Mistakes happen. This is the reason why my Volvo has large rubber bumpers and why pencils have rubbers. Indeed, I, amongst many others, may have been born for this very reason. It has long been recognised that mistakes also happen when providing health care. Of course, there are lots of reasons that such mistakes occur, including sleep deprivation, being rushed, having illegible hand writing, delivering complex care, language barriers and treating elderly patients who have lots of interacting comorbidities.
Estimates vary, but the now famous Harvard Study concluded that 1% of hospital admissions result in an adverse event.1 I must admit to being sceptical about this figure, which would equate to more Americans being killed in US hospitals every 6 months than died in the entire Vietnam War. By contrast, on ward rounds it is clear to me that what actually kills people is being old, frail and diseased. And herein lies the root of my scepticism. The study takes no account of the expected risk of death in the absence of a medical error.
A rarely quoted study by Hayward and Hofer,2 tried to examine care more objectively. The authors derived three interesting and illuminating conclusions.
First, most deaths reportedly due to medical errors occurred at the end of life or in critically ill patients in whom death was likely regardless of the care received. Indeed, two-thirds of cases reviewed in their study had a "do-not-resuscitate" order in place at the time of death.
Second, the level of agreement that a death was 'preventable' was poor (kappa value 0.25). The authors concluded, "…most of the 'errors' identified represent outlier opinions in cases in which the median reviewer believed either that an error did not occur or that it had little or no effect on the outcome." They dryly commented that if there were a large enough reviewer panel, there would always be someone of the opinion that each death was preventable.
Third, the probability that an error definitely had contributed to a death was considered rare. The reviewing clinicians estimated that only 0.5% (95% CI, 0.3–0.7%) of patients who died would have lived an extra 3 months if their care had been optimal. This would represent roughly 1 patient per 10,000 admissions to the seven Veterans Affairs medical centres included in the study.
The above conclusions will not be surprising to doctors who look after patients. But surely we should always be alert and pounce on every medical mishap to continuously improve practices and reduce the risk of further errors? The hunt for hospital mishaps may, however, have unforeseen consequences. I am aware of three problematic issues.
The first issue is that identifying particular patterns of care that result in truly preventable deaths is difficult. It's far easier to identify 'minor' problems and the 'minor' individual responsible. For this reason nursing staff are the most frequently reprimanded team members. Paradoxically, nurses are the least likely not to internalise the reprimand and forgive themselves, and are the most likely to become depressed afterwards. I have seen nurses leave our NHS after a reprimand. There is already a national shortage of nurses because young people don’t want to spend a lifetime working shifts, doing physically and emotionally draining jobs in a punitive environment that comes down hard on every mistake. Qualified nurses don't need a push through the door; they are already leaving frontline services for less demanding jobs in the private sector. Personally, I regret not going into the financial sector where, for example, bringing your bank into financial ruin or the whole world into deep recession is rewarded with a warm handshake and a fat pension.
Trying to fix problems in complex settings using hindsight and anecdotes is the second issue as it may lead to processes that worsen care. British nurses now seem to spend more time completing care plans and paperwork than directly caring for patients. To reduce prescription errors, we now have a bewildering 6-page colour coded prescription chart with a myriad of tiny boxes to prescribe anything from oxygen, to compression stockings, to drugs, with ample opportunity to get confused and make a mistake. In my opinion, almost all processes that are put in place to reduce risk, result in an increase in the complexity and time it takes to achieve the task. Surely, the reverse should be the case? Never before in the history of medicine have so many patients with so much disease been given such complex care by so few nurses within such a short space of time!
The third and final issue is the overestimation of life expectancy. It is often lamented that doctors overestimate the life expectancy of their patients, but sick patients and their relatives have even more unrealistic expectations—they don't expect to die at all! Paradoxically, it is the elderly who have the least to gain from receiving numerous medical interventions, are the most likely to suffer an adverse event and the least likely to survive when something goes wrong.
In spite of all this, when the 'unexpected' happens, it is presumed that it must be due to an error. The creeping mistrust is fuelled by our medical obligation to disclose every medical mishap. The relatives of an octogenarian who succumbs to a hospital acquired infection after a hip replacement will have little doubt that, "The reason my grandfather died was because the nurse didn't dress his wound on time/he didn't get his tablets in the evening/the hospital gave him pneumonia."
How can we continuously reduce errors, encourage more young people to become healthcare professionals and provide sufficient time to complete every task, whilst encouraging the elderly to be realistic in their expectations? Unfortunately, this is a circle that cannot easily be squared. In the future, health care will undoubtedly be provided by robots!
- Brennan TA, Leape LL, Laird NM et al. Incidence of Adverse Events and Negligence in Hospitalized Patients — Results of the Harvard Medical Practice Study I. NEJM 1991; 324: 370–376.
- Hayward RA and Hofer TP. Estimating hospital deaths due to medical errors: preventability is in the eye of the reviewer. JAMA 2001; 286: 415–420.