How comparable is NHS sickness rate data?

On Wednesday we published a Factcheck on a number of claims made in the media about NHS sickness absences.

One such claim was that the number of sickness days taken by professionally qualified (PQ) ambulance staff was five times that of the national average.

An eagle-eyed reader pointed out a typo in our heading which originally read "Is 15 days per year five times the national average?"

Considering that this was referring to PQ ambulance staff, not NHS staff as a whole, it should have read "Is 22.6 days per year five times the national average?" and has subsequently been ammended.

However, rather more interestingly, the same reader pointed out that the two different statistics used to measure this didn't seem to add up.

Considering that the average number of days lost per year among ambulance staff was five times that of the UK labour market as a whole (22.6 compared to 4.5), why was the sickness absence rate for ambulance staff only 3.4 times that of the entire labour market?

Explaining the difference

The first thing to note is that the statistics on UK labour market and ambulance staff are from two different sources: the former being from the ONS and the latter from the NHS Information Centre.

We returned to the reports mentioned in our original article and had another look at their methodology, and found some quite important differences that help to explain this anomoly.

UK Labour Market (ONS)

The ONS's estimate is centred on weekly sample surveys in which respondents record the number of hours missed during the week due to sickness and the total hours that they usually work during the week.

Absence rate

The ONS calculated the absence rate by taking the total hours lost due to sickness or injury and dividing it by the total hours usually worked by employees.

This is then multiplied by 100 to convert into a percentage.

Average number of days lost per year

First the ONS calculates the number of hours lost by the individual throughout the year. This is done by taking the total hours lost due to sickness/injury during the sample week and multiplying it by 52.

Next, the ONS calculate the average number of days lost by the individual throughout the year. This is based on an average day being 7.5 hours long.

Therefore, dividing the total hours lost due to sickness/injury for each individual throughout the year by 7.5 gives us the total days lost.

Finally, if we divide the total days lost by all workers by the total number of workers we get the average days lost per person.

Ambulance staff (NHS IC)

In contrast, the NHS calculates the absence rate and average number of days lost per year using full time equivalent (FTE) data.

Absence rate

The absence rate is calculated by taking the total number of FTE days lost due to sickness and dividing it by the total number of FTE days available including non-working days.

This is then converted into a percentage.    

The NHS IC accept that choosing to calculate the figures based on a 365 day calendar may produce a slight underestimate of the actual absence rate as some workers may not record days that they were ill but were not expecting to come to work anyway.

It is worth noting that many other government departments calculate their statistics based on a 225 working day year.

Average number of days lost

If we take the number of FTE days available and divide it by 365 (the number of days in a year), we arrive at the figure for the number of NHS FTE employees.

Subsequently, we can divide the FTE days lost due to sickness absence by the number of FTE individuals employed by the NHS to arrive at an estimate of the average number of sick days per FTE employee.

Conclusion

So where does this leave us?

The key point to take away from this is that the ONS statistics and the NHS IC statistics do not produce straightforwardly comparable and thus a change in the number of days lost will not necessarily produce an equivalent change in the absence rate of the same magnitude.

So while the orginial claim that the number of days lost to sickness by ambulance staff may have been five times more than other NHS workers is an accurate portrayal of the information available to us, there is much more to the picture than this might suggest, and we need to be careful about drawing too many conclusions from any comparison of these two datasets.

We'd like to finish by thanking Full Fact reader Paul Brown for highlighting this issue, and hope that this might clear up any confusion caused by this complicated issue.