Apparently, NHS patients admitted to hospital on a Sunday are 16% more likely to die in hospital than those admitted on any other day. This fact nicely demonstrates the danger of leaping to conclusions without looking at the whole picture, and shows why statistical analysis is absolutely crucial to effective clinical trials.
There are several questions you need to ask to try and get to the bottom of the situation…
- is there anything different about the way a hospital operates at the weekend?
- when do most hospital patients die?
- is there anything different about the patients who are admitted to hospital at the weekend?
There are many other questions you could ask, such as, what is public transport like at the weekend – might patient have deteriorated whilst waiting for the bus – but let’s stop with the three above.
The difficulty is to know when to stop asking questions. It can be easy to stop too early just because you think you have found the answer. This is why trial protocols are so important and are created BEFORE the trial begins. Trials are expensive and complex - it would be awfully tempting to stop when the “facts” seem to have been borne out by the initial results.
The UK press has stopped at the first question – because fewer of the most experienced staff are present at the weekend the lack of expert skill is clearly causing these deaths.
This view has permeated the media and even government rhetoric – with a drive to a fully 7-day NHS – in part to resolve this lethal situation.
It is a shame that few people seem to have asked the second and third questions…
Apparently most people who die in hospital do so on a Wednesday.
OK - the dangerous weekend theory could still be true. As long as you assume that the poor medical care people received on a Saturday and Sunday takes a few days to cause death.
However, the answer to the third question blows everything out of the water...
People admitted to hospital at the weekend are much more likely to be emergencies and are on average, far more seriously ill than those admitted during the week.
One fact about the way the NHS operates at the weekend does seem to have a bearing on the situation – but it’s nothing to do with the availability of consultants.
Terminal patients who have elected to die at home may be admitted to a hospice if they take a turn for the worse during the week. If this happens at the weekend, they will go into hospital instead.
I’m not suggesting that the situation is quite this simple but it is certainly not the simple result of dangerous weekend understaffing.
So back to experimental design…
The example above shows just why experiments should never be designed to prove a point but to explore the facts. The rules of Good Clinical practice are clear on the vital importance of experimental design and accurate, thorough analysis. The rules of GCP state that “Clinical trials should be scientifically sound, and described in a clear, detailed protocol.”
For those who need a quick reminder, you can download all 13 principles of ICH GCP here...
The information is taken from the online course ICH GCP in English
Pet theories and ideologies have no room in true science – leave those to politicians and reporters.
|
|
 |
|
|
Share on LinkedIn |
Share via Email |
