Medical errors: a common problemBMJ 2001; 322 doi: https://doi.org/10.1136/bmj.322.7285.501 (Published 03 March 2001) Cite this as: BMJ 2001;322:501
It is time to get serious about them
Medical errors continue to dominate newspaper headlines. There is rarely an informed comment on likelihood or cause, rather a tacit assumption that they should never happen—and an implicit conclusion that they are getting more common. What is the truth? Firstly, errors have always happened. Secondly, there has been no clear indication as to how common they are in the United Kingdom—though a pilot study in this week's issue represents a first attempt to quantify the size of the problem (p 517).1 Alongside this is the difficulty of indicating risk. To a bereaved relative the knowledge that there was a 1 in 1000 risk is no consolation—for them it was 1 in 1. In a country where millions are spent every week on the national lottery the concept of risk is obviously alien. What is clear, however, is both that we need to know more about errors and to do more about them.
How common are errors? Can they be minimised? And how should we tackle risk management? One problem in assessing the frequency of errors is that we are deeply immersed in a blame culture, so it is hard to persuade people to report them. Many errors do not cause harm, but in many ways these are as important as those that do. They indicate a breakdown in the system or a wrong decision. If we are to learn from mistakes then we need to know about as many as possible so that corrective action can be taken. This requires a cultural change and sensitive handling of the individual making the report. A recent report from Chesterfield has shown a 150% increase in error reporting by threats of disciplinary action—apparently effective but perhaps not the best approach.2
Few reliable studies of adverse events exist. Two seminal studies were reported some years ago from the United States 3 4 and Australia5 showing adverse event rates of 3.7% and 16.6% of admissions respectively, with intermediate rates in Colorado and Utah. 6 7 In the Colorado study rates were higher in the elderly.8 Problems arise because of definitions; and retrospective analysis can be subjective. What appeared to be clinically reasonable at the time may be second guessed if an adverse event occurs. Nevertheless, a figure of 5-10% is worrying, particularly since a half or more of these events were deemed preventable.8 Similar rates were found for interpreting emergency radiographs.9
Finally, we now have some British data from London based on retrospective record review. In their study of over 1000 records in two acute hospitals, Vincent et al found that almost 11% of patients experienced an adverse event, over half of which were deemed preventable judged by ordinary standards of care.1 More worryingly, at least a third of these events led to disability or death. This was a pilot study but there is no reason to believe that the results are unrepresentative. The frightening extrapolation of these data suggests that in England and Wales adverse events lead to an extra 3 million bed days at a minimum cost of £1bn per year. Only a full scale study can substantiate this estimate, and if the NHS is serious about learning about and reducing errors it should fund such a study.
What can be done about these errors? They cannot be ignored. Once errors are recognised their causes must be analysed so that preventive measures can be applied. Some of the mistakes are caused by systems failures—this has been shown, for example, with drug errors or wrong transfusions. Clear definition of clinical responsibilities is needed. Fatigue may also cause problems, as does the use of inappropriately junior staff. The main causes of adverse events relate to operative errors, drugs, medical procedures, and diagnosis. Each of these is amenable to prevention. Better surgical training is obvious. This has been taken on board by the Royal College of Surgeons, though concerns remain that, because of shorter training and tighter working hours, young surgeons are less experienced than previously. Better training programmes will also help with medical procedures. Fewer operations and procedures during the night may also help. Drug errors remain a problem—no one can remember all the possible drug interactions that may occur, and incorrect dosages are also a recurrent problem. A computer linked pharmacology system, such as that described from Birmingham,10 seems an ideal preventive and learning tool. This system sends warnings when incompatible or otherwise dangerous drugs are prescribed, and the introduction of such a system nationwide could prevent hundreds, indeed thousands, of errors. Errors in diagnosis could be minimised by better training and wider use of protocols and diagnostic algorithms.
Errors are problems that will not go away. A pilot study by the Royal College of Physicians into deaths after admission for medical emergencies suggests that some error occurred in as many as one in five cases, although not necessarily leading to an adverse event (unpublished). These data should be interpreted cautiously but do suggest that actual recorded adverse events are the tip of the iceberg. Analogies are often drawn with airline pilots. These are overinterpreted in that an aeroplane should behave predictably on all occasions, whereas every patient is different and the same disease can present in myriad ways. Nevertheless, we can learn from the airlines, as David Johnson suggests on p 563.11 They spend a much higher proportion of revenue on training and they report all incidents, with “blame” being minimised. This is a habit which we should adopt, but it requires a much more sympathetic approach from management than has pertained in the past.
Even more important, we need, as suggested by Vincent et al1 and England's chief medical officer12 to put in place a national system for recording adverse events. This is an enormous undertaking and could be introduced initially in high risk areas—but in the end it should be a matter of course in every medical setting, public and private, in the United Kingdom. Only then will we really learn and improve our practice to the ultimate benefit of the public.