This graph shows what the infection rate in a Scottish hospital cancer ward really looks like. The blue column represents the number of patients on any one day. The tiny marks along the bottom represent the average level of infections reported each day.

In April to June 2018 there were around 70 episode counts of bacterial infection in the blood samples of some of the 400 children being treated for cancers in a Glasgow hospital. This was the worst case, April to June 2018, from the graph below. The graph was used to suggest a spike in infection which then contributed to later deaths reported extensively, morbidly, in the Scottish media but there’s a problem with it if you don’t know how to read it and how to think about what it means.

For convenience, it has been massively cropped and squeezed to let us see something that in reality is miniscule. There are typically 400 children in the ward each day and who knows how many more over the three-month periods used for this graph, yet the vertical axis stops far short, at 25 simply for presentation reasons. Rather than show us the daily count, the graph below has also grouped them into ‘quarters’ of 91 or 92 days in order to get some data that can be seen. In Quarter 2 of 2018, there were around 70 counts in the 92 days or less than one per day on average. This was done to help you see something that is in reality barely noticeable and in so doing has created a fake spike where there is really none.

In the graph above, I’ve taken just April 2018 and inserted a single count every day even though there will have been less than that in reality. Because, for accuracy, the graph has to be scaled vertically to at least 400, I’ve had to stretch it to almost a full page to allow it to show even the slightest presence of the bacterial infections.