National Statistical

There have been claims reported in the news media that around 120,000 young children are not registered with a local doctor. Rich Pereira explains why that is not quite the case.

Over the last week there have been several news stories highlighting the difference between the number of children registered for GP services across England and the number of children in our population estimates.

These stories have wrongly concluded that there are more than 120,000 babies and small children (those aged 0-4) in England that are not registered with a doctor and are consequently missing out on important health checks.

Looking at the data more closely reveals the difference between the GP registrations is considerably smaller, nearer to 18,500. This is primarily because the numbers that the press have compared related to different time periods, but there are other reasons too.

Population size changes over time

All the news stories have compared the latest population estimates for 0-4 year olds, for the middle of 2017, with the latest publication of GP registrations for England, for 1 January 2019. This comparison is problematic as it is comparing two different groups of people.

For example, all those aged 3 and a half or over in mid-2017 would have been aged over 5 on 1 January 2019 and all those aged under 1 and a half years on 1 January 2019 had yet to be born in mid-June 2017.

In short, we would expect there to be a difference between the number of patient registration for 0-4 year olds for 1 January 2019 and the population estimates for 30 June 2017.

Further, given that there has been a downward trend in the number of births in recent years we would expect the number of 0-4 year olds in 2017 to be higher than the number of patient registrations for 0-4 year olds in 2019.

We make regular use of data on GP registrations as part of the production and quality assurance of the population estimates. Part of this involves comparing our population estimates with GP registrations for a comparable period, for mid-2017 this found a difference of 18,500 between the number of 0-4 year olds.

Better late than never, the impact of delays in registration

A second important issue relates to how people interact with GP services. If we look at the difference between the numbers of 0-year-olds and the number of 1-4 year olds in the population estimates and on GP registrations for 2017, we find that there is very little difference for the 1-4 year olds (2,500 more GP registrations) and a relatively large difference for 0-year-olds (20,900 fewer GP registrations).

The explanation for this? Registering with a GP requires action from individuals, the process does not occur automatically. Immediately following the birth of a child there may be a short delay before they are registered with a GP as new parents are understandably pre-occupied.

Because of this the GP registration data will always tend to underestimate the number of 0-year-olds in the population as they will never all be registered on the same day they are born. Given there were around 12,600 live births per week in the year up to mid-2017, an average delay of say 2 weeks between birth and registering with a GP would mean the GP registrations would be 25,200 lower than the population.

In response to this when we use GP registrations to produce population estimates we always allow a month’s grace so that late registrations can be accounted for. Further, if a child is registered solely for private healthcare they will not appear on GP registrations at all, resulting in under coverage.

One final consideration is that the population estimates are estimates of the population. For 0-4 year olds at the England level the estimates will be of high quality but there will inevitably be some small difference between our estimates and the true number of 0-4 year olds.

In a nutshell

While reports that 120,000 babies and toddlers are without a GP make good headlines, they are misleading. While we can never be 100% certain as our statistics are estimates, more comparable data puts the numbers considerably lower.