Identifying Hazards and Communicating Risks

Identifying Environmental Problems

Environmental hazards (see Definition box below: Hazards and risks) cause a wide range of diseases and environmental factors are involved in the aetiology of virtually every illness. Although a high index of suspicion is essential for diagnosis, once diagnosed, environmentally-induced diseases often respond to simple environmental solutions. Moreover, many patients worry about elements of their environment that have little or no effect on health, so physicians need to be able to distinguish those that are hazardous from those that are not. To diagnose an illness as predominantly environmental in origin, physicians have to include a thorough environmental assessment in their history-taking. Clues that an illness is caused by an environmental factor include:

• The patient suspects it
• The pattern of illness is atypical; for example, the patient is not in the usual age group, the usual risk factors are absent, the course of the illness is unusual, the symptoms do not respond to the usual treatments
• The temporal pattern of the illness suggests it (for example, symptoms improve when the patient goes on holiday and worsen on returning home, or they worsen when the patient is at work and improve at home)
• There is no obvious other cause for the illness
• The signs and symptoms suggest specific toxins, such as lead or mercury poisoning.

Once an environmental cause is suspected, the physician should take a detailed environmental history to identify all hazards the patient may have been exposed to, as shown in the box “Taking an environmental history”. All of the patient’s activities and environments should be explored. The chronology of events, patient’s proximity to the presumed source, and whether or not other people are affected should be explored for evidence that supports or discounts the hypothesis of an environmental cause (i.e., establishing Time, Place, and Person).

Once the physician has established that the problem is likely related to the environment, he should ensure that other people exposed are protected. He can call the public health service for advice on how to manage health problems linked to the environment. Indeed, some provinces require physicians to notify the public health department of particular environmental illnesses or possible outbreaks. In the case of a hazard that is likely to affect the population at large, the public health service is responsible for controlling the problem. However, even when the problem is community wide, the public health service may only be able to educate and inform. At times, public health services advise on the safety of projects proposed by industry or governments.

The patient’s work health and safety board (see Chapter 12, Occupational health service) is responsible in the case of a work hazard. In an emergency, as with a case of carbon monoxide poisoning, public security services (fire and police) would coordinate control efforts.

Taking an environmental history¹

Ask about the following topics (CH₂OPD₂ is a useful mnemonic):

• Community or neighbourhood sources of hazard: pollution; industry; waste storage
• Home: year of construction, renovations; materials used in construction and decoration; moulds; garden and house plants; use of cleaning products, pesticides, herbicides
• Hobbies and leisure: exposure to chemicals, heavy metals, dusts, or micro-organisms
• Occupation: current and previous occupations; work with known hazards; air quality
• Personal habits: hygiene products; smoking
• Diet: sources of food and water; cooking methods; food fads
• Drugs: prescription, non-prescription, and alternative medications; health practices.

If a scanning question reveals a possible hazard, ask detailed questions to find out as much as possible about the nature and level of the hazard and then check Time, Place and Person:

• Time: When did symptoms begin? When did the exposure begin? When do symptoms get worse? When do they improve?
• Place: Where is the patient when symptoms get worse? Where is the likely hazard? What is the channel through which the hazard reaches the patient?
• Person: Does anyone else have similar symptoms? Who? When? Where?

Reducing Risk

The three main activities in addressing health problems possibly linked to environmental hazards are risk assessment, risk management, and risk communication. The clinician addresses these when discussing environmental disease with a patient; a public health officer applies them in responding to community wide problems.

Risk assessment

Risk assessment involves a series of steps in evaluating the likelihood of occurrence and probable severity of health effects due to a hazard. Various agencies assess hazards, including occupational health agencies, environmental protection agencies, and public health authorities. Clinicians confronted with a case of disease with possible environmental links might use some of the following risk assessment steps to arrive at a diagnosis.

Steps in risk assessment for clinicians²

1. Hazard identification: Are environmental hazards involved? What are they? How likely are people to be exposed and what will the effects be?
2. Exposure assessment: Is the patient’s exposure to the hazard sufficient to cause these symptoms? Which population groups are likely to be at most risk?
3. Risk characterization: Given the patient’s level of exposure, is the hazard likely to cause these types of symptoms in this type of patient? What are the dose response curves for different routes of exposure (inhalation, ingestion, etc.)?
4. Risk estimation: How much has the hazard contributed to the patient’s condition? What is the risk from this hazard for this specific population?

Risk assessment, step 1: Hazard identification

The first step identifies the agents possibly responsible for the problem. Each hazard is defined by its ability to cause adverse effect, under what conditions and to which population groups. Hazards are classified as biological, chemical, physical, ergonomic, psychosocial, or related to safety. Table 10.1 gives examples of the various types of hazard and their possible health effects.

Table 10.1: Examples of the various types of environmental agents and their associated health effects

Hazards can be found almost everywhere in the environment, including the air, water, soil, and in food. Hazards may affect health directly, or indirectly when they produce a change that puts a person or population at risk, for instance climate change. Table 10.2 lists some of the common hazards.

Table 10.2: Examples of hazards in the environment

Indoor hazards that cause health problems include household chemical products, carbon monoxide, radon, mould, lead, and consumer products (cosmetics, perfumes, hygiene products). Second hand smoke is still a problem in some households.

The clinical picture may indicate what kind of hazard to search for. Airborne irritants can cause itchy or sore eyes, runny nose or coughing. Irritants that come into direct contact with the skin can cause dermatitis. Allergens can cause numerous manifestations of allergy including dermatitis, asthma, sneezing. Asphyxiants cause different respiratory problems depending on the type of asphyxiant. For instance carbon monoxide blocks oxygen transport. Other gases can accumulate in high concentrations displacing oxygen in the air. Certain ingested or absorbed substances may damage organ systems in pathognomonic ways. For instance, severe lead poisoning causes neurological changes, abdominal pain, or anaemia, whereas poisoning with mercury typically produces tremors, among other neurological symptoms. Mesothelioma is linked to exposure to asbestos.

The effects of a hazard may be delayed, sometimes for years. This is especially true of cancer-causing agents. Identifying this type of hazard and evaluating claims about it can be difficult. Furthermore people can be exposed to more than one hazard, each contributing to their health problem. For instance, in the case of a person who works in construction and who smokes, it can be difficult to assess the relative contributions of his occupation and his smoking to his chronic lung disease.

If an outbreak of a disease of environmental origin is suspected, public health officials will collect initial information on the possible sources. Once a case definition (see CASE DEFINITION in Glossary) is established, cases are sought and information is collected on when and where they were exposed, when they started showing effects, and demographic characteristics and predisposing factors (Time, Place and Person). Information from all cases is then collated to produce a picture of the distribution of cases in time and place (see Chapter 11: Detection and control of outbreaks).

Risk assessment, step 2: Risk characterization

This step describes the potential health effects of a hazard and answers the clinical question of whether the identified hazard could have caused this patient’s symptoms. As far as possible, effects on molecular, biochemical, cellular and organ systems are described. A chemical hazard may only cause health problems when in a specific form: ingested elemental liquid mercury is virtually non-toxic, however mercury vapour is toxic, as are all mercury organic and inorganic compounds. The route of entry into the body may also be an important determinant of the damage caused: there is no convincing evidence that ingested asbestos is toxic, whereas inhaled asbestos causes asbestosis, mesothelioma and other cancers.³  Furthermore, a person’s response to a hazard is mediated by the factors that influence the hazard’s toxicokinetics and toxicodynamics (See Definitions: Kinetics and dynamics). The individual’s genetic makeup and metabolic state as well as environmental factors moderate the toxicokinetics and toxicodynamics, so susceptibility to damage by a hazard varies from one person to another; malaria offers an example.

A low dose of some substances may be beneficial for a person, yet a higher dose is toxic. For instance, fat soluble vitamins, such as A and D, are essential for health, but too much of either is toxic. Sunlight in small doses increases the production of Vitamin D, while in high doses it can cause skin cancer. Some hazards, such as heat and noise, must reach a threshold level before damage occurs. Others, including many cancer-causing agents, are presumed to cause damage even at the lowest measurable levels. The effects of some agents, such as X-rays, are cumulative over a lifetime, while other agents, such as alcohol, allow the body to recuperate somewhat during temporary breaks in exposure.

For information on hazardous substances, clinicians can look to the scientific literature, check with the local public health department or call the toxicology or poison centre. If the probable source of the hazard is an industrial product, information is likely to be available via the Workplace Hazardous Materials Information System (WHMIS), which sets out the labelling requirements for hazards, see Table 10.3. The product label also indicates if a ‘Material Safety Data Sheet’ is available for the product. This sheet contains further details about the hazard, how to handle it safely and what to do in an emergency.

Table 10.3: The Workplace Hazardous Materials Information System Symbols 6

	Class A – Compressed Gas	Contents under pressure. Cylinder may explode or burst when heated, dropped or damaged.
	Class B – Flammable and Combustible Material	May catch fire when exposed to heat, spark or flame. May burst into flames.
	Class C – Oxidizing Material	May cause fire or explosion when in contact with wood, fuels or other combustible material.
	Class D, Poisonous and Infectious Material, Division 1: Immediate and serious toxic effects	Poisonous substance. A single exposure may be fatal or cause serious or permanent damage to health.
	Class D, Poisonous and Infectious Material, Division 2: Other toxic effects	Poisonous substance. May cause irritation. Repeated exposure may cause cancer, birth defects, or other permanent damage.
	Class D, Poisonous and Infectious Material, Division 3: Biohazardous infectious materials	May cause disease or serious illness. Drastic exposures may result in death.
	Class E – Corrosive Material	Can cause burns to eyes, skin or respiratory system.
	Class F – Dangerously Reactive Material	May react violently, causing explosion, fire or release of toxic gases, when exposed to light, heat, vibration or extreme temperatures.

Risk assessment, step 3: Exposure assessment

Exposure assessment is the step that quantifies the exposure of a person or population to a hazard. The levels of some hazards can be directly measured, either in the environment or in the people exposed. More often, however, exposure must be estimated from a careful history of the patient’s activities, as well as an inspection of the environment (see box on thalidomide). Details of activities, working practices, and processes during which a person is likely to be exposed must be examined to determine how and how much the person was exposed to the hazard.

When the problem may have been caused by an agent that produces delayed effects, such as a carcinogen, the exposure history should trace exposures back twenty years or more. Occupational disease can occur in retired people; in workers, it can be caused by previous employment.

Risk assessment, Step 4: Risk estimation

Risk estimation quantifies the likelihood that a hazard will affect a specific person or population; it also estimates the size or severity of the effect. In this step, the information obtained during the previous steps is summarized and collated. The epidemiological triad (see Chapter 2, Figure 2.8) can be used to assemble the information on what the hazard is, where it comes from, how the environment allows it to come into contact with the host, and the host’s susceptibility to the hazard in order to arrive at a conclusion on who is at what level of risk. Table 10.4 illustrates what needs to be considered in this step.

Table 10.4: Examples of various host, agent, and environmental risk factors for selected categories of health problems

The risks for an individual and for a population should be balanced against the costs and risks of intervening. In individuals it might be a choice between the risk of asthma and the loss of a beloved pet or, if the risk is work-related, it could be a choice between staying healthy or staying employed. In populations, the options can affect a number of sectors and lead to political discussion. For instance, a population may be at risk from pollutants emanating from a factory. It might, however, be at greater risk from the socioeconomic fall-out from closing the factory. Once the levels of risks due to the different hazards are documented, possibilities for risk reduction can be explored.

Occupational Hazards

Most people spend a considerable part of their life at work, and if hazards are present the workers’ exposure to them is likely to be prolonged. In some occupations, specific hazardous agents must be used, putting workers at risk of occupational diseases associated with them. Physicians should ask patients four simple questions: “What sort of work do you do?” “What sorts of work have you done?” “Do you think your work may be a cause of your problem?” And “What job are you going back to?”

Common, non-specific work-related diseases include dermatitis, asthma, and musculoskeletal disorders. It is also accepted that mental health can be influenced by stressful conditions at work.

Table 10.5: Some examples of occupational disease

Injuries do not occur randomly; groups at particular risk of occupational injury include:

• • Young workers lacking in work experience or safety training—particularly those in temporary or summer jobs;

• • Workers obliged to work long hours at a fast pace. This could include piece-workers or those working on production lines;

• • Workers given work responsibility without the authority or control necessary to meet that responsibility; and

• • Workers in certain high-risk industries such as construction and forestry.

According to Canadian federal and provincial laws, specific services must be provided for the protection of workers (see Chapter 12, Occupational Health Services) and for compensating those who are harmed as a result of work.

Risk Management

The source-path-receiver model, derived from energy engineering, is used to propose ways of controlling risk. The source is the equipment or process that is directly responsible for a hazard. The hazard could be a form of energy, such as acoustic, thermal, etc., or it could be a substance, such as toxic fumes or dusts. The path is the conveying medium (air, water, etc.), while the receiver is the human being, the worker. In contrast to the epidemiological model, which allows for bidirectional influences, the source-path-receiver model is unidirectional.

As complete elimination of risk is generally impossible, risk management aims to reduce the risk from the hazard without causing another problem. Strategies to reduce risk can target the source of the hazard, the path or the receiver, whether the risk involves a person or a population. Strategies can be educational or advisory, technical or engineered, economic or regulatory. As illustrated in Figure 10.1, strategies aimed at the source can modify or substitute the hazard (for example, replacing asbestos with other materials), or enclose it if it cannot be removed. Legislative strategies can be used to reduce the emission of hazardous pollutants.

Modifications to the path include environmental changes; for example, closing windows to prevent smog from entering the home; using bed nets to prevent transmission of malaria, or a film of environmentally safe soap or oil on standing water to drown mosquito larvae.

Finally, strategies can alter individual and population susceptibility to the risks of hazards. Examples include immunization against infectious disease, ensuring the nutritional status of disadvantaged children, and counselling for behaviour change. Personal safety equipment can reduce exposure. For example, power line workers use special tents in winter to reduce exposure to cold, and airport apron workers use ear protectors to prevent noise damage. Doctors may also promote other safety measures: for instance safety harnesses, helmets and steel-capped boots for construction workers and miners.

Haddon’s matrix and injury prevention

Dr. William Haddon combined the epidemiological triad (see Chapter 2, Figure 2.8) with a time dimension to create Haddon’s matrix, which sets out the factors that determine the level of risk associated with a hazard and the severity of its effects, in order to identify modifiable risk factors for injury.¹¹ It is commonly applied to the analysis of road traffic collisions, as illustrated in Table 10.6. In this example the host is the person injured (driver or other), while the agent is the equipment that determines how much energy (mechanical or thermal) is transmitted to the host. The environment refers to the physical and social environments in which the injury occurs. The time axis distinguishes factors that operate before the injury (e.g., recent snow had made the road slippery) from those that operate at the moment of injury (e.g., the driver was wearing a seat belt). Other factors operate following the collision (e.g., bystanders knew CPR).

Table 10.6: Illustration of Haddon’s matrix applied to a motor vehicle collision

The matrix is now being used in other situations, including the assessment of medical errors. An advantage of the matrix is that it gives the full picture of the problem instead of focussing blame on individuals; it leads to a more constructive consideration of ways to reduce this type of error. Table 10.7 illustrates a situation in which potassium was used instead of saline for dissolving a drug that was then administered intravenously, causing severe cardiac arrhythmia and leading to the collapse of the patient.

Table 10.7: Illustration of Haddon’s matrix applied to an incident of erroneous intravenous administration of potassium

 As a way to enhance patient safety, completing the matrix identifies hazards as well as protective factors. The protective factors, such as nurses monitoring patients while they administer injections, calling promptly for help, and quick response of the cardiac team should be reinforced, while the risk factors—storage and packaging of solutions, poor lighting, inadequate supervision, and harassment by other team members—can be rectified. (To prevent this type of mistake, most hospitals do not store potassium on the ward.)

Active versus passive interventions

Interventions to reduce risks are categorised according to the participation required of the person at risk. Active interventions rely on continued compliance for their effectiveness, whereas passive interventions are applied once only, or else their effects last for some time. In general, passive interventions are more effective because people do not have to remember to use them. For instance, administration of influenza vaccine (passive) confers immunity for the whole influenza season, whereas hand-washing (active) must be repeated several times a day. An airbag (passive), once installed in a car, stays there until it is needed, whereas seat belts (active) must be put on every time. Even in well-managed workplaces, people’s compliance with the use of personal safety equipment can be poor. For example, ear protectors can be hot and uncomfortable, and may mask auditory danger signals, so that reducing environmental noise levels is preferable, where possible. Safety harnesses can impede mobility, but in some situations they can be replaced by guard rails or safety cages.

Risk Communication

Once a risk has been identified and quantified and methods of reducing it have been found, the information has to be communicated to people at risk to allow them to understand their risk and take steps to mitigate it. Communication is the exchange of information; good communication is the exchange of information in a way that the recipient understands what the sender intends (Figure 10.2). A model of communication derived from communication technology offers a useful basis for thinking about interpersonal communication. This model distinguishes six elements in the communication process: the message, the messenger, the encoding, the channel, the decoding, and the recipient.

The message

How risk is perceived depends on factors beyond the simple statistical level of risk; perception varies according to the nature of the hazard (the exposure) and according to its possible effects (the outcome). Factors that increase the perception of danger are listed in Table 10.8.

• Involuntary
• Not under personal control
• Unnatural (e.g., terrorist attack)
• Unfamiliar
• Mismatch between risk and benefit (people who suffer the consequences experience no gain from the activity)
• No trust in institutions involved
• Media attention

• Catastrophic (instead of chronic)
• Affects children or future generations
• Unknown or uncertain outcome
• Affects identifiable people, not statistics
• Dreaded outcomes (e.g., cancer)
• Immediate (vs. delayed)
• Irreversible
• Media attention

These characteristics can transform a minor statistical risk into a major perceived risk. For example, in the U.S., the risk of death per mile travelled in a car is about 10 times the risk of death per mile travelled in a commercial airplane, but people are appalled when they hear about an airplane crash. News of a car crash has much less effect. From Table 10.8, the difference in reaction is due to the unfamiliarity of the event and the fact that it is beyond the control of the traveller. Furthermore, the death toll from an airline crash is generally in the hundreds and so they are seen as catastrophic, receiving a great deal of media attention. By contrast, road traffic fatalities are familiar, under the control of road users, and each only affects a few people.

Table 10.8 also offers some insight into why some people may worry about environmental risks and yet refuse to change their own high-risk lifestyles. For example, smokers can be anxious about small environmental risks that they cannot control, while continuing to ignore the great risk of smoking.

Finally, the message may be obscured by uncertainty. Clinicians should be as precise as possible when giving information:

• Words can be ambiguous and their meaning can vary: terms such as high or low risk are unclear. Clinicians should make certain that their patients understand the precise meaning of the possible outcome and the course of action suggested.
• If numbers are used instead of words, be aware of the biases inherent in a positive versus a negative framing of statistics (see the next section).
• The definition of the risk may lack specificity. The risk should always include a time frame. A 10% chance of death from a lung cancer over a lifetime is not the same as a 10% chance in the next five years. A 20% chance of loss of function could be interpreted as a 20% reduction in function, or that 20% of people will have complete loss of function, or that 20% of people will have some loss of function.

Framing the message

In relation to communication and decision-making, framing refers to the way in which information on risk is presented. Crucially, large changes in preferences can be caused by minor variations in the words used in presenting a choice.¹⁴

Emphasis on gain versus emphasis on loss

People tend to be averse to loss and will do more to avert a perceived loss than they will to achieve a gain. See Nerd’s corners: Treatment A or B and Glass half full.

The default option

The option that is presented as the usual choice is more likely to be chosen than one presented as the alternative.

Numbers versus proportions

Data expressed as proportions tend to be seen as relatively benign, whereas data expressed as a frequency tend to engage people much more.

The messenger

People respond more to the attitude of the messenger than to his or her status as a professional or authority. People tend to disregard information given by a recognized expert if he shows a lack of caring or empathy (Figure 10.3). Therefore, to get a message across to a patient, a clinician needs to show a caring attitude rather than trying to impress with science.

The recipient

The recipient is an active participant in communication. The recipient’s prior knowledge, beliefs, attitudes and experience affect his understanding of the message (see Nerd’s corner People and their perceptions). When communicating with patients, a clinician needs to assess the patient’s state and adapt the message accordingly.

The channel

The usual channel for clinical communication is the spoken word: clinician and patient talk to each other. However, words can be supplemented with visual aids, such as posters, leaflets, and, occasionally, videos. The recipient must have access to the channel of communication: written materials may not reach people who have difficulty reading.

Encoding and decoding

Information must be coded before it can be passed on. For communication to be successful, the messenger and the recipient must share a common understanding of the code. Even though a clinician speaks the same language as the patient, differences in socio-economic milieu, education and experience can limit their shared understanding of terms. When a physician discusses an appendectomy with a patient, the physician is talking about a routine procedure that the patient is likely to recover from in a few days. The patient, on the other hand, is hearing about an alarming and unique experience that is likely to be painful, will leave a permanent scar, and will disrupt life for, at the very least, several days.

When communicating with their patients, clinicians should:

• Use words and concepts that their patients can understand;
• Remember that clinicians are familiar with medical conditions and procedures, whereas their patients are not and that the patient’s perception is clouded by apprehension;
• Modify their tone and body language to conform to the common code, bearing in mind that gestures may mean different things in different cultures. In North America, a Greek who shakes his head may be asking a question, not saying “no,” while an Italian gesturing for someone to come closer may appear to be waving goodbye. Some people feel that it is polite to look an interlocutor in the eye, others find it threatening.

Risk information is often presented as numbers or graphs. Most people can grasp the information in these forms, but there are a number of points to remember when using numbers to communicate information:

• People don’t come in halves
  People are more at ease with whole numbers, so fractions and decimal places should be avoided. Few estimates of risk or of therapeutic benefit are so precise that they merit decimals.
• Numerator and denominator
  People tend to focus on the numerator and ignore the denominator. A disease that afflicts ten people in a hundred tends to be seen as less common than one that afflicts a hundred people in a thousand. When discussing a single risk with a patient it can be helpful to express it in several ways: 10%, or one in ten, or ten in a hundred. However, when asking people to compare risks of different outcomes, one should keep the denominator constant. It is difficult for most people to understand the difference between one in five risk of one outcome and a 25% risk of another outcome. They will find it easier if a 20% risk is compared to a 25% one or if a one in five risk is compared with a one in four risk.
• Relative and absolute risk
  Relative risks may be particularly difficult to interpret because people rarely know the context. Although hormone replacement therapy doubles the risk of breast cancer, it causes only eight additional cases of breast cancer in 10,000 woman-years. For an individual woman, doubling the risk does not greatly increase her absolute risk because the baseline risk is so small, whereas, on a population level, a doubling of the risk may be significant. It is advisable to use only absolute risks when communicating with individuals, because proportional changes may often obscure a lack of substantive importance.¹⁶ (See Chapter 5 for a discussion of relative and absolute risk).

When discussing risk with a patient, the information should be communicated in different ways, using both positive and negative framing, to help the patient arrive at a fully informed decision.

An effective message contains more than simple information; it also implies what the recipient should do with the information. For example, after discussing the risks and benefits of exercise with a patient, a clinician should conclude by relating this information to the patient’s personal situation and making it clear that the patient should take more exercise.

Self-test Questions

References

1. Marshall L, Weir E, Abelsohn A, Sanborn MD. Identifying and managing adverse environmental health effects: 1. Taking an exposure history. Canadian Medical Association journal. 2002;166(8):1049-55.
2. Porta M, editor. A dictionary of epidemiology. New York (NY): Oxford University Press; 2008.
3. World Health Organization. Asbestos in drinking-water: background document for development of WHO guidelines for drinking-water quality 2003 [cited 2016 June]. Available from: http://www.who.int/water_sanitation_health/dwq/chemicals/asbestos/en/.
4. Aidoo M, Terlouw DJ, Kolczak MS, McElroy PD, ter Kuile FO, Kariuki S, et al. Protective effects of the sickle cell gene against malaria morbidity and mortality. Lancet. 2002;359(9314):1311-2.
5. Dorland WAN. Dorland’s illustrated medical dictionary. 37th ed. Philadelphia, PA.: Saunders Elsevier; 2007.
6. Canadian Centre for Occupational Health and Safety. WHMIS 2015 – Labels. 2015 [cited 2016 June]. Available from: http://www.ccohs.ca/oshanswers/chemicals/whmis_ghs/labels.html.
7. Theoret JJ. Chronology of the Thalidomide story. Canadian Medical Association journal. 1962;87(18):981-2.
8. Cox LA, Jr. Does concern-driven risk management provide a viable alternative to QRA? Risk Analysis. 2007;27(1):27-43.
9. Ontario Ministry of the Environment. Drinking water system at the Kashechewan First Nation: Technical Report. Toronto, Ontario: Ontario Ministry of the Environment, 2005.
10. Eggertson L. Despite federal promises, First Nations’ water problems persist. Canadian Medical Association journal. 2008;178(8):985.
11. Haddon W, Jr. A logical framework for categorizing highway safety phenomena and activity. Journal of Trauma-Injury Infection & Critical Care. 1972;12(3):193-207.
12. Slovic P. Perception of risk. Science. 1987;236(4799):280-5.
13. Covello VT, Peters RG, Wojtecki JG, Hyde RC. Risk communication, the West Nile virus epidemic, and bioterrorism: responding to the communication challenges posed by the intentional or unintentional release of a pathogen in an urban setting. Journal of Urban Health. 2001;78(2):382-91.
14. Tversky A, Kahneman D. The framing of decisions and the psychology of choice. Science. 1981;211(4481):453-8.
15. Slovic P, Monahan J, MacGregor DM. Violence risk assessment and risk communication: the effects of using actual cases and employing probablility vs. frequency formats. Law and Human Behavior. 2000;24(3):271-96.
16. Spasoff RA. Epidemiologic methods for health policy. New York: Oxford University Press; 1999.