Poor hospital care is “needlessly killing 1,000 NHS patients a month”, The Daily Telegraph headline reads. It says that the largest ever study of errors in British hospitals has found that one patient in 10 is affected by potentially serious medical errors, with half of them dying as a result.
This study was a review of the patient records of 1,000 adults who died in 10 hospitals across England in 2009. Medical reviewers who examined the records considered that one death in 20 had a greater than 50% chance of being preventable.
The researchers defined a “preventable death” as:
A recent example of a preventable death, highlighted by the Telegraph, was the tragic death of Kane Gorny. His inquest found that he died from dehydration due to a combination of misdiagnosis and inappropriate treatment.
Most of these “possibly preventable” deaths occurred among elderly, frail patients with multiple other medical problems. This raises debate over whether these deaths were actually “preventable”.
Based on these figures the reviewers estimated that almost 12,000 (11,859) adult deaths would have been preventable in England during 2009. These are important findings, but are estimates only – the reviewers only examined the records of 1,000 patients from a sample of hospitals.
While a single preventable death is one too many, the researchers actually found that the number of preventable deaths was far lower than previously thought. Some previous estimates put the number of preventable deaths occurring every year in England as high as 40,000. The researchers were keen to stress, “this does not mean that preventable deaths should be ignored and no attempt made to improve our understanding of their causes”.
This study was conducted by researchers from the London School of Hygiene and Tropical Medicine, the National Patient Safety Agency, Imperial College London and the University of Newcastle. The study was funded by the National Institute of Health Research, Research for Patient Benefit Programme. It was published in the peer-reviewed British Medical Journal Quality & Safety.
The media coverage is generally representative of this research, but the Independent’s headline states that doctors are to blame for the deaths. While doctor-related factors such as misdiagnosis or treatment errors were considered to have contributed to some of the deaths, the study has not reported the specific errors, or implied any responsibility for the errors.
This was a review of the medical records of adults who died in hospitals in England during 2009. The researchers say that previous national and international studies have given wide estimates of the number of preventable deaths that occur in hospital, with estimates for England ranging from 840 to 40,000 deaths a year. However, these studies have not assessed whether adverse events could have contributed to death. This is what the current review aimed to assess.
In the current study, trained medical reviewers examined the records and identified issues with care that could have contributed to the deaths. The researchers say that a retrospective review of medical records is the most sensitive approach in determining the proportion of hospital deaths that are preventable. They based their study design on previous similar reviews that have been performed in the UK, the Netherlands and the US.
As only a random sample of hospitals and patient records was examined, the number given of annual preventable deaths across England is only an estimate. Also, though the researchers ensured all medical reviewers were fully trained, and checked their assessments, the review will unavoidably contain an element of subjective judgement.
The researchers identified deceased patients from 10 randomly selected English acute hospital trusts. The random sampling had been stratified to ensure it contained:
They chose to sample 1,000 patient deaths based on their estimate of the number of deaths they expected to be preventable (6%). From each of the 10 hospitals the medical records of 100 patients who had died in hospital during 2009 were randomly selected using the hospital administration system in each trust.
The researchers wanted to focus on general medical and surgical admissions, so excluded paediatric, obstetric and psychiatric hospital admissions.
Judgement of preventable deaths was carried out in two stages. First, the reviewers were asked to judge whether there had been any problem in care that had contributed to the patient’s death. Such care problems were defined as:
Second, for each case where a problem in care was identified, the reviewers then judged whether the death could have been prevented.
This two-stage process was used because some care issues that contributed to death may not necessarily have been the result of poor practice. For example, if a patient with a heart attack was appropriately given an anti-clotting drug, but giving that drug then caused them to die from a brain bleed, the death would not be considered preventable. Reviewers judged preventability on a six-point scale ranging from one (definitely not preventable) to six (definitely preventable). Deaths were judged to be preventable if reviewers gave a score of four, five or six on the scale. That is, there was a more than 50% chance that the death was preventable.
The reviewers were general medical doctors recruited through the Royal College of Physicians, who received training in the review process. To validate their assessments a sample of 25% of the notes was examined by another reviewer, and each case that was considered to be a preventable death was discussed with the principal investigator and an expert reviewer.
In the first stage of review, 131 patients were identified as having experienced a problem in care that contributed to their death. In the second stage of review 52 of these deaths (5.2% of the total sample reviewed) were judged to be preventable (95% confidence interval 3.8% to 6.6%). This was 39.7% of the 131 cases identified to have had a problem in care contributing to death. These 52 deaths had received a score of four to six suggesting there was a greater than 50% chance the death was preventable. There were no significant differences between the proportions of preventable deaths found at each of the 10 hospitals.
Patients with preventable deaths were more likely to have been admitted under surgical specialties, and most of the problems occurred during ward care. In 73% of preventable deaths more than one problem in care was identified. The most frequent problems were related to:
Most preventable deaths (60%) occurred in elderly, frail patients with multiple other medical problems who were judged to have had less than one year of life left to live.
The researchers considered that if 5.2% of deaths in hospital are preventable, there would be 11,859 preventable adult deaths in English NHS hospitals each year (based on 228,065 adult deaths in hospitals in England in 2009).
The researchers conclude that the incidence of preventable hospital deaths in England is lower than previous estimates, though the burden of harm from preventable problems in care is still substantial. They say that “a focus on deaths may not be the most efficient approach to identify opportunities for improvement given the low proportion of deaths due to problems with healthcare”.
This was a well-conducted study that has important findings. Around 5% of the 1,000 patient deaths examined were judged to have been preventable due to issues with healthcare. The reviewers used this figure to calculate that almost 12,000 deaths a year are preventable – the 1,000 a month figure quoted in the media.
The main thing to be aware of is that these are estimates based on a relatively small sample only. The researchers reviewed only 1,000 deaths from 10 English hospitals. However, the researchers did make careful attempts to ensure that their selection was a representative sample of hospitals from across England.
The researchers also ensured that the medical reviewers were fully trained in the review process, and they also validated their assessments by performing a second review of a sample of 25% of the notes. Additionally, each case that was considered to be a preventable death was discussed with the principal investigator and an expert reviewer. Despite this, there will still have been some subjective analysis, and a different set of reviewers may have come up with different figures.
Related to this is the use of the six-point scale. Scores of four to six were considered to be preventable deaths, though the researchers say that using a more stringent definition of preventable (scores of five and six only) gave an estimate of 2.3%, rather than 5.2%. Though, similarly, using a more relaxed definition of preventable (scores of three to six) would raise the proportion of possibly preventable deaths to 8.5%.
Despite the alarmist media headlines, the researchers conclude that the number of preventable hospital deaths is actually much lower than previous estimates. They consider that “given the low proportion of deaths due to problems with healthcare”, focusing on patient deaths may not be the best way of finding ways to improve healthcare. This seems to be a sensible conclusion.
Of interest is the fact that the independent National Confidential Enquiry into Patient Outcomes and Death (NCEPOD), an independent charitable organisation, commissioned by the Healthcare Quality Improvement Pathway (HQIP), performs regular reviews of medical and surgical practice in UK hospitals and makes recommendations for improving the quality of healthcare. It does this through extensive confidential surveys and research covering many different aspects of care, including the review of medical and surgical records of patients who have died, and interviews of the treating consultants. NCEPOD produce several reports a year focusing on specific aspects of healthcare. These typically involve the review of several thousand records. It would be well worth comparing NCEPOD’s findings with those of this study.