In 1662, English businessman John Graunt studied London’s mortality records to understand causes of death over the years. This was one of the first times data was used to identify broader population trends.
This approach underlies many public policies today. Governments use data to understand patterns, pinpoint needs and design better measures. This has been especially crucial in the pandemic, when nations had to learn and adapt quickly to cope with a public health crisis.
What lessons can we glean from some of these data initiatives in the past year? Data chiefs from Singapore and New Zealand shared their insights at the recent webinar on ‘How governments can equip public servants to use analytics’.
Analytics in action
Data can be useful when serving a population with diverse needs. Singapore’s Central Provident Fund Board (CPFB) oversees a comprehensive social security system that covers all citizens young and old, rich and poor alike. Citizens can use CPF savings for retirement or to pay for healthcare and housing.
When the pandemic hit last year, the Board created a data dashboard to monitor the impact on citizens and identify groups who might need support, shared Gregory Chia, Group Director of Policy, Statistics & Research at CPFB.
The dashboard highlighted that some citizens who used CPF to pay for housing may face difficulty sustaining their mortgage payments if they lost their jobs or had their wages cut, he revealed.
CPFB shared its analytics with the relevant agencies, which then rolled out new relief measures. These include payment deferments to give citizens more time to repay loans, the Board says. It also targeted outreach to this group to inform them of these measures.
The Board has started using prescriptive analytics to plan for service centres resources. This is a relatively new approach: instead of just predicting outcomes, data can tell us how to achieve a desired outcome, explained Soo-Kiat Loo, Head of Advanced Analytics & Big Data at NCS.
CPFB built a simulation model that looks at the number of appointments a centre can take in a day, how long each appointment should be, and how many counters should be open. The model then advises the optimal balance.
More broadly, prescriptive analytics will come in handy as nations continue to allocate Covid resources and schedule vaccinations, NCS’s Loo added.
Data has certainly been useful in coordinating Singapore’s pandemic response. Healthtech agency IHiS managed to quickly set up a system for consolidating Covid test results from across labs and e-health records.
The team then standardised and interpreted these results so the data could be passed on to contact tracing teams, noted Christine Ang, Deputy Director, Emerging Capabilities-Health Insights at IHiS.
The agency will be exploring personalised care next, she shared. Hospitals could pull together genomic, clinical and lifestyle data to design more targeted treatments for patients.
Trust and transparency
The best algorithms would fail if citizens didn’t trust them enough to use them. As governments build new data tools, it must remember to involve citizens in the transformation journey.
IHiS is working on creating explainable AI, which will help make algorithms fair and transparent. For instance, if an algorithm predicts that a patient is at risk of a particular disease, doctors would be able to explain how it came to that conclusion.
New Zealand is also prioritising transparency in its algorithms. Stats NZ, its official data agency, developed a series of guidelines for government agencies to explain simply how they use data and what algorithms they use to make decisions.
The guidelines will be reviewed in July 2021, a year after it was introduced. “We absolutely recognise that there are areas that will need to evolve over time, particularly as technology changes,” said Eleisha Hawkins, General Manager for Data System Strategy and Capability at Stats NZ.
Build data capabilities
With these possibilities in mind, how can governments begin to transform with data?
For starters, it will take more than a small team of data scientists or a few mavericks on the leadership board. “You need almost everyone in the organisation to be comfortable working with data,” said CPFB’s Chia.
He encouraged giving employees a safe space to experiment. “A lot of the data work sometimes is about just playing with it and seeing where you get. Sometimes you don’t get anywhere, but you learn something along the way,” he said.
Loo from NCS noted the importance of making training relevant for public servants’ everyday work. Organisations that adapt training materials for their own use cases often achieve better results than those who dole out standard explanations of a tool’s features, he shared.
New Zealand will make its upcoming data maturity assessment framework as simple as possible. “These maturity type assessments are often really, really detailed. There’re lots of questions and so only a true data geek would get into them,” Hawkins said. Stats NZ wants to make it easier for “even the chief executive” to understand and engage.
The agency is also working with universities to launch data courses for public officials. These micro credentials could help them build confidence in working with data.
Data remains a key tool for fighting disease outbreaks, after Graunt’s first venture almost 400 years ago. As governments tap analytics for the pandemic and beyond, they will need to build citizen trust and public sector skills.