Letters

RAPID RESPONSES FROM BMJ.COM

As of June 25th, 62 rapid responses had been posted on bmj.com in response to the editorial on banning "accidents." Barry Pless and I distilled the negative commentary into 10 key arguments, and prepared responses to each of them, which we posted as our own rapid response. A sampling of the original rapid responses (after editing), along with our reply, is presented below.EDITOR

"Accident" should not be purged

Accident looking for somewhere to happen

BMJ takes leadership role

Some daft terminology

Accidents do happen!

Injevents?

If you need more convincing, look to Katmandu

"Accident" should not be purged

This article originally appeared in BMJ USA

EDITORThe irrationality of the lay public, who need to be educated in the "facts" of injury epidemiology, is a seductive thesis, but sadly it has no basis in evidence. Detailed qualitative work ^{1 2} has demonstrated that the "lay" public largely share the understanding of public health professionalsthat accidents are predictable, and ultimately preventable, at least in theory, and that luck has little part to play in the distribution of injury. They also understand the prevention paradox: that population-level knowledge of the risk factors for accidental injury is of little help in explaining any specific individual event, except perhaps in retrospect.

Attempts to create a terminology with unnecessary neologisms, unique to medical journals, are likely to alienate further those professionals who work with communities on accident prevention.³ We already have a perfectly adequate generic term to cover the range of events that may injure people. It is "accident," a term which has shifted in meaning considerably over time² to reflect changing medical and social understanding of risk and predictability.

Judith M Green, senior lecturer in sociology. 
London School of Hygiene & Tropical Medicine, London, UK judith.green{at}lshtm.ac.uk

Accident looking for somewhere to happen

EDITORMy first reaction to attempts to replace simple words whose meaning is innate to all native English speakers with circumlocutory contrivances ("prostitute" with "commercial sex worker" springs to mind) is to reach for my dictionary. To quote from Chambers English Dictionary (7th ed, 1989): "n. ac'cident that which happens: an unforeseen or unexpected event: a chance: a mishap: an unessential quality or property: unevenness of surface."

If the BMJ editors agree with these definitions, I suspect that they will seldom find cases where "use of the term is inappropriate or misleading" because your new stricture on the use of accident invokes the exact instrument of hindsight.

For example, if a health care worker unintentionally sticks herself or himself with a contaminated needle, she or he might look back on the incident and think, "Although I now realize I ignored years of training and acted carelessly, I failed to foresee the event at the time." That event is then clearly an accident by the dictionary definition.

If, on the other hand, colleagues had seen the health care worker handling the needle carelessly, and had immediately communicated their foresight, and the injury had still taken place, that is not an accident by the dictionary definition because the specific event was foreseen.

Before trying to do away with accident, I urge you to use the benefit of hindsight. Deliberate attempts to change the use of words have failed almost without exception, whereas constant actual changes in language are an accident of history.

John McConnell, editor. 
The Lancet Infectious Diseases, London, UK john.mcconnell{at}lancet.com

BMJ takes leadership role

EDITORCongratulations to the BMJ for taking a strong, though overdue, stance on an important public health issue by recognizing in your editorial policy that injuries are not "accidents." It was common practice in the nineteenth century for an infectious disease outbreak to be described as an "act of God." One wonders when and by whom editorial policy was changed on that issue?

In most instances alternative words exist to replace "accident." Perhaps, rather than coining a new word for those instances where the English language fails us, we should explore the many other existing languages for a word with the required meaning and adopt it into English.

Joan Ozanne-Smith, professor and chair of injury prevention. 
Monash University, Victoria, Australia joan.ozanne-smith{at}general.monash.edu.au

Some daft terminology

EDITORAttempts to substitute politically correct words for those in common parlance can lead to some daft terminology. One local newspaper in South West London persistently refers to pedestrian casualties as having been "in collision" with vehicles, suggesting a rather more equal contest than that experienced by the pedestrian, who would probably say that he or she has been knocked down or run overif still in a position to comment at all.

With hindsight, the pedestrian might agree that it was not a good idea to cross the road without looking, but he or she has probably done the same thing on many previous occasions without mishap. Statistically, there may have been the same probability of meeting a car each time, but humans do not conduct a Bayesian analysis every time they cross the road.

After crossing Cromwell Road twice a day for 10 years, I felt that a close encounter with a No. 74 bus was becoming a near certainty. Having now moved to Orkney, getting mowed down by one would qualify as an achievement rather than an accident.

Anna Whelan, corporate services officer. 
Orkney Islands Council, UK whelan{at}zetnet.co.uk

Accidents do happen!

EDITORI was surprised and disappointed at the suggestion of censoring the richness of the English language from your journal. Accidents do occur, and to remove this word from our vocabulary is to deny that sometimes no one is to blame when things go wrong. Some "accidents" are the consequences of other people's actions, but the very word "accidental" gives us the opportunity to accept that injuries may not be as a result of a deliberate plan to harm. Accidents will and do happen!

Lucy-Jane Davis, study administrator. 
National Perinatal Epidemiology Unit,University of Oxford, Oxford, UK jane.davis{at}perinatal-epidemiology.oxford.ac.uk

Injevents?

EDITORCongratulations to the BMJ for this bold step. The trick will be to develop a commonly accepted substitute (injevents?). Whatever the choice, be prepared for a long wait for general recognition!

Susan P Baker, professor. 
Johns Hopkins School of Public Health, Baltimore, Maryland, USA sbaker{at}jhsph.edu

If you need more convincing, look to Katmandu

EDITORBanning incorrect use of the word "accident" is one more outstanding move by one outstanding journal. Incorrect use of the word causes more harm than good. The ultimate rejection of the word occurred just before publication of your editorialin Katmandu when the Nepalese palace spokesperson announced that the royal family, who had been mowed down by gunfire from a semi-automatic gun and then systematically shot through the head to finish them off, had died from an "accident." The Nepalese people then rioted in the streets. Can't we do better than that?

Mark L Rosenberg. 
Task Force for Child Survival and Development, Decatur, Georgia, USA mrosenberg{at}taskforce.org

Ron Davis, North American editor. 
BMJ

Barry Pless, editor, "Injury Prevention". 

IN REPLYWe appreciate the thoughtful comments from readers about our editorial on banning "accidents." As of June 25th, 62 rapid responses had been posted, divided equally among those supporting and opposing our new editorial policy. We have had a similar debate about the policy among the editorial staff of the BMJ, reflecting strongly held views on both sides of the issue.

After carefully considering all of the comments we've received, we continue to believe that we have chosen a proper course of action. Here are some of the key arguments against our new policy and our responses:

1. "Accident" is a perfectly good term and its meaning is not ambiguous.

We disagree. The commentary running through the first 48 rapid responses has confirmed that the word has many different meanings. One responder noted that "accident" can refer to the event, the injury, or the precursors of the event. Another added that it covers "everything from spilled milk to death."

2. Some "accidents" occur when no one is to blame. "Accidents" do occur.

We agree, and as we pointed out in our editorial, "some injury producing events may seem to be attributable to bad luck or acts of God and thus not preventable." But in the case of most injuries, risk factors were present before the injuries occurred and could have been eliminated or diminished.

In some cases, the injury-producing event is difficult or impossible to predict or prevent, but steps can be taken to reduce the risk or severity of injury. It may be impossible, for example, to avoid a collision with a drunk driver who crosses into opposing traffic. But wearing a safety belt, driving an automobile with an air bag, and avoiding excessive driving speeds will almost always reduce the severity of injury when an unavoidable collision occurs.

3. At the population level, most injuries are predictable and preventable. But at the individual level, they are not. Injuries at the individual level are predictable and preventable only by hindsight.

We disagree. Safety experts see injuries waiting to happen as soon as they enter unsafe environments. The challenge is to extend that same sensitivity to all those who control the safety of homes, vehicles, schools, worksites, medical facilities, and other venues where injuries often occur. If schools do not have a soft surface below playground equipment, if parents store medications where they are easily accessible to children, if factory workers do not use proper protective gear, and if hospitals store concentrated potassium chloride (instead of appropriately diluted solutions) on patient care units, then injuries will predictably occurwith foresight as well as hindsight.

4. Most people understand that "accidents" are preventable.

In a national survey of 943 adults in the United States, Girasek found that 83% believed that "accidents" usually could have been prevented.¹ This means, of course, that 17% did not hold that belief. In addition, the survey showed that only 25% of respondents believed that "accidents" usually could have been predicted, and that 26% believed that "accidents" are controlled by fate. Moreover, subjects with less education were less likely to believe that "accidents" are predictable, and more likely to believe that "accidents" are controlled by fate.

These findings do not contradict the statement in our editorial that "An accident is often understood to be unpredictablea chance occurrence or an `act of God'and therefore unavoidable." If many people hold that belief, especially within disadvantaged strata in the population (as the findings by Girasek would seem to indicate), then the case for eschewing "accident" is supported. But beyond the matter of preventability and predictability, the imprecision of the meaning of "accident" is reason enough to avoid using it in scientific discourse.

5. You are censoring the English language; your policy is Orwellian and totalitarian.

Some of this commentary, we suspect, was a reaction to the word "ban." Although the title and first paragraph of the editorial announced a "ban" on "accident" without qualifiers, the rest of the editorial explained that we are banning "inappropriate use" of the word, that "implementation of the ban will not be draconian," and that there will be exceptions when "accident" is allowed (see item #10 below).

We believe that in most instances, the word is used inappropriately or it can be replaced easily with alternative language. Thus, we expect that the word will become uncommon inbut will not disappear fromthe BMJ.

We see nothing Orwellian about the policy. Every language, including English, varies from place to place and changes over time. Differences within a language are seen among ethnic, geographic, and socioeconomic groups in a population. Occupational disciplines develop their own professional languages, which often redefine or reject common lay terms.

To insist that a scientific publication use a lay term with its ambiguous lay meaning strikes us as much more Orwellian and anti-language than our new policy. One of the responders compared the word "accident" with the word "heart attack." If a cardiology journal rejected use of the term "heart attack" in its research papers, would we cry "censorship"?

6. Deliberate attempts to change the English language have failed in the past.

We agree. But our main goal is not to change the English language wherever it is spoken. Our primary objective is to speak clearly and properly within the BMJ. We hope that other communicators in medicine and science will do the same. When that happens, perhaps the move away from "accidents" will trickle down into the general population, but we will not hold our breath waiting for that to happen.

7. There is no evidence that banning "accident" will reduce the burden of injuries.

We agree. But there is no evidence that avoiding the word will cause harm. As we see it, the benefit of speaking clearly and properly is justification enough for the new policy, in the absence of "hard evidence" regarding good or bad outcomes. As we state in the editorial, "we believe that correct and consistent terminology will help improve understanding that injuries of all kinds . . . are usually preventable. Such awareness, coupled with efforts to implement prevention strategies, will help reduce the incidence and severity of injuries." (emphasis added)

8. Banning "accident" contributes to a culture of blame.

This argument doesn't get to the crux of the issue. The word "accident" has an effect on the attribution of blame, whereas "injury" is neutral in regards to blame. Because of its connotation that the event was unpredictable or unavoidable, "accident" protects against blame. When we use "accident," we exonerate from blame the person injured, the parent, the manager, the facility, the company, or "the system." In so doing, we discourage the search for root causes and their remediation. On the other hand, as one of the responders noted, "The use of the term injury rather than accident focuses attention on the damage to the person and what can be done to prevent this damage."

9. If we ban "accident," we will ignore the lessons to be learned from examining "near misses."

Abandoning "accident" and focusing on the injury and its antecedents does not preclude attention to "near misses." It simply means that we will not refer to near misses as "accidents." Schemes for reporting near misses have been institutionalized in several non-medical sectors such as aviation, nuclear power technology, petrochemical processing, steel production, and military operations; and similar reporting systems are being developed in medicine. These systems employ many terms besides "accidents" to describe near misses, including "near misses," "close calls," "errors," "unsafe actions," "critical incidents," "dangerous occurrences," "sentinel events," "non-consequential events," and "adverse events" (eg, "adverse drug events").²

10. Banning "accident" will waste hours of editors' time searching for clumsy replacement language.

We have heard this concern from some BMJ editors, and one of the rapid responses (from a copy-editing manager) stated that "most will have great difficulty . . . knowing which word to use when." Despite those misgivings, we are optimistic that editors and copy-editors (sub-editors) will have little difficulty in finding replacement language. With experience, the task should become easier.

Some journals have been working for many years to abstain from "accidents," such as Injury Prevention (published by the BMJ Publishing Group) and Morbidity and Mortality Weekly Report (published by the US Centers for Disease Control and Prevention). Neither the editor nor the technical editor of Injury Prevention has had difficulty in replacing "accident" with "injury" or some other phrase. When MMWR is forced to use "accident"for example, when it publishes national mortality data for which the term is still usedit adds this footnote: "When a death occurs under `accidental' circumstances, the preferred term within the public health community is `unintentional injury.'"³ The "injury control community" in North America has been successful in communicating about injuries without "accident."

If other journals and people who work in the field can find a way to avoid using the word, then the BMJ and other journals should be able to do so as well. Our editorial provided suggestions for alternative wording. We are willing to serve as resources to BMJ editors and others when they are struggling to avoid "accidents." If instances occur when the clumsiness quotient is too high, then "accident" should be allowed. And as mentioned above and in our editorial, there will be exceptions when "accident" is considered acceptable and alternatives need not be sought. MMWR, for example, published recommendations recently on responding to fecal accidents in swimming venues⁴; we would not be troubled to see "fecal accidents" in the BMJ (on the BMJ would be a different matter).

One of us (RD) went through the exercise of purging "accident" from this issue's Lesson of the Month (p 401) on handlebar injuries from bicycle "accidents" in which the word appeared several times.⁵ Here are the results:

• Paragraph 1, sentence 1: Change "Accidents" to "Unintentional injuries."
• Paragraph 1, sentence 2: Change "after bicycle accidents in children" to "when bicycle falls and collisions occur among children."
• Paragraph 1, sentence 3: Change "Bicycle accidents" to "Bicycle falls and collisions."
• Paragraph 1, last sentence: Change "accident" to "incident."
• Paragraph 9, sentence 2: Change "motor vehicle accidents" to "motor vehicle collisions."
• Paragraph 10, sentence 1: Change "Bicycle accidents" to "Bicycle falls and collisions."
• Paragraph 11, sentence 3: Change "Bicycle accidents" to "Bicycle falls and collisions."
• Paragraph 11, last sentence: Change "accidents" to "incidents."

1.	Girasek DC. How members of the public interpret the word accident. Injury Prevention 1999; 5: 19-25[Abstract/Free Full Text].
2.	Barach P, Small SD. Reporting and preventing medical mishaps: lessons from non-medical near miss reporting systems. BMJ 2000; 320: 759-763[Free Full Text].
3.	US Centers for Disease Control and Prevention. Mortality patternsUnited States, 1997. MMWR 1999; 48: 664-668[Medline].
4.	US Centers for Disease Control and Prevention. Responding to fecal accidents in disinfected swimming venues. MMWR 2001; 50: 416-417[Medline].
5.	Lam JPH, Eunson GJ, Fraser DM, Orr JD. Delayed presentation of handlebar injuries in children. BMJ 2001; 322: 1288-1289[Free Full Text].