The HSE and the Government provide details of the daily number of confirmed Covid patients in hospitals as well as the number admitted and discharged in the previous 24 hours. The daily data for these measures are available to view at their website[1] since mid-March to early April. I speculated that the admissions and discharge data might indicate how many patients are getting the virus while hospitalised for other reasons as they will not be recorded in the Covid patient admissions data.

It is simplest to look at the data from the period in the summer when the number of confirmed Covid patients in hospitals was close to zero. By the end of July, the figure was only 6 confirmed Covid patients nationally and that was chosen as the convenient starting point for the analysis. Using the Covid confirmed number of patients in hospitals on July 31st of 6 and adding the 1 admission in the subsequent 24 hour period and subtracting the 1 discharge the calculated number of patients on August 1st was 6. The next admission figure was added to this and discharges subtracted from it for the following day’s number of hospitalised Covid patients and so on until December 22nd. This is the calculated series of the number of Covid admitted patients to hospitals nationally shown in the graph below in green. Also shown in blue is the series (from the same website) that officially records the actual number of confirmed Covid patients reported in hospitals each day.

 

Clearly, there is a very large divergence between the 2 series (but a very similar pattern). This leads to one of two possible conclusions: either the admissions/discharge data is highly unreliable or else many of the confirmed Covid patients didn’t get admitted to the hospital as ill with Covid and requiring hospital care for same. While the data is likely to have some minor quality issues, it is highly improbable that official Government & HSE pandemic data is so misleading. Furthermore, there is consistency between the Covid admissions and discharges data with the figures for the period indicating that there were 63 Covid admitted patients in hospitals nationally by December 22nd when all admissions and discharges for the period are taken into account. But the officially recorded number of Covid patients on December 22nd was 237 so where did the other 174 come from? Either the patients must have been admitted to hospital for another reason and were subsequently tested and found to be Covid positive (either having become infected before hospital admission or else getting infected while there). The data suggests that on December 22nd 73% of Covid patients in hospitals weren’t admitted to hospitals for having Covid.

Looking into this issue a little further, there were 1,586 Covid patients admitted in the period and 10,120 patient-days in the calculated series (2 patients for 5 days would be 10 patient-days). This gives an average hospital stay of 6.4 days per patient. A recent journal article in BMC Medicine Journal[2] estimates hospital stays as generally between 3 and 9 days internationally so this average appears reasonable. However, the same calculation for the officially recorded series has 23,532 patient-days. This means that 13,412 patient-days (57%) were not due to patients admitted to hospital for Covid in the full period. Using the same average stay per patient, these patient-days equate to around 2,110 patients in the period. But it is likely that patients only detected as Covid positive in hospital are not as ill as those admitted for it so their patient-days are likely to be less which would mean there were more than 2,110.

Given events over the last few days and plans to re-enter a Level 5 lockdown shortly, it is interesting to note that the uptick in hospitalised Covid patients recently is only in the non-Covid admitted ones and not the Covid-admitted patients. It is an indication of detection or infection issues in hospitals and not increased infections and subsequent illness in the community. At least not yet.

If the official government data is reasonably error free, then it is a reasonable conclusion that the majority of Covid patients in our hospitals were not admitted to them for Covid care. Given the well-known problem of infections in hospitals generally and the infectiousness of Covid-19, it is also reasonable to conclude that many of the non-Covid admitted confirmed Covid patients are getting infected in the hospitals themselves.

NPHET and their modellers are desperate to find out how most Covid infections are occurring and it is ironic that hospitalisation may very well be one of the riskiest activities one can do when it comes contracting Covid!

 


 

Joe Henry

[1] https://covid-19.geohive.ie/pages/hospitals-icu–testing
[2] COVID-19 length of hospital stay: a systematic review and data synthesis, Rees et al