What bed is that, then? (part 4)

A closer look at the non-COVID staff absence trend.

In part 3 of this series I indicated that we observe a gradual rise in the number of non-COVID daily staff absences over the period Apr ’20 – Jan ’22 that was hard to eyeball on the slide provided. Here are the data again but with a suitably scaled y-axis to bring out this detail…

Whilst the dashed red line represents one of those ubiquitous ordinary least squares regressions (OLSR) that spreadsheets like Excel kick out I’ve gone and run a Generalised Linear Model over the distinctly non-Normal data. A Tweedie distribution with identity link yielded an overall slope estimate of 17.5 additional absences per day (p<0.001): our staff are getting sicker!

This slope primarily arises because of elevated levels of staff absence during the winter of 2022/21 compared to 2021/20. It is worth pondering on why this is so – a return of some serious influenza strains perhaps? The steady increase from Apr ’21 to Oct ’21 is most interesting and suggests we cannot pin everything on influenza-like season…