Data matters: How do we know what’s happening with breastfeeding?

Data matters: How do we know what’s happening with breastfeeding?

Indicator 10 of any World Breastfeeding Trends Initiative report is about that country’s monitoring and evaluation systems. It may not seem an exciting topic but it’s essential to collect robust data on infant feeding to know what the breastfeeding rates are and how mothers are experiencing services. Without having monitoring data how can services be evaluated and then improvements planned?

For small projects, feedback from mothers may be the most effective evaluation but for larger projects and sizeable areas, the percentages of babies being breastfed (called prevalence) at particular ages help to monitor what is happening. Figures don’t capture the ripple effect of support, though, such as a mother who’s been helped on her breastfeeding journey then supporting friends or being more likely to breastfeed a subsequent baby or deciding to train as a peer supporter.

The WBTi UK report in 2016  found that data collection and analysis had reduced considerably since the ending of the 5-yearly infant feeding survey. There is variation between the four nations with England collecting the least data.

Changing systems in England

In England, record level data on infant feeding is currently submitted by service providers to NHS Digital as part of the Maternity Services Dataset (initiation/first milk feed) and the Children and Young People’s Health Services Dataset (6-8 weeks). Whilst these new datasets are reaching full maturity, NHS England and Public Health England are publishing official statistics on an interim basis for breastfeeding initiation and breastfeeding at 6-8 weeks respectively. Both sets of data are assembled from aggregate data submitted on a voluntary basis by service commissioners. These two breastfeeding indicators are included in the Health Improvement part of the Public Health Outcomes Framework (https://fingertips.phe.org.uk/profile/public-health-outcomes-framework).

Data on infant feeding at birth are submitted by maternity units. Babies were previously counted as breastfeeding if they went to the breast at least once in the first 48 hours. However, this is now changing to a record of the baby’s first milk feed. This is captured as part of the Maternity Services Dataset and NHS Digital publishes monthly reports on the statistics of all the indicators in the data set: as well as annual data.

Local authorities have responsibility for ensuring that their commissioned providers of the universal health visiting service submit infant feeding data for babies aged 6-8 weeks as part of the Children and Young People’s Health Services Dataset (which is about to be renamed the Community Services Dataset). As this data collection is not yet mature, aggregate infant feeding data is submitted to PHE on a quarterly basis. Official statistics are produced annually and quarterly

The Early Years part of PHOF, including these indicators, is maintained by the National Child and Maternal Health Intelligence Network.

In theory, the information at 6-8 weeks provides a picture of what is happening for the whole population but, disappointingly, there are quite a number of gaps. In most cases the local authority submits data to PHE but it cannot always be published as official statistics as a result of validation failures. This is most often due to too many records of ‘unknown’ breastfeeding status.

In one area, the health visiting service has found a solution to this by including a mandatory field on infant feeding in the electronic record of the questions about maternal mood at 6-8 weeks.

There needs to be information on at least 95% of the eligible population of babies (called coverage) to be valid and thus included. This results in some gaps in the published statistics even though the underlying data is available.

 

Try it yourself!

There is now a facility to compare different sets of the annual data, such as comparing gastrointestinal or respiratory infection rates with breastfeeding rates. §If there is sufficient correlation between two datasets, a red line appears. It doesn’t prove there’s a causal link though. You can try this for yourself here:
You select the Region, Area and Indicator you are interested in and then another Indicator for the Y-axis. You need to tick ‘add regression line’ to see if there is a correlation.

Using data to advocate for services

However, even if a service is well-evaluated, that does not guarantee its continued existence, as occurred with the Blackpool peer support Star Buddies programme, although this can be a useful tool in challenging actual or proposed cuts. Zoe Walsh, in her speech to Blackpool council (at 4mins 30s in the recording) in September 2017 used data as part of her clear explanation of why the service needs to be reinstated.

Data collection counts!

 

Cover photo credit: Paul Carter

30. Photo for WBTi MAINN presentation

Patricia Wise is an NCT breastfeeding counsellor and a member of the WBTi Steering Group.