First off, a sincere thank you to all front line medical workers across the globe, particularly the NHS. You are all heros in the face of a dangerous situation and your contribution is felt by your patients and the public alike. We can never repay you for the sacrifices you are making and we all pray an event like this will never be seen again. To everyone else, if you are so moved, why not join the NHS Volunteer Army and help save lives today!
The UK has been criticized for its inability to accurately and confidently report on the COVID-19 death rate, facing accusations from ineptitude to cover up. I posit that it is the same issue most of us face or experience in our workplaces everyday, complacency.
If you work in an office, you know that spreadsheet that manages everything. You know the one. It takes ages to open, is covered with lots of colours and tabs. And what would you do without it, eh? I mean if that went down you and your co-workers would be in trouble. I suspect this is a image many of you can relate with. We love our data and value it highly, but it is always on a knife’s edge, just one right click away from oblivion. Why is that? Is it lack of skill with excel, lack of time? Does it suit its purpose so no need to upgrade? No seriously, I am curious, please message me.
How does this relate to COVID-19 you ask? Let’s go on a roundabout journey about data processing systems. When I briefly worked for the Nursing and Midwifery Council I was shocked at coroners reports for two reasons;
First, there was a website with all of them on it! 😊
Second, they were jsut links to PDF’s (sigh)… 😞
Just halfway to a useful data source.So I built and XML schema to grab the PDF’s so I could see what was new each day, but colleagues still needed to trudge through the body of the PDF to identity relevant facts that required investigation. Again, a half measure. I am certain when that API was set up to publish reports people were high fiving; “Look at this improvement we made!” And they should feel good for that. Some access is better than no access. But data systems have a tendency to evolve this way, organically, piecemeal, problem driven. It is only when you realize you need to mangle two sources together that you sigh and get on with it.
So COVID-19… The uK’s death rate is unfortunately one of the highest in Europe according at the moment, sitting at over 21,000 as of writing this post. But there has been criticism as well as a spotlight shown on this;
Coronavirus: Far more deaths from COVID-19 crisis than headline numbers show - Sky News
Can we trust the Covid-19 death toll? - Channel 4
Coronavirus in the UK: why calculating the death toll is so difficult -The Times
That second headline is a read duzey, nothing like a bit-a-clickbait Channel 4. The third link is actually very good and you can read that article here. Now lets, do as we do here at Known Unknowns, strip the passion and the rhetoric from the politics of reporting and information dissemination and get to the heart of this problem. First off, why is reporting a death toll so important? There might be many answers to that; one might be pandemic planning and containment, another might be national defence, social services, economic forecasting, etc (As to why the public needs to have a minute to minute perfect figure is beyond me but I am happy to be corrected).
So assuming we need a death toll how do we calculate it? Well the simple A to B would be if someone dies, you add a tick mark to the board. Simple! But wait a minute, we need to track it by day. Okay, record the day. And the time. Why not location? Okay wait a second. We said, death toll, not detailed analysis of this subject. And therein lies the rub ladies and gentlemen. We just took a simple problem and made it very complex very quickly. Just like that office spreadsheet that does everything. It used to be a list and now it is so much more!
But let’s keep going with our death toll example! Each of the NHS’s 207 clinical commissioning groups will have their own staffing (that will also vary with size, sickness and availability), reporting systems/ clerical management. If they are rural it will take longer for autopsy or records to get finalized. Do they report live or do they report daily, weekly? And what date do we report, the day someone died, then a doctor made the official call? If they died of something else but they had COVID-19 at the time what do we put down? I could go on, but I hope you get the picture.
Problems we think are simple are actually hard, and while it is not perfect, it is the best we got. Much like that critical office spreadsheets. Its no one’s fault, it is a by-product of that organic development of data I spoke about at the start. That is not to say we shouldn’t demand better, absolutely the government should be able to provide better figures faster. But what are we doing with our own data that makes us any better? Do we keep everything up to date all the time? Have we explored a better system for managing, analysing and protecting our data? Is it secure, integral and accessible? It is always easy to throw rocks at others, especially when we look up at the government. We say “but you are in charge! So why don’t you have this?” The same could be said for us.
So what’s the solution? One thing I always try to encourage others to do in their data systems is create a roadmap of what they ‘think’ they might want to ‘know’ or ‘show’ in the future and what that might look like. The government likely didn’t think a global pandemic was going to hit them but I can guarantee you they will now have a roadmap for better data capture and analysis in the future, ie; the NHS tracking app that is due to be launched. I would encourage everyone reading to stop and think next time they read a headline like “Can we trust the Covid-19 death toll?” to ask; How would I answer that and what pitfalls might I face in getting to that solution? Not only will it make you more sympathetic to people who hold spreadsheets, just like the one at your work, but you will further the innovation and efficiency of you business by imagining the art of the possible.
Demand better from your data, but recognize we are all on that journey.