Do NHS death rates rise when new junior doctors start their jobs?
"Death rates for the week in August when junior doctors typically start are 8 per cent higher than at other times of the year".
Daily Mail, 17th July 2012
"Hospital death rates go up by 8 per cent when junior doctors start their jobs in what a top NHS executive has labelled a 'killing season'".
Daily Mail, 22nd June 2012
Following Full Fact's recent analysis of the scale of preventable deaths occurring in NHS hospitals, today's claim in the Daily Mail caught Full Fact's beady eye.
The assumption that it is more dangerous to be admitted around the time that recent medical graduates are making their shaky debut across England's hospitals, was apparently somewhat legitimised in June by Sir Bruce Keogh, Medical Director of the NHS, who announced plans to implement sweeping training reforms for junior doctors.
Sir Bruce told the Mail that the plans, which will see junior doctors compulsorily spending at least four days shadowing senior colleagues before starting work, were intended "to end the so-called killing season" around the time new doctors begin.
But how much does early August merit this ominous nickname? Do death rates really rise by eight per cent around this time as the Mail reports?
Analysis
The figures given by the Daily Mail can be traced to a study done by researchers from Imperial College London in 2009, entitled: "Early In-Hospital Mortality following Trainee Doctors' First Day at Work".
The study reviewed NHS hospital statistics from 175 hospitals over the period 2000 to 2008. For each year, researchers compared two groups of hospital admissions.
The first group of patients consisted of emergency admissions into hospital on the last Wednesday of July, while the second group consisted of emergency admissions a week later, on the first Wednesday of August.
The two sets of patients were then followed for a week, and all deaths in each group within the next seven days were totted up.
As medical graduates begin work on the first Wednesday of August, the aim was to pinpoint any difference new junior doctors may have on patient mortality.
Overall, the researchers found that once factors such as age, gender and socio-economic deprivation had been taken into account, there was a six percent increase in death rates for the first Wednesday in August compared to the previous Wednesday between 2000 and 2008.
Of course, Full Fact's astute readership will have noted that this is two percent less than the figure quoted in the Daily Mail.
However, this errant eight per cent can be accounted for by considering that the study also calculated the increase in death rates between the two groups was two per cent greater if patients with cancer and those which undertook surgery were factored out.
So, although the Daily Mail chose to ignore this distinction and preferred to report the higher variable, it does at least quote this correctly.
But does this increased death rate mean it would be wise to avoid hospital in early August if at all possible? And how much blame can we really lump onto fledgling doctors?
Firstly, we should note that although the rise in death rates between the two weeks forms a fairly regular pattern throughout 2000 to 2008 and is undeniably statistically significant, it is still a relatively modest increase.
As Dr Paul Aylin, the senior author of the report noted:
"Our study does not mean that people should avoid going into hospital that week. This is a relatively small difference in mortality rates, and the numbers of excess deaths are very low."
So much so reassuring, but what about the cause of this rise in mortality rates?
Yes, the increase parallels the influx of junior doctors into the hospital, but as the oft-cited phrase goes, correlation does not necessarily imply causation.
The 2009 study does not include an extensive breakdown of each patient death. As such, it gives no indication of the cause of each death and therefore does not differentiate between deaths which were non-preventable and deaths which could have been prevented if it were not for an error on the part of a junior doctor.
The study acknowledges that it could not ensure that the only difference between the two cohorts was the presence or absence of new medical graduates. There were other factors, such as the severity of the ailment or disease, which were not adjusted. We therefore cannot be sure that any difference in mortality rates was in part, or solely, due to the turnover of junior doctors.
The researchers make clear that further and more detailed studies are needed to draw definitive conclusions about why there is a rise in mortality rates in the first week of August and stop short of blaming new junior doctors for this rise.
Conclusion
The Daily Mail's claim accurately references the findings of a 2009 academic study by researchers at Imperial College London, which found a small, but statistically significant rise in death rates between the first Wednesday in August and the previous Wednesday over eight years.
Although it is true that this rise correlates with the first week of work for new medical graduates, the study does not provide a conclusive account of the cause of each death. We therefore cannot be sure that early August is a "killing season", only that it is an "increased death rate season".
However, the training reform pledged by the Medical Director of the NHS suggest that even without definitive conclusions, the NHS deem the rise likely, at least in part, to be due to these new doctors.