Had Pipple fulfilled the greatest wish of GGD GHOR Nederland, the Eindhoven econometricians would have written world history. But even they could not predict how the coronavirus pandemic would unfold. But with their extrapolation model, they still contributed to the averaging of the crisis.
Testing and vaccinating and testing with big data
Organizing a vaccination program, every local Municipal Health Service (GGD) could already do that with their eyes closed. After all, they already vaccinated babies on the assembly line. Analyzing statistics was also a piece of cake: population screening is one of their core tasks. But streamlining processes around corona tests and vaccinations with big data was a bridge too far for them at the time.
Not surprising, notes Nicolette Verduin, who started working as a corona manager at GGD Hart voor Brabant shortly after the corona outbreak. ‘The analyses that GGDs normally make are based on historical data. There weren’t any now. In addition, babies present themselves according to time-honored cohorts. Hordes of snotty people with an unknown virus are a lot more difficult to interpret.’
Forecast model
As a corona manager, Verduin was constantly bombarded with questions: how many people can we expect? How many come by car? How many m2 do you need? And how many staff and equipment? Nicolette and her team asked Pipple to create a forecast model with which they could calculate scenarios themselves. Their main guide: the reproduction rate of the RIVM (the infamous ‘R’). ‘If we test an x number of people with today’s R, then an R that is expected in six weeks will probably report x people for a test’, Verduin explains the model.
Nowadays Verduin works at GGD GHOR Nederland: the association for the 25 directors of Dutch GGDs and the Medical Assistance Organizations in the Region. As one of the program directors of the national Corona Program Organization, Verduin looks at and indicates the continuity in the care chain at a national level. She does this by uncovering, sharing and interpreting relevant and accurate figures with her teams. When Verduin started nationwide, she was looking for a tool to identify bottlenecks in the care chain at an early stage and with which she could visualize any developments: tools with which she could enter into discussions with the chain partners. The model of GGD Hart voor Brabant could be an excellent starting point for this, she thought. And so she also brought Pipple on board at GGD GHOR Nederland.
Bottlenecks in the chain are ahead of
Together with the healthcare professionals and the information experts of GGD GHOR, Pipple expanded the extrapolation model in such a way that it could calculate various consequences for a wider range of parties deep into the chain. Every time GGD GHOR encountered a new relevant variable, the model was enriched with it, Verduin explains. ‘For example, if there were x people in hospital, we knew that a percentage of them would return home after so many days. A large part of them needed extra oxygen. To ensure that we had sufficient oxygen supply in the right places, we added that variable to the model.’
Multi-purpose
The project made a significant appeal to Pipple’s flexibility. It turned out to be big. “They ended up in an unspread bed. Even our own teams didn’t fully know each other. Because everything was so ad hoc, we could only tell Pipple what the end result should be. And that it had to be done quickly.’ After only one week there was a first draft, and after four weeks a working model. ‘That version had all kinds of open ends,’ says Verduin. ‘There was no other way, because we didn’t know much yet. We were already very happy with it, but for Pipple it felt unnatural. They want it to be right for their customer. It’s amazing how patient they’ve been with us.’ The extrapolation model is still being expanded. “Maybe it will even survive covid. It is useful for many more logistics processes.’
