Safety Sciences

The Tragedy of Radiation Therapy Errors

Radiation Offers New Cures, and Ways to Do Harm

Walter Bogdanich, New York Times Jan 23, 2010

Radiation is a common therapy for cancer and is delivered using a linear accelerator, a machine that delivers beams of high-energy radiation. It is an effective treatment for many types of cancer, but when errors occur in its administration, the results can be disastrous to the patient. The New York Times just published an article following an investigation into radiotherapy errors. The paper “examined thousands of pages of public and private records and interviewed physicians, medical physicists, researchers and government regulators.”

The article describes in graphic and heart-breaking detail the stories of two patients who suffered from inadvertent overdoses of radiation therapy. It also reports that more than 90 prostate cancer patients received the wrong radiation dose at one Philadelphia hospital, and that at a Florida hospital 77 brain cancer patients received a 50 percent higher dose than they were supposed to receive. The article cites some statistics from 621 incidents (from 2001 – 2008) of radiotherapy errors in New York state. In 133 of these incidents, devices used to shape or modulate radiation beams were left out, wrongly positioned or otherwise misused. In 284 incidents, radiation missed all or part of its intended target, or was directed to the wrong body part. In one case, radioactive seeds were mistakenly implanted in the base of the penis instead of the prostate. One patient was treated for prostate cancer when he actually had stomach cancer. Fifty patients received radiation intended for someone else.

Why did these errors occur? Four types of causal factors are described in the article: (1) Problems with the computers that control the linear accelerator – flaws in the software or programming; (2) Poor safety procedures; (3) Inadequate staffing and training; (4) Failure to follow a good quality assurance plan. The tremendous increase in the complexity of linear accelerators and treatment planning over the past 20 years has apparently increased the number of opportunities for error. Hospitals and health professionals are too trusting of the computer systems that run the accelerators. In the Florida case, the linear accelerators had been programmed incorrectly for nearly a year before the error was discovered.

Human performance too was at fault. For example, in one case, the therapist mistakenly programmed the computer to not use a ‘wedge’ (a device to make the radiation beam narrower so that it strikes only a small area of the body) when the wedge should have been used. This error was caught neither by the medical physicists who checked the treatment records weekly, nor by other therapists (on 27 occasions!), even though the computer screen clearly showed that the wedge was missing, and even though the state health department had recently sent hospitals a notice that therapists should closely monitor their computer screens.  

In another case, the computer crashed multiple times as the radiation therapy dose was being programmed into the linear accelerator. As a result, the multi-leaf collimator (a device that narrows the radiation beam to focus it solely on the tumor) was wide open, and the patient was exposed to a wide swath of radiation that targeted his entire neck instead of just the base of his tongue. This was not caught by the medical physicist until it was too late. Also, a test to check that the computer was programmed correctly was not run before exposing the patient to the first dose of radiation. Finally, even though the computer screen was displaying the collimator as being wide open, this was not noted by the therapists. As a result, the patient received three excessive doses of radiotherapy that ultimately led to his death.

Other points the article makes are: (1) the exact frequency of medical errors with radiation therapy is unknown; (2) the exact degree of harm caused by radiation overdose is hard to determine because often the adverse effects occur years later; and (3) inadvertent underdosing of radiation therapy is hard to detect, because there is no apparent harm. Because of the last two points, lawsuits related to radiation therapy errors are apparently rare.

This article is a sobering reminder of the power of radiation therapy to hurt patients as well as to heal. It also reminds us how computers can cause serious errors and how blind faith in computers can lead to serious errors being overlooked.  

What's in a Name........?

Categorizing Errors and Adverse Events for Learning: A Provider Perspective. Ginsburg L et al. Healthcare Quarterly 2009; 12: 154 - 160

'Near miss,' 'medical error,' adverse event,' 'harm,' 'preventable adverse event,' 'variance,' 'sentinel event' - a bewildering number of terms like these is used in patient safety to indicate different types of events. (An umbrella term for all these is 'patient safety events.') Often there is confusion among health professionals about what to call an event or an incident and much misunderstanding may result from this. 

To understand what front-line providers and managers thought about such terms, and to identify which types of events they thought were most suitable for learning in order to improve patient safety, the authors of this paper conducted focus groups with patient safety officers, patient care managers, nurses and other allied health professionals. 

The authors presented the following categorization of events to the participants: (1) near misses, (2) preventable adverse events causing minimal harm but no prolonged hospitalization, (3) preventable adverse events that resulted in prolonged hospitalization but no permanent harm to the patient and (4) sentinel events (events that resulted in permanent disability or death).  The focus group participants:

1. preferred a simple classification of events into 'major events' and 'minor events,' with a further sub-classification of the events by the degree of harm that was actually observed or could have occurred. They did not like the above categorization - it was not intuitive, was too fine-grained, and too restrictive. 
2. believed that preventable events should be the focus of learning opportunities, as opposed to non-preventable events (such as an allergic reaction to a medication that occurs for the first time).
3. thought that near-misses were an important way of learning how to improve patient safety.
4. disagreed on which type of events were more valuable for leaning - events actually causing severe harm, or those that could have caused severe harm (but did not). 

Other findings from the focus groups were:

• Attribution of harm to an event is sometimes difficult. For example, it may be difficult to determine whether an event truly caused prolonged hospitalization or whether the disability it caused is permanent.
• Preventability might be hard to determine in certain cases. 
• The word 'preventable' is bothersome because it implies that the health professional is at fault, and not other aspects of the system.
• Near-misses are often unreported or unnoticed, and are complex to identify and categorize.
• Near-misses should be categorized based on the severity of the harm that could have resulted, and how close to the patient they were caught.
• When an even causes serious patient harm, it can lead to significant and long-lasting learning, but can be hard to discuss. In contrast, near-misses that could have potentially caused harm are easier to discuss, but may not be taken as seriously as actual catastrophic events.

Based these results, the authors developed and refined a new typology of patient safety events, to be used for all patient safety events except those that are clearly non-preventable: 

Minor Event: An event involving no harm or very minimal temporary harm to the patient 

Moderate Event: An event that causes discomfort sufficient to interfere with usual activity and requires additional specific therapeutic intervention, but poses no significant or permanent risk of harm to the patient. 

Major Event: An event involving death or serious physical/ psychological injury.  

Minor Near-miss: An event that would have resulted in no harm or very minimal temporary harm to the patient but did not because it was caught or because of good luck. 

Major Near-miss: An event that would have resulted in death or serious physical or psychological injury but did not because it was caught or because of good luck. 

What's the Most Dangerous Month in Hospitals?

Rate of undesirable events at beginning of academic year: retrospective cohort study

Haller et al BMJ 2009; 339: b3974

"Don't get admitted to the hospital in July" is advice that is commonly passed around, based on the belief that the quality of patient care is poorer in July, when the new batch of residents typically starts in the US. This transition period when trainees begin their training has been called 'the July phenomenon' in the US and the 'August killing season' in the UK based on the belief that this is the most dangerous period of the year to be admitted to the hospital. 

This belief persists, even though numerous studies that have not been able to prove that patient outcomes are worse in the months immediately after an influx of new trainees arrives. In a recent study however, Haller and colleagues were able to show that there is indeed some truth to the belief that patient care by new trainees leads to more adverse events. 

In this study, the authors analyzed data on adverse events related to anesthetic procedures. In their hospital (Alfred Hospital in Melbourne, Australia), a special form is completed for every anesthetic procedure - this form asks the trainee to report, for every single procedure, whether or not each of 44 predefined incidents occurred. Examples of these events, which the authors call 'undesirable events,' are central and peripheral nerve injury, inadequate oxygenation, vomiting/aspiration in the OR, technical failure of endotracheal intubation, accidental upper airway obstruction, respiratory arrest during local-regional anesthesia, technical failure of arterial line insertion, uncontrolled hypertension, and uncontrolled hypotension.  

The authors analyzed the accumulated data from such reporting, on procedures performed by trainees between October 1995 - December 2000. In Australia, the new trainees usually start on February 1.

After adjusting for case mix using a clustered multivariate analysis, the authors found that the number of adverse incidents per 1000 patient hours of procedures was higher in the first month of academic training than in the rest of the year (137 vs 107 per 1000 patient hours, a 40% higher adjusted rate, p < 0.001). 

Surprisingly, they also found that trainees at all levels of training, even those in their fifth year of training had higher rates of undesirable events in the first month compared to the rest of the year. (Note: all these trainees, even those at higher levels, were working in this hospital for the first time. In Australia it is common for trainees work in multiple hospitals through their years of training, and not stay in one program throughout). 

The authors' speculate that their results are different from those of other previous studies on this topic because they collected detailed data on each patient (instead of analyzing administrative databases like some other studies), had a larger sample size than other studies, and studied anesthesia care, which is inherently a riskier specialty than other medical specialties. 

This study clearly shows the existence of a learning curve for trainees (at least in anesthesia) and also emphasizes that trainees at more advanced levels are still at higher risk of committing errors in the first month, perhaps because of unfamiliarity with the work environment (remember that even the advanced trainees were new to this hospital), excessive cross coverage, poor handovers, and insufficient supervision. 

The authors suggest that adverse events in the early part of the academic year can be prevented by improving the orientation of trainees, using simulation, not assigning clinical tasks to trainees right from day one of training, systematic use of well documented standard working practices, teamwork training, and closer supervision. 

Are Physicians like Cobblers with Bad Shoes? Part 1.

“The cobbler always wears the worst shoes” is a phrase that reminds us that physicians may not really take care of their own physical and emotional health. Such lack of physician self-care can have negatively affect patient care, the functioning of hospitals and clinics and the costs of healthcare. These topics are addressed by a recent Lancet article by Wallace and colleagues. 

In this article the authors point out that physicians commonly suffer from physical and emotional health problems, but are not good about seeking help and getting care for such problems. As a result they often function in a sub-optimal state of health and burn-out that can lead to low job satisfaction, increased turnover and low morale in the healthcare organization. More importantly, such ‘professional malaise’ leads to poor patient care, increased healthcare costs and medical errors. The authors propose that therefore the ‘wellness’ of physicians is an important quality indicator that healthcare systems should routinely track.

Here are the main facts that I gleaned from the article:

Frequency and Types of Health Problems among Physicians

Studies on physician health have mostly focused on illness. Such studies have shown that they often suffer from burnout, stress, depression, substance abuse and relationship troubles. In some studies, up to 75% of physicians were burnt out. Suicide rates are six times higher among doctors than in the general population. Doctors have a higher than average cardiovascular mortality. Stress and burn out are common among medical students and residents as well. 

However, health or wellness should not be considered just the absence of disease or sickness, but a positive state of well-being, including being challenged, thriving, and achieving success in various aspects of personal and professional life. Studies measuring this positive state among physicians are sparse, so we do not know exactly how often this ideal state is achieved. 

Reasons for Physician ‘Unwellness’

The reasons for physician illness and stress include external and internal factors:

External Factors

• Heavy workloads that are increasing over time (increased patient-care demands)
• Fatigue and sleep deprivation
• Having to deal with emotionally charged situations (that are associated with suffering, fear, failures and death)
• Difficult interactions with patients, families and colleagues
• High cognitive demands (the need to quickly process an overwhelming amount of information over long periods)
• Growing bureaucracy
• Increased accountability 
• Having to negotiate conflicts between the needs of the organization and patient needs, and 
• Loss of autonomy from restrictions on practice placed by external regulators and payors.

Internal factors

• Physicians commonly have certain personality traits such as perfectionism, workaholism and ‘Type A’ attitudes. These traits are associated with burnout, depression, anxiety, eating disorders and cardiovascular disease. 
• Female physicians often have to balance work and family responsibilities, which can increase work-family conflict and stress.
• Immigrant doctors can suffer from additional stressors and strains in addition to the stressors experienced by all trainees and physicians.

[to be continued.........]

Reference

Physician wellness: a missing quality indicator.Wallace et al. Lancet 2009; 374:1714

Are Physicians like Cobblers with Bad Shoes? Part 2.

[continued from previous post]

Physicians’ Self-care Practices

Physicians are not very good at caring for themselves and at seeking help for their problems. Many physicians themselves do not have a primary care provider and neglect to have physical examinations. Even when they develop a problem such as depression, very few physicians seek treatment for it. Most physicians continue to work when they are  unwell and expect the same from their colleagues. Such attitudes have been reported in trainee doctors as well as with older, established ones. In short, doctors rely on denial and avoidance to cope with their problems. 

Doctors may also not seek help out of fear that might be stigmatized as being incompetent, or unable to cope. Or they may be afraid that seeking treatment may cause licensing boards to investigate them, thus placing their medical license and career in jeopardy. They may also feel uncomfortable in the role of a patient.

The culture of the medical profession is an important deterrent to doctors taking care of their emotional and physical problems. Physicians are educated in a system that rewards individual achievement, self-reliance, independent judgment, industry and self-sacrifice. This system can also inculcate the notions that the best doctors have few needs, make no mistakes and are never ill. Physicians are under pressure from colleagues and patients to appear physically well, even when they are sick, because they believe their health is interpreted as an indicator of their medical competence. 

Colleagues’ Behavior and Support

Colleagues of unwell or impaired physicians are often not helpful to the affected physician. Often, when a physician is affected by illness or impairment, neither the physician himself nor his colleagues are willing to talk about the illness or impairment - a conspiracy of silence. Even when the impaired physician’s illness or impairment poses a threat to patient safety, and even when the affected physician is at risk of committing suicide, his colleagues may not report or address the problem, because confidentiality between colleagues is valued in the medical profession. Colleagues are more likely to report a substance-abusing physician than one who is emotionally or cognitively impaired. 

Healthcare Institutions

Healthcare organizations may not provide even basic resources for physician wellness and self-care, such as adequate rest, recovery and nutrition. 

Negative Consequences of Physician Unwellness

Burnt out or ill physicians can leave the healthcare organization or even leave the medical profession prematurely. This represents a significant loss to the organization and the profession. Replacing a physician with a new hire and training a physician are both expensive. When an organization has to employ a new physician to replace one who has left, the estimated cost is US$ 150,000 - 300,000.  

Even if physicians do not leave an organization, physician burnout causes them to be less productive and efficient. Manifestations of these include increased absenteeism, ordering unnecessary tests and procedures, reduced practice revenue, less time spent with patients, and early retirement. Suspension of a physician because of mental health problems also is costly to an organization.

Physician fatigue, stress and illness also significantly degrade the quality of medical care provided to patients and lead to medical errors. Even in the absence of overt physical or mental illness, studies have shown that physicians dissatisfied with their jobs have riskier prescribing profiles, less adherent patients, and less satisfied patients (which likely are indicators of lower quality of care). 

Residents suffering from burnout or depression reported providing poor quality care - examples of this were failure to fully discuss treatment options or answer patient questions, treatment or medication errors (not due to lack of knowledge or inexperience), and reduced attentiveness or caring behavior towards their patients. Studies also show that medical students and physicians who have a poor personal health profile are less likely than those who are healthy to recommend preventive interventions and a healthy lifestyle to their patients. 

Conversely, physician wellness and satisfaction are positively correlated with patients’ adherence to treatment and to satisfaction with care. 

Measuring Physician Wellness as a Quality Indicator

Because physicians who are unwell can adversely affect patient outcomes and the healthcare organization’s functioning, physician wellness should routinely be measured as a Quality Indicator that reflects the organization’s health. A standardized questionnaire can be used to measure this. An example is the quality of work competence survey published by Arnetz et al. This survey measures ten items - mental energy, work climate, work tempo. work-related exhaustion, skills development, organizational efficacy, and leadership. A single total (weighted) score can then be calculated. 

Interventions to Improve Physician Wellness

Some studies have reported that structured interventions to improve physician wellness can reduce burnout and sick leave. Others have shown significant reductions in medication errors and malpractice claims. More research is needed to study the impact of physician wellness programs on patient outcomes and organizational performance. Of course, individual physicians should also strive to take better care of themselves - ‘Physician heal thyself.’ 

Reference

Physician wellness: a missing quality indicator.Wallace et al. Lancet 2009;374:1714