What Was COVID-19, Exactly? (part 5)

I investigate a cohort of 6,865 adult in-hospital deaths using split file, staged multivariate logistic regression and present an estimate for 'residual COVID'

In part 3 of this series I began on a staged multivariate logistic regression model in the prediction of an ICD10 COVID-19 diagnosis in the EPR of 3,558 adult in-hospital deaths that occurred between 2020/w10 and 2020/w21 (first wave), and 3,307 adult in-hospital deaths that occurred between 2020/w39 and 2020/w50 (second wave) within an unknown NHS Trust.

We discovered what appeared to be ‘discrimination’ between the sexes, with females far more likely to receive a COVID diagnosis than their male counterparts during the first wave. Such discrimination evaporated during the second wave which made me a wee bit suspicious so I took a detour in part 4 that opened a can of worms with regard to COVID and cancer. But… back to that there model!

That There Model

With the stage one and stage two demographics in place within the model it is now time to roll out the third stage, and that is to submit the following 12 diagnostic indicators in a conditional forward selection procedure in order to dete…