Is breast best in the west of England? Use of linked routinely collected data sources to quantify the protective effect of breastfeeding against infectious gastrointestinal and respiratory illness during the first year of life

Abstract

BackgroundThe potential health benefits of breastfeeding, including the protective effect against common childhood infectious diseases, are well known. This project sought to quantify the reduced risk of admission to hospital during the first year of life due to these illnesses, in Bristol, UK, to help to build the business case for local breastfeeding promotion, and to generate information specific to the population of the city. MethodsTwo routine data sources were linked: child health records and hospital episode records for 17 260 Bristol resident children born during 2009–11. Children missing any of the required variables were excluded, as were low birthweight and preterm infants, and babies admitted to hospital in the first 48 h of life, to reduce the effect of reverse causality on any association observed. The primary exposure variable was feeding status at the 6–8 weeks check, and the primary outcome was hospital admissions for gastrointestinal and respiratory infectious illness during the first year of life, based on diagnoses codes used in similar studies. Relative risks of admission were estimated with a multiple regression model, adjusting for a range of known confounders: maternal age and smoking status at delivery, ethnicity, deprivation, birthweight, gestation at delivery, and parity. Ethics approval was obtained from South Central NHS Research Ethics Committee and the University of the West of England Faculty Research Ethics Committee. FindingsThe cohort consisted of 12 918 children (75% of all births in the population). Significant (p<0·05) risk reductions associated with a reduced risk of admission for infectious gastrointestinal illness were found for any, exclusive, and partial breastfeeding, compared with no breastfeeding (odds ratio 0·371, p<0·001; 0·361, p<0·001; 0·359, p=0·003). They were also found for partial and any breastfeeding and risk of admission for infectious respiratory illness (0·766, p<0·001; 0·633, p=0·005). InterpretationBreastfeeding was associated with a significantly reduced risk of admission for infectious illness during the first year of life. The findings suggest that an increase in breastfeeding prevalence could bring considerable savings to the local health-care economy. The data linkage method using pre-existing routinely recorded data provided useful and robust findings in line with research typically requiring costly and lengthy primary data collection, and offers potential for more research of this type to be conducted closer to local service provision in future. FundingThis research was conducted as part of a postgraduate study funded by Bristol Primary Care Trust. Their successor organisation (Bristol Clinical Commissioning Group) acted as host with their agreement obtained through the Avon Primary Care Research Collaborative. None of these organisations had any influence on the design or conduct of the research. I had the final decision to submit the abstract for publication. I had and still retain access to all data used for the analyses that contributed to this research.