I've just read the key bit of the Public Health Scotland report about the statistics of discharges from hospitals into care homes and outbreaks of the virus.

It's pretty clear to me it says discharges do increase the chance of an outbreak.

https://beta.isdscotland.org/media/6224/2020-10-28-discharges-from-nhsscotland-hospitals-to-care-homes-report.pdf

The key information at the bottom of Table 10. The 2.88 tells us that, all else being equal, hospital discharges into the home nearly triple the probability of an outbreak occurring and the fact the confidence interval is 2.31 to 3.60 says this is statistically significant.

But they adjust this to account for other factors (not clear how). Then the hazard ratio drops to 1.21, ie the probability increases by 21% if there are discharges. But, the confidence interval is 0.94 to 1.54; the 0.94 is why it's termed not statistically significant. But...

...what does 0.94 mean? It means that if you discharge patients from hospital into a care home you decrease the risk of an outbreak! What this confidence interval is really telling us is that the data *and* the model used to adjust it are being pushed to a limit of reliability.

Another point of concern is that the size of the care home and occurrence of discharges are different types of characteristics. The former is very broad and not specific to any one aspect of introducing the virus whereas the latter clearly is. That makes comparison problematic.