By Rick Payne, PhD researcher at The University of Portsmouth, and Nicole Holgate, Communications Manager at DataKind UK

The volunteers and staff at Starlight's DataDive weekend

“It really helped us understand how we could make better decisions as an organisation.”

Dr Sandy Gulyurtlu
Head of Insights and Impact, Starlight

Hospitals and healthcare providers in the UK are struggling with a long-term, upward trend in the need for children’s services. But, wonderfully, there is already evidence that play in healthcare significantly improves children’s experience of healthcare. Fear and distress are reduced, and health play specialists’ support can mitigate the pain and trauma of treatment. Whether during treatment, care, or recovery, play can help children’s wellbeing and resilience, and reduce treatment times and costs. It’s crucial to the health and wellbeing of children.

That’s why DataKind UK was excited to work with Starlight, a UK charity focussed on ensuring that every child can experience the power of play in hospitals and other healthcare settings. The charity provides a range of boxes containing toys and activities to help children engage with play. This empowers them to cope with their illness, or with potentially traumatic treatment and procedures. Last, but not least, Starlight collaborates on research on the impact of play in hospitals and advocates for more and better play services.

What did they want to find out?

‘How can we best allocate our resources?’ was the key question Dr Sandy Gulyurtlu, Starlight’s Head of Insights and Impact, had for DataKind’s volunteer data scientists. They want to understand where there is overlap between their services, hospital play provision, and donors so they can decide what services to provide based on need, and advocate or fundraise appropriately. They are particularly interested in communities and locations that show gaps in provision, or disparities between different ethnicities and demographics.

Much of the data (such as number of child admissions, level of play provision, specialist staff, and play budgets) had to be collected through freedom of information (FOI) requests that came from 140 hospital trusts and health boards across the UK. The response rate was 87%, and there were some gaps in the data for those that did respond. This suggests that potentially useful data is not being collected.

Data collection and ‘wrangling’ (cleaning data to make it useable) are the most time-consuming part of most data science projects. Geospatial data, used to produce map visualisations, was processed so it could be used more efficiently — a lesson learned from a previous DataDive project with Material Focus.

DataDive weekend volunteer Luke said, “I love connecting with people doing similar roles in different organisations, it’s my community. It’s fantastic to help people do good with data and see what is possible.”

What did they learn?

How to collect more useful and accurate data

Collecting the information through FOIs was laborious and time-consuming, so Starlight were thrilled to find out that a small part of the data for England could be extracted directly from NHS Digital. Not only does this mean that some of the analysis can be updated more regularly, but it also helps Starlight to verify and check their existing data. As the analysis progressed, Starlight also learned how they might improve their FOI requests to gain more precise and useful information.

Deprivation doesn’t affect service provision

An interesting finding was that Starlight’s resources were spread evenly over deprived and less deprived areas. Some regions had more eligible trusts than were receiving play provisions, while others had a lot more hospital admissions than there were boxes distributed. This indicates that there is lots of opportunity for Starlight to be more strategic about directing its help to places with the greatest need.

They love dashboarding!

Starlight also came away with prototype dashboards that would help staff across the organisation quickly assess provision and fundraising in different regions, and be able to drill down to data about specific hospital trusts to support their local knowledge. “We love dashboards — they have opened up a whole new world of how we can showcase what we do. They will completely revolutionise our work,” said Sandy.

This was also the first DataDive project for DataKind’s new Data Science Project Manager Caitlin Loftus, who said “I found my first DataDive weekend a really inspiring experience. The amount of time, effort and care that the volunteers put into making sure the DataDive project was a success was amazing to see, and Starlight’s energy, enthusiasm and openness made them a joy to work with.”

What’s next?

Starlight got a lot out of seeing what could be done with their data and visualisations, and the potential for creating dashboards. They will be able to use the analysis and prototypes to examine service provision and follow up on why some hospital trusts are providing significantly more play for children than others. They can also tackle why donor levels are growing in some areas, but declining in others.

In terms of the data itself, Starlight can improve their data collection, and further examine hospital admissions data. Some of the methods used during the project, run with updated data, will help them to monitor changes over time. Three of their staff are training further to maintain this data collection and analysis. All of this will greatly help their team to interact with NHS trusts and hospitals around the country, so that more seriously ill children are able to experience the benefits of play.

Sandy added, “You’ve had such tremendous impact on our organisation, and my mind is blown with how talented and skilled you all are, and what a difference you have made to us.”

Huge thanks to the project’s Data Ambassadors, Tom, Jess, Lakshmi, Claudia, and troubleshooter Péter. And a big thank you to Softwire for hosting the DataDive weekend in their offices, in particular to Laura and her colleagues, who gave up their weekends to make sure everything ran smoothly.

Improving children’s recovery with Starlight