege Review entitled ‘First Do No Harm’ (IMMDSReview 2020) includes recommendations on strengthening the post-marketing surveillance and long-term monitoring of outcomes related to the use of medicines and medical devices. Two of the cases they investigated were about the safety of tests and medications in pregnancy which were thought to be linked to miscarriages, birth defects and developmental delays in offspring. Anonymised primary care patient data can be valuable for monitoring the impact of maternal exposures on their offspring provided the mother’s health record can be linked to her children in a robust way without compromising patient confidentiality. The aim of this study is to develop a methodology to link mothers and their babies in the CPRD Aurum database to facilitate such studies. More specifically, we will evaluate the feasibility of linking mothers in CPRD Aurum to their children using information on birth and delivery details.
This is a descriptive feasibility study of delivery and birth codes recorded in CPRD Aurum to assess the feasibility to link mothers with their children in CPRD Aurum.
We will select two cohorts of patients from a sample of 100 CPRD Aurum practices: 1) women of childbearing age in CPRD Aurum, defined as women aged 12 to 49 years old who have a Medcode that indicates pregnancy or delivery any time after their recorded start date, and 2) all patients in CPRD Aurum whose recorded start date was within the first year of life (i.e. baby). We will then develop an algorithm using practice ID and family ID to match women of childbearing age to their children using delivery/ and birth Medcodes. All matches will be scored according to the level of matching (e.g. exact delivery / birth Medcode date match, same month and year match, same year match) to provide a level of confidence in the match. We will test the algorithm by reviewing the electronic records of a sample of mothers and their matched children. Once the algorithm has been developed and tested it will be shared with CPRD for implementation of a mother-baby link in the entire CPRD Aurum database so that it can benefit public health surveillance and the wider research community.
Linkage of mothers to their children in CPRD Aurum through development of a mother-baby linkage algorithm.
Susan Jick - Chief Investigator - BCDSP - Boston Collaborative Drug Surveillance Program
Catherine Vasilakis-Scaramozza - Corresponding Applicant - BCDSP - Boston Collaborative Drug Surveillance Program
Eleanor Yelland - Collaborator - CPRD
Jennifer Campbell - Collaborator - CPRD
Katrina Hagberg - Collaborator - BCDSP - Boston Collaborative Drug Surveillance Program
Puja Myles - Collaborator - CPRD
Rachael Williams - Collaborator - CPRD
Rebecca Persson - Collaborator - BCDSP - Boston Collaborative Drug Surveillance Program