Are 1,000 NHS patients a month dying due to 'basic errors'?

With the sound of the recent coroner's ruling on the death of 22 year-old Kane Gorey still ringing in the public ear, the Independent reported that the neglect at the heart of Mr Gorey's case may not have been an isolated incident highlighting a new study on the scale of preventable deaths in NHS hospitals.

The headline rang:

"Doctor's basic errors are killing 1,000 patients a month."

This was backed up with some shocking extracts from the report's findings:

"..something went wrong with the care of 13 per cent of the patients who died in hospitals. An error only caused death in 5.2 per cent of these — equivalent to 11,859 preventable deaths in hospitals in England."

So where does this data come from and should we be worried?

The study quoted by the Independent can be traced to a report by researchers at the London School of Hygiene & Tropical Medicine , published on the 7 July in BMJ Quality and Safety, entitled: "Preventable deaths due to problems in care in English acute hospitals: a retrospective case record review study."

The researchers took a sample of 1,000 adult patients who died across 10 acute hospitals in England in 2009. They then used a method called Retrospective Care Record Review (RCRR), which involves an analysis of the 1,000 cases by medical professionals, who judge whether the deaths could have been preventable.

The judgement was done in two stages. The first stage involved, the reviewers considering each case and flagging up any perceived "problems in care" which may have occurred, which could either result from an act of omission or a direct mistake by a clinician. They identified 131 of the 1,000 patients as having a problem in their care.

The second stage of review ranked the 131 cases where a problem had occurred in care on a six point scale, according to how preventable the death is considered to be. Overall, the reviewers judged 52 of the 1,000 deaths to be in some part caused by a problem in care.

The report then compared their results with the overall number of deaths in acute hospitals in 2009, and calculated the number of preventable deaths per year in acute hospitals as 11, 859 (5.2 per cent of the total).

The Independent has therefore correctly quoted the findings of the report — 13 percent of patients had experienced a problem in care, with 5.2 percent dying from these problems, which when extrapolated to all deaths in acute care in England gives a figure of 11,859 preventable deaths per year.

While this certainly sounds like a large number, there are a few caveats that we need to bear in mind.

Firstly, we do need to be careful about extrapolating a reliable number of deaths from negligent care from a case study of 1,000 cases in 10 NHS hospitals. There are over 400 hospitals in England, and we don't know how representative the ones sampled are of the wider standard of care.

Secondly, we should consider what these "problems in care" were. The table below shows a breakdown of the problems in the 52 cases of a preventable death, indicating that the most common type of problem was clinical monitoring. This can include failure to act upon test results, or failure to increase the intensity of care when required.

 What we can see is that the majority of the problems in care were acts of omission, rather than an active mistake on the part of a clinician.

Most importantly, however, the report makes clear the majority of the 52 patients who died a preventable death were elderly or frail and were experiencing numerous complications. The average age of those who experienced a problem in their care was 76.7 years of age.

Overall, the average life expectancy of the 52 patients was only 2.1 more years, with over 60 percent of the patients already judged to have less than a year to live prior to any problems in their care occurring.

So while the Independent might be correct to point to "basic errors" as one of the things that led to a patients' death, a number of factors may have complicated or exacerbated any lapses.

Most previous estimates of the scale of harm and number of deaths caused by problems in care have been extracted from international studies and have typically reached much higher figures. For example, in 2000, the Chief Medical Officer of England estimated that 60,000 to 255,000 NHS patients suffered preventable disability or death.

While this would suggest an even bleaker picture of the standard of care in English hospitals, the 2009 Select Health Committee report, commenting on the methods used to measure patient safety in the NHS, claimed that extrapolation from international figures is extremely limited.

The Committee considered the methodology used by the London School of Hygiene and Tropical Medicine study as the most reliable, but it still highlights some limitations, some of which still apply to this report.  

Firstly, the Committee takes issue with the RCRR method itself, pointing out that professionals reflecting back on a case can lead to what is called "hindsight bias": the notion that it is difficult to judge what constitutes a clinical error after the event.

Secondly, and in a similar vein, the Committee claimed that "preventability is in the eye of reviewer": using individual professionals will necessarily involve some level of subjective opinion, and judgements could be contested.

So while the Independent correctly cites the data accumulated by a recent study on preventable deaths in NHS acute hospitals by the London School of Hygiene and Tropical Medicine, there is more to the story than the headline might suggest.

There are several noted limitations with the methodology used by this study. However, as the first study of it kind into patient harm and death by problems in care in England, it may nevertheless be the most reliable indicator of the scale and type of occurrences which lead to preventable deaths in acute NHS hospitals.

What was found is that the majority of the problems occurring involved omissions rather than active mistakes made on the part of a clinician. Additionally, the vast majority of patients who died a preventable death were old or had a very poor prognosis outside of any problems experienced in their care.