Background The National Data Opt Out Programme was implemented in 2018 to enable users of the English NHS to electronically opt out of sharing their patient information for purposes other than their direct care. It has been reported that opt outs may affect the reliability of data used to evaluate services and conduct public health research; however, biases arising from opt outs have not previously been quantified. The aim of this study was to describe the extent to which rates of birth and maternity outcomes at Clinical Commissioning Group (CCG) level may be biased by patient opt outs.

Methods We selected one common and one rare maternity/birth outcome: the rate of deliveries with caesarean sections and rate of births with very low birth weight. Average 2016 rates (per total number of deliveries/births) for both childbirth indicators are published online by Public health England. The percentage of total patients that have opted out as at 31 December 2018, by CCG, is publically available through NHS Digital. We simulated outcome rates across each CCG had opt outs not been applied to the data.

Results As at December 2018, the median CCG opt out rate across England was 2.4%, ranging from a minimum of 0.3% to a maximum of 10.1%. The average published proportion of deliveries with caesarean section was 27.3% (95% CI 25.6, 29.1) and births with very low birth weight 1.18% (95% CI 0.84, 1.67). For the caesarean section indicator, our simulation produced an average minimum value of 26.6% and a maximum value of 29.3%. For the very low birth weight indicator, the average minimum was 1.15% and average maximum 3.88% (2.32 times the published upper confidence interval).

Discussion There is substantial geographical variation in the proportion of patients opting out of sharing their NHS data for research and planning. As shown by this simulation, even in areas with average rates of opt outs, published health indicators may be biased and not reflect the true picture of health outcomes. This is particularly the case for rarer – and potentially more severe – events. It is essential that patients are informed of the implications of opting out when deciding whether to consent to sharing their health data for research and planning purposes.