The situation has now become so bad that America's Medicaid health insurance system will, from next October, refuse to pay hospitals for eight common errors that regularly occur and which they believe are avoidable. The insurers' umbrella group, the Centers for Medicare & Medicaid Services (CMS), may eventually add another nine situations to the list, which would include the hospital-acquired killer infection caused by Clostridium difficile and deep vein thrombosis (JAMA, 2008; 299: 2495-6).
The CMS' Acting Administrator Kerry Weems commented: "Medi-care can and should take the lead in encouraging hospitals to improve the safety and quality of care and make better practices a routine part of the care they provide not just to people with Medicare, but to every patient they treat."
This extraordinary step, which may jeopardize the survival of some hospitals across the US, has been a long time coming. The first rum-blings of discontent were made by the Institute of Medicine in 1999 with its influential report, To Err
Is Human: Building A Safer Health System, in which it pinpointed the extent of the problem and the actual numbers of patients being harmed.
The CMS announcement came within days of a similar sounding report from the US Government Accountability Office (GAO), which criticized the country's health bureaucrats for failing to improve safety standards in hospitals. This was the response to the GAO's discovery that 87 per cent of hospitals have failed to implement all of the recommended safety measures that would minimize the occurrence of medical errors.
Not many dead
The true extent of medical errors may never be known, and it's accepted that the official figures are a low approximation. One study, for example, estimated that medic-al errors kill more like 180,000 Americans every year and went on to describe the rate as "equivalent of three jumbo-jet crashes every two days" (JAMA, 1994; 272: 1851-7).
But the American health insur-ance group, Healthgrades Inc., believes that the number of deaths due to errors in US hospitals is closer to 195,000 each year. These figures are based on a survey of 45 per cent of all hospital admissions between 2000 and 2002, and includes failures to rescue dying patients and deaths from infection in low-risk patients, two categories which have been included in the official figures. However, in addition to deaths, 1.14 million patients every year also suffer what is referred to as a 'safety incident', which represents one in four Medicaid patients admitted to hospital (Reuters, 27 July 2004).
In the UK, around 850,000 errors occur in hospital every year, and 40,000 patients die as a result. But here, as in the US, these are conservative estimates, and are based on an error rate of 2.2 per cent of all admissions. A more accurate rate, based on a meta-analysis of several studies, is around 4.75 per cent which, if correct, means that the true figures should be twice the numbers being officially reported (BMJ, 2004; 329: 369).
Not surprisingly, a Canadian government commission has concluded that hospitals are more dangerous than mines and factories. The commission was created after 44 patients died in Canadian hospitals because of the SARS (severe acute respiratory syndrome) infection (Lancet, 2007; 369: 264).
The items on the new 'no-pay' list from America's health insurers concentrate on surgical errors or their consequences. In general, these are related to:
- air embolism
- blood incompatibility fromtransfusions
- catheter-associated urinary tract infection
- decubitus ulcer (pressure sores)
- vascular catheter-associated infection
- surgical site infection
- mediastinitis (infection of the chest area after surgery)
- falls and trauma
- objects left in the body after surgery.
It has been estimated that errors are responsible for $3 million of additional costs every year in a typical 100-bed to 300-bed hospital in the US, which is accustomed to paying for prolonged stays and additional care. As there are 2765 hospitals across America that fall into that category, excess costs due to medical errors amount to $8.29 billion every year, a sum that is higher than the annual defence expenditure of Taiwan.
Overall, Medicaid reckons that its total bill for all errors and their consequences amounts to a stag-gering $40 billion every year.
Air or gas embolism-where gas bubbles get into the bloodstream-is the only concern on the 'no-pay' list to have an improved error rate, probably because of improvements in monitoring and/or surgical techniques.
In one study covering the years 1980 to 1995, death due to pul-monary embolism decreased by 50 per cent in the Minneapolis-St Paul, MN, metropolitan area (J Clin Epidemiol, 2000; 53: 103-9). In addition, a separate study looking at trends across the US found that deaths due to pulmonary embolism had fallen since 1998, reversing the trend seen between 1979 and 1989, when mortality rates were on the increase (Am J Cardiol, 2004; 93: 1197-9).
Giving the patient the wrong kind of blood during a transfusion-known as 'ABO incompatibility'-is a major concern, and one that can result in the death of the patient.
In one study involving 777 hospitals across Japan, 115 of them, or 20 per cent, reported ABO errors between 1995 and 1999. Most of the errors-specifically, 42.8 per cent of them-were caused by wrongly identifying blood bags, while the incorrect identification of the patient's blood type was responsible for 15 per cent of the errors, and giving the transfusion to the wrong patient accounted for 11 per cent of errors. The majority of these mistakes happened during holiday periods, during the night shift or in an emergency situation (Rinsho Byori, 2003; 51: 43-9).
The risk of blood transfusion error is reported to be "significant" in American hospitals, according to a study that assessed incidents in New York State over a 22-month period. Of the 1,784,600 transfu-sions that were monitored at this time, 92 cases were made in error and 54 involved the wrong blood type, three of which proved fatal. Allowing for the strong likelihood of underreporting, the researchers concluded that errors occur in one per 12,000 transfusions (Transfusion, 1992; 32: 601-6).
However, another team of scientists-this time from Ikoma, Japan-believes that the rate is far higher, possibly up to 253 errors per 100,000 transfusions. In this case, these figures also included giving the wrong blood type to the patient as well as giving the blood to the wrong patient (Rinsho Byori, 2003; 51: 146-9).
The most likely error is giving the wrong type of blood to the patient, and this happens in one out of every 2262 transfusions, researchers from the Cleveland Clinic in Ohio estimate. In a review of transfusion errors, 57 per cent were related to the wrong blood type (Am J Clin Pathol, 2006; 126: 422-6).
In the UK, 46 people died as a direct result of a blood trans-fusion between 1996 and 2003, according to figures gathered by the SHOT (Serious Hazards of Transfusion) survey. In addition, 253 patients suffered a serious reaction, including acute lung injury and infection (BMJ, 2005; 330: 104-5).
Infection in hospital-and especially following surgery-is relatively common, with around 7.6 per cent of hospital patients developing some sort of infection during their stay.
In fact, around two million people each year fall victim to an infection while being a patient in an American hospital, and around 100,000 of them die as a direct result (JAMA, 2008; 299: 2495-6).
According to The Third Preva-lence Survey of Healthcare Asso-ciated Infections in Acute Hospitals, which was jointly sponsored by the Hospital Infection Society (HIS) and the Infection Control Nurses Association (ICNA), and based on data collected from a total of 75,694 patients from February to May 2006 in the UK and Republic of Ireland, the most common infec-tions involved the gastrointestinal tract. This accounted for 20 per cent of all infections, followed by infections of the urinary tract, which affected 19.9 per cent patients.
Surprisingly, considering the many headlines they've attracted, the hospital-acquired 'superbug' infections-MRSA (methicillin-resistant Staphylococcus aureus) and C. difficile-together were responsible for only 2.3 per cent of all infections (J Hosp Infect, 10 June 2008; epub ahead of print).
Catheter infection is another common problem. In a review of 147 hospital patients who were fitted with a catheter, 15 per cent developed an infection. But this rate increased dramatically to 46 per cent if the presence of inflam-mation around the catheter site was included. The rate also rose the longer the catheter was kept in place (Crit Care Med, 1996; 24: 1660-5).
Wound infection of the sternum following coronary artery bypass grafting is not often reported and, while it is a relatively rare occur-rence, it can also be deadly. One study that tracked 9201 patients following bypass surgery revealed that 65 of them developed sternal wound infection, and seven died while still in hospital (BMC Infect Dis, 2007; 7: 112).
Pressure ulcers have become so commonplace that they have come to be considered a likely conse-quence of any lengthy hospital stay, especially among the elderly. In one German study of 6473 hospital and nursing-home patients, 61 per cent of those in nursing homes had pressure ulcers compared with 38 per cent of those in hospital. In the majority of cases, the causes of these infections were found to be within those facilities (Pflege Z, 2008; 61: 90-3).
A different pattern emerged in a study of 695 patients who were either in a university hospital, a general hospital or a nursing home. The prevalence of pressure ulcers was highest, at 23.9 per cent, in the university hospital, followed by 20 per cent of those in the nursing home in comparison to just 13 per cent of patients in the general hospital. Most of the ulcers were assessed as grade I, which suggests that they were especially painful. However, despite this fact, only half these patients had the use of a pressure-reducing mattress (J Wound Care, 2004; 13: 286-90).
Mediastinitis, or chest infection, is another rare, but serious, compli-cation following heart surgery. In one study that tracked 1700 patients after heart surgery, 45 of them-or 2.65 per cent-developed mediastinitis. Those who were most likely to suffer from such infection were either obese or had chronic obstructive pulmonary disease (J Cardiothorac Surg, 2007; 2: 23).
A slightly higher rate of infection was seen in a study of 331,429 patients who underwent coronary artery bypass surgery. Although 11,636 patients, representing 3.5 per cent of the total, suffered a major infection after surgery, medi-astinitis was responsible for only a quarter of these cases. About one-third developed septicaemia. Over-all, around 17 per cent of patients died as a result of their infections (Circulation, 2005; 112 [9 Suppl]: 1358-65).
Objects left in the body
Stories of objects being left in the body after surgery are the stuff of urban legends. Statistics prepared for insurance companies in the US reckon that one unlucky patient out of 1500 undergoing surgery will end up with some material or implement used in the procedure still inside him afterwards. Fabric items such as swabs are left inside the patient in 70 per cent of cases, and the remainder are metal objects used during surgery. Metal objects cause a quick, acute react-ion prompting immediate surgery, whereas fabric items can cause a slow, chronic problem that can last for years. Tragically, operations to remove the object are themselves considered high-risk procedures, as up to 35 per cent of patients die afterwards (Chirurg, 2007; 78: 7-12).
Another study that looked at 30 cases discovered that sponges were left in 16 of these patients, while surgical instruments were left in the rest. Most of these objects were left in the chest cavity. Of the 25 patients who underwent operations to remove the objects, all survived the procedure. Researchers also found that patients were likely to have more than one object left behind, either because they had undergone multiple surgical proce-dures all at the same time or had several surgical teams working on them (J Surg Res, 2007; 138: 170-4).
Falls and trauma
Falls in hospitals and care homes are common, especially among the elderly, but they're not immediately associated with medical errors. Yet, studies that have looked into the problem have found that they are often the consequence of surgery combined with a failure to provide patients' aids to prevent falls.
An Austrian study monitored the progress of 935 patients who had undergone surgery to repair a fractured femur after an initial fall, and were able to identify those who were most likely to fall again. Of these, 11.8 per cent fell again, and most of these falls occurred in the second week following surgery, when the patient was becoming mobile again. The researchers con-cluded that it was possible to build a risk profile for predicting future falls (Z Gerontol Geriatr, 2003; 36: 16-22).
A study carried out by the University of Hamburg, in Germany, found that falls were more common in people who were staying in hospital for an extended period of time-in this case, 28 days. This meant that they were also under-going more complicated treatment (Pflege, 2005; 18: 39-42).
A similar picture was painted in a separate study carried out at the same university. In this case, they found that 'fallers' were more likely to be those who stayed in hospital for more than 27 days (Z Arztl Fortbild Qualitatssich, 2007; 101: 617-22).
The buck stops
The coming 'pay-out strike' by America's health insurers is the result of years of paying out for errors that they maintain are avoidable. Hospital errors are so rife, the CMS believes, that they add $15,000 to every patient's bill, resulting in an annual payout of $40 billion for avoidable costs. In the current economic climate, it seems that hospitals and doctors are virtually being rewarded for making mistakes, says the CMS.
Unfortunately, this unprecedent-ed action will probably not reduce the number of medical errors, nor will it encourage doctors and/or hospitals to improve their standard protocols. Instead, it seems it will merely force hospitals to close. This means that, either way, it's the patients who will ultimately have to pay the price.
Because of the fear of ensuing litigation, one in five doctors has admitted that he wouldn't own up to an error if it resulted in a patient's death. In fact, according to a mailed-out survey of medical and surgical physicians in the US and Canada, only 40 per cent claimed they would openly admit that an error had occurred, while the remainder would either mask the error by describing it as an 'adverse event' or, in 20 per cent of cases, not mention any error at all. This suggests that many deaths that occur in hospital are never classified as being due to a mistake within the hospital system (Arch Intern Med, 2006; 166: 1585-93).
Yet, it appears that a third of all hospital doctors make at least one major mistake every year that can harm, or even kill, the patient. As a result, the physician may become so upset by the blunder that his chances of committing another mistake increase dramatically, researchers found.
In all, 34 per cent of doctors admitted that they made at least one 'major' medical error in a year of working at the Mayo Clinic in Rochester, MN. However, researchers fear that the true picture could be even worse, as many doctors refused to participate in the survey, and even those who did may not have admitted to all of their mistakes (JAMA, 2006; 296: 1071-8).
Most mistakes aren't reported because the so-called 'adverse event' takes place in the patient's home, where he is not being monitored. French researchers have discovered that only one out of every 24,433 adverse reactions is ever reported, often because either the doctor or patient doesn't realize that a drug is to blame or because the doctor doesn't have the time to prepare a full report (Br J Clin Pharmacol, 1997; 43: 177-81). According to another study, around 6 per cent of all adverse drug reactions are the result of a prescribing or other medical mistake (Swiss Med Wkly, 2004; 134: 664-70).
Most errors are made by junior doctors. In a review of 240 incidents involving junior doctors and trainees, the vast majority was the result of an error of judgement and a lack of technical competence (Arch Intern Med, 2007; 167: 2030-6).
Doctors also blame the high error rate on their punishing schedules, which often result in working for 24 hours without sleep. Among medical interns, it was found that, due to sleep-related fatigue, they were three times more likely to make medical errors, some of which resulted in the patient's death (PLoS Med, 2006; 3: e487).
In addition, the risk of a car crash after a long shift increased by 168 per cent and a road crash near-miss by 460 per cent. Interns have admitted to falling asleep during lectures, during hospital rounds and even during surgery (National Public Radio, 13 December 2006).
Most medical mistakes are related to the administration of drugs, either because of a wrong dosage or because it is an inappropriate drug for the condition. In one study, researchers discovered that two-thirds of drugs prescribed in a hospital were 'inappropriate' (J Am Med Inform Assoc, 2008; 24 April, epub ahead of print).
In all, researchers from the University of London have identified 88 reasons why a prescription can go wrong. Each of the errors is down to mistakes or misjudgements by hospital staff because of fatigue, or complete ignorance of the drug, and its correct usage and dosage.
The researchers spent several months in a hospital pharmacy unit, and counted 88 separate errors. In interviews afterwards, the doctors involved in these mistakes admitted to a lapse in concentration, usually because they either had too much work to do or had been interrupted when writing the prescription (Lancet, 2002; 359: 1373-8).
A study from Australia discovered that 4 per cent of all hospital admissions, and up to 30 per cent of admissions of elderly patients, were the direct result of a medication error. In fact, it was found that 26 per cent of all prescribing errors in hospital involve mistakes related to medication. Overall, errors occur in 20 per cent of hospital ward prescriptions (Int J Qual Health Care, 2003; 15 Suppl 1: i49-59).
Another study has presented an equally disturbing picture and estimated that, for every 1000 prescriptions written in hospital, 111.4 were wrong. Most of them involved a dosing or frequency error, where the drug was prescribed for too many or too few occasions per day, but 52 errors were caused by the doctor's poor handwriting [Proc (Bayl Univ Med Cent), 2004; 17: 357-61].
The intensive care unit (ICU) is where many drug prescribing errors occur. One study discovered that errors were twice as likely in an ICU compared with a general ward although, when adjusted to accomodate the greater number of drugs used in an ICU, the difference disappeared (Crit Care Med, 1997; 25: 1289-97). Nevertheless, according to another study, there was an error in one out of every five prescriptions made in the ICU (Crit Care Med, 2006; 34: 415-25).
Aside from patients in critical care, elderly patients are also regular victims of prescribing errors. But, again, this is perhaps not surprising as they are also taking more drugs than any other sector of the general population. Indeed, older patients are often given multiple prescriptions, a popular practice referred toas 'polypharmacy'.
In one study of 389 geriatric patients, 107 were given prescriptions for 116 drugs that were totally inappropriate for use by the elderly, according to an expert consensus list known as the 'Beers Criteria'. However, although the researchers recorded a total
of 131 adverse drug reactions, only 9 per cent were because of a drug named in the Beers List (Am J Geriatr Pharmacother, 2006; 4: 297-305).