To include foodborne illness in the passive surveillance system, the following steps have to occur: A person becomes ill, and the patient must go to a doctor. The doctor must request specimens for analysis. The laboratory must analyze the specimen, the positive results must be reported to the health department, and the reports must be forwarded to the CDC. (Editor's comment: Some reporting systems do not require lab confirmation. Also, some states require that clinically diagnosed ailments be reported, even when laboratory confirmation is not available.) Some labs can test for a variety of microorganisms and use biomolecular typing to determine whether isolates are related to the same outbreak. Usually, tests for toxins, viruses, and parasites are run only if specifically ordered. Other labs have a more limited capacity and can run only routine tests.

Another example of passive surveillance is when someone becomes ill after eating at a restaurant and then complains to the health department. People who are concerned enough to take the time to file a report to a surveillance program want assurance that the appropriate person will be notified and that immediate action will be taken. Surveillance programs should be organized to receive and respond to complaints. The process for reporting should be quick and traceable to ensure that the appropriate person is informed regardless of the point of the first contact. The surveillance system should be "triage-oriented" (respond differently depending upon the circumstances) and should be able to proceed smoothly from one level of action to the next.

The surveillance system should have a referral process that ensures a timely and competent public health response. And finally, the system must be able to terminate effectively when resolution is reached and have a mechanism for feedback notification. If you try to analyze calls and complaints on an individual basis, it's difficult to determine the source of an exposure. That's where the surveillance log comes in. A surveillance log is simply a record of illness complaints.

Basic information in surveillance log:

Part of surveillance is being prepared for emergencies because we all know, sooner or later, you'll have an outbreak. Being prepared for an outbreak includes having your equipment and team ready to go. The investigation will go a lot smoother if you have the necessary equipment, documentation forms, and sampling kits ready for use when you need them. In this module we're going to talk about developing an equipment checklist and collecting the necessary supplies so you can be ready to go at a moment's notice.

Make sure that you have your aseptic sampling equipment all set to go. An important point to remember is some sterile equipment and supplies have a limited shelf life, so pay attention to expiration dates. Talk with the laboratory and develop replacement schedules for supplies. Restock supplies as they are used up or exceed the expiration date. Sample integrity can be questioned if improper equipment or out-of-date supplies are used to collect the sample.

By following proper sample collection procedures, you will ensure that any microbial contamination found during analysis did not come from the sampling equipment. After all, it's better to be part of the solution than part of the problem! Sampling guidelines will identify the equipment used for collecting various samples. You can use the guidelines for organizing the outbreak equipment.

Food and environmental samples are collected primarily to identify or verify the agent. In other words, if your suspect pathogens are, let's say, Campy or Salmonella or Shigella, your environmental samples are taken to prove your suspicions. There is an extensive list of forms and equipment in the course manual.

Having the necessary equipment to perform the inspection and aseptically collect samples is important. If improper equipment is used to collect samples, then sample integrity will be questioned and samples will be meaningless. The following is a suggested equipment checklist. The list should be modified depending on product and inspection requirements.

How your team is formed with the breakdown of roles and responsibilities will depend upon how it works best for you. For example, not every jurisdiction has an epidemiologist, but every state has one. Many teams will have to establish protocols in consultation with their state epidemiologist, lab, and nurses and others. Someone has to be the leader and central hub for the flow of information and open the lines of communication. The position of team leader is not necessarily a supervisory role. Typically the team leader may not have direct authority over the individuals comprising the team. The members should view the arrangement as a temporary job matrix during the outbreak where team members provide input, perform specialty functions, and report findings to the team leader. At the conclusion of the investigation, individuals return to their routine assignments.

The team atmosphere should be open for consensus building. Encourage constructive input and brainstorming to generate all those wild and sometimes crazy ideas and potential "could be's" or "what if" scenarios in trying to figure out what happened. Also with this process, the pros and cons of a decision can be weighed by the group before its implementation: a kind of check-and-balance process to reduce the chance of error or overlook possibilities that should be considered.

Between outbreaks, team meetings can be regularly scheduled to build rapport within the team and fine-tune procedures. After your next outbreak, evaluate the investigation as a team, keep what worked well, refine what didn't work, and conduct in-house training to enhance skills. Attending seminars keeps staff current and provides opportunities for networking with other agency personnel. It's also a good idea to start a library with information on foodborne illness and keep it current as new and emerging pathogens are reported. Once an outbreak occurs, the team should be prepared to go. Holding daily meetings to review findings and keep everyone up to date is important.

The foodborne illness investigation can be visualized as a three-legged stool. The investigation process has three components: EPI, laboratory, and environmental.

Without all three legs secure, the stool will fall over, toppling the team. Sometimes an outbreak will cross jurisdictional lines and require cooperation between investigation teams on the local, state, and federal levels. For example, in 1997, a Hepatitis-A outbreak in the Midwest involved strawberries served with the school lunch program. What started out as a local outbreak went multistate, and the investigation ended up including many state and federal agencies. When multiple agencies are working on the same outbreak, interagency communication becomes extremely important. Before the next outbreak, identify cooperating agencies, establish two-way communication and coordination procedures with these agencies, and include a contact person, emergency phone and fax numbers, and e-mail addresses.

The success of an investigation comes down to thorough work, notification protocols, and networking. Leaving a department out of the loop can hinder an investigation; besides, it's an exercise in courtesy and diplomacy. Would you rather be questioned on a pending outbreak by the press or informed by a counterpart? Everyone knows the answer to that question. Typically, large outbreaks initiated at the local level progress to the state level.

Depending on the circumstances such as interstate commerce, the size of the outbreak, and the agent involved, federal notification by the state may take place.

If the implicated or suspected food item is meat, poultry, or some egg products from a USDA-regulated plant, then the USDA would be notified through compliance officers in the field or directly into USDA's Consumer Surveillance Information System. Consumer complaints are also received on USDA's Meat and Poultry Hotline. So there are several channels through which information can flow.

Generally speaking, all other food products except domestic meat, poultry, and some egg products in interstate commerce fall under the FDA's jurisdiction. Typically, the state would contact its FDA District Office, and the district office would then notify the Division of Emergency and Investigational Operations in the FDA's headquarters. The district offices and DEIO also receive complaints from the public. If a complaint goes to the wrong agency, both FDA and USDA forward appropriate complaints back and forth and work together in investigations as necessary. Another example of teamwork!

In waterborne and environmental related outbreaks, the state may notify the Environmental Protection Agency (EPA). If there is a potential of contaminated drinking water coming into contact with foods in a processing plant or slaughterhouse, then both FDA and USDA would become involved. An example of interagency cooperation was the Cryptosporidium waterborne outbreak in Milwaukee, Wisconsin.

Depending on the size and nature of the outbreak, the pathogen involved or the need for certain lab tests, CDC can provide expert advice and assistance when requested. All federal agencies network and assist each other and the states, as appropriate, for foodborne illness investigations. The federal, state, and local agencies have developed a network called FORC-G, (pronounced "force-gee") "Foodborne Outbreak Response Coordination - Group." This outbreak evaluation group consists of the heads of various federal, state and local agencies to improve and coordinate the approach to multistate outbreaks. The FORC-G group reviews operations and networking after interstate outbreak investigations to see what worked well and what procedures could be improved or refined to develop standard operating procedures.

Good teamwork and strong communication are vital for resolving foodborne outbreaks. If your department does not already have a foodborne illness investigation team, then develop a plan. If you already have a team, then reevaluate procedures to make sure everything is the way it's supposed to be. Now that we've discussed planning for the investigation, let's move on to beginning the investigation.

Reports of illness can come into the health department piecemeal. When a complaint is received, try to get as much information as possible up front. Trying to establish the relevant facts is difficult enough when you have all the information, but trying to establish the facts when reports are incomplete is next to impossible. To standardize data collection, most departments use a general complaint form to record information.

The booklet entitled Procedures To Investigate Foodborne Illness, published by IAMFES, the International Association Of Milk, Food and Environmental Sanitarians, shows an example of a general illness complaint form.

Categories in illness complaint form include the following:

Start with basic identification information on the caller such as name, address, and phone number for home and work. Additional information may be needed from the caller as the investigation progresses. Plan a follow-up contact. If the caller wants to remain anonymous, tell them it may be difficult to keep in touch as the investigation continues. Attempt to identify additional cases. Ask if the caller knows of anyone else who is ill. If you suspect an outbreak, get information about the event, the exposure, the number affected. Ask for the names of both ill and well individuals who attended the event, and be sure to get telephone numbers of the persons affected. The goal of a case investigation is to get as much relevant information as possible. Timing is important. Each bit of information may lead to more cases or contacts. If you can't get all the information on the first contact, then get the basics on the caller, the illness, and the food operation.

Knowing demographic characteristics allows you to describe the people affected by an outbreak. Some demographic characteristics include, for example, age, gender, occupation, race, and education.

Take time to talk with the caller about their clinical signs and symptoms. Discuss initial impressions and find out how the caller learned about the illness, possible diagnosis and any medical assistance sought, specific symptoms, time of onset, duration of illness. Have they seen a doctor? If yes, who? What was the diagnosis? Were they hospitalized? What tests or samples were done? What were the results? Was any treatment provided? Also ask about chronic conditions, allergies, or medications that could mimic the symptoms of foodborne illness.

Find out if samples were submitted for analysis and if results are available. If samples have not been submitted, you may want to request a sample. A follow-up call to the physician can confirm the information you have, and you may be able to obtain the results of any pending tests.

Let the caller tell you what they believe happened. Then work to obtain the full history. When you're busy, you may not want to spend a lot of time on a single case that could be a dead-end. But you never know. The single report could be the tip of an iceberg. If a second case is identified and verified, you won't want to waste time re-interviewing the initial caller. So when a complaint is received, get as much information as possible up front. Remember, the better the initial information, the more likely your chances of early identification of an outbreak.

For a variety of reasons, when foodborne disease is suspected, information of foods consumed in the 72 hours prior to the onset of symptoms is requested. People have trouble remembering what they have eaten. It's not easy to obtain a food history, but the details contained in the food history are necessary for identifying potential exposures. The 72-hour time frame is used because recall gets very unreliable for foods consumed more than two to three days before an interview. Most people lead fairly routine lives, and a five- to seven-day food history could capture virtually everything they ever ate, increasing the difficulty in implicating a food vehicle. Also, the incubation period for many foodborne pathogens is 72-hours or less. Of course, there are exceptions. The incubation period for Hepatitis-A is 15 to 50 days; E. coli O157:H7 has an incubation of 3 to 8 days; and Trichinosis has an incubation range of 5 to 45 days.

The 72-hour food history should include all meals, snacks, and beverages, including water and ice eaten at commercial operations, as well as in the home. Find out where the foods came from. Were meals prepared on-site or catered, and did food come from outside sources? Ask if they noticed anything unusual about the foods, such as, off taste, texture, color, or odor. Find out if hot foods appeared fully cooked and were served hot and if cold foods were cold. If there is leftover food, give instructions for labeling and storing the food before it is taken to the lab for analysis.

When individuals can't recall the specific foods they ate, see if you can help them out: Ask about their food preferences, what they usually eat and where they have eaten lately. If you can't obtain a full 72-hour history, then try to obtain a facility-specific history. Ask about meals eaten out and then ask about meals eaten at home. A good wrap-up question could be "Is there any other information that may be relevant that you could provide?" Make sure that the caller knows that you may be calling them back if you need additional information. Also ask them to call you if they remember any other information about the event.

Many callers will attribute their illness to the last place they ate and may have already decided what food made them ill. We often assume that illnesses are foodborne in nature, but we need to keep an open mind to the idea that other factors may play a role. You don't want to put all your eggs in one basket. For example: there was a Salmonellosis outbreak, but the investigation did not reveal any food associations. The EPI evidence showed the common item among the cases was marijuana. The lab found the same type of Salmonella in both the marijuana and cases. So remember, the source of illness may not be food-related. Ask the caller if they have done anything unique or different recently. Ask about domestic or international travel and whether they've had contact with ill persons or with animals.

When an illness complaint is received, examine the symptoms, onset times, and the 72-hour food history. Use the food history to see what was consumed prior to onset of symptoms. Try to match up foods, incubation times, and symptoms that could be associated with the possible pathogens. Of course, you should go through this exercise when the caller reports a self-diagnosis and the pathogen and source are unknown. But even if you do have a medical diagnosis with a known pathogen and possible source, play "devil's advocate" and review the data for consistency. (Editor's comment: Make a distinction between an investigation that supports the first plausible explanation you think of and a legitimate investigation. It is bad form to identify a food vehicle and cause of illness based on incubation period and symptoms and then to design an investigation to prove the association.)

For example, say the illness report concerns a person complaining of tingling and burning sensations around the mouth, facial flushing, dizziness, headache, and vomiting about one hour after consuming a large tuna steak. Is this complaint plausible? A review of reference materials would indicate they are plausible. These symptoms and onset time suggests a toxin or chemical poisoning. Specifically, the symptoms and onset are similar to that of scombroid-type poisoning, where histidine is converted to histamine. That type of poisoning is most often associated with fish in the Scombroid family, of which tuna is a member. As a general rule, symptoms of chemical and toxin poisonings occur within one hour of ingestion. But keep in mind that some toxins and poisons can have a longer incubation period.

Here's another situation. Say the illness report concerns a person complaining of only nausea and vomiting. You have a 72-hour food history. What part of the food history would you concentrate on? Look at what was consumed in the six hours prior to onset. Pathogens such as Staphylococcus aureus or the emetic form of Bacillus cereus would be good candidates for this time frame.

Say the illness report concerns a person complaining of only cramps and diarrhea. What incubation times would you expect? In general, for cramps and diarrhea only, look at what was consumed between six and 20 hours prior to symptom onset. Possible suspects may be Clostridium perfringens or the diarrhetic form of Bacillus cereus. If fever and diarrhea are symptoms, consider a possible infection. Illness with fever generally indicates an infection rather than intoxication.

Diarrhea and Fever, 12 to 72 Hours After Eating

Pathogens that infect typically require a longer incubation time to allow the organism to multiply in the body. Look at foods consumed about 12 to 72 hours prior to onset. Keep in mind that these are general rules. They give us a place to start; they're not absolutes.

Decide whether further investigation of a complaint is warranted.

There's a lot to consider here, including the number of cases, the agent, the severity of illness, and how long ago the illness occurred, not to mention the department's policies, investigation criteria, and resources. Determining whether you actually have a foodborne illness outbreak and deciding when to initiate action can be challenging. If you' re investigating a serious illness, the investigation could begin based on a single report. But for a less-serious illness, follow-up may be postponed until additional reports are received. Analysis and interpretation of surveillance data can help identify potential problems. For example, if a decision to follow up is made based on analysis of log entries, the investigator should determine whether an increase has actually occurred and whether the increase can be linked to an obvious common exposure. Evaluate the data to ensure that there is a plausible basis for the potential association and verify the diagnosis. It may be necessary to find more cases, collect more data.

Case report data from the surveillance log can be organized with respect to time, place, and person to find possible associations among cases with the same diagnosis or similar symptoms. For example, when you review reports, trust that "deja-vu" feeling. If individuals experience similar symptoms and onset times, they may have a common association with a particular place or event.

If there appears to be a meal in common that is associated with the illnesses, then the time between ingestion of the agent and the onset time provides the incubation period. The incubation period and symptoms are helpful clues in determining which diseases "could be" or "definitely are not" involved with the illness. Incubation times will vary among cases. That's why incubation periods are given over a range of time. The time frame the investigation team is concerned with is related to the suspected agent's incubation time. For example, let's say several reports of confirmed Salmonellosis were received over the last week and onsets for these cases occurred within 60 hours of each other. It is possible that these cases could share a common source of exposure, because all the onsets fall within one incubation period for Salmonella, typically 12 to 72 hours.

When illness complaints are organized by time, the data may be displayed on a graph. Graphs show visually the relative size and trend of the problem and can be constructed to show daily or seasonal trends or time intervals that span several years. Graphing the data a few different ways will help you decide on the most appropriate and revealing time interval to use. Graphs are very useful for showing past trends and for predicting future trends. They can also provide insights into what may have caused the problem.

Let's say several reports for a cluster of suspected Shigella illnesses were received, and the diarrhea onset times for the ill college students were graphed. The illnesses occurred Saturday, Sunday, and Monday. The typical incubation period for Shigella is between 24 and 72 hours. It is possible that the students shared a common source of exposure, as all of the cases fall within one incubation period. So if they ate any meals together Friday, one or more of the meals could be associated with the illnesses.

Reviewing illness reports with the same diagnosis or similar symptoms and onset times may show a common association with a particular place or event. Association by place refers to the attributes or factors that describe the environment in which the disease occurred, such as geographic location, county, town, food-service facility, business, residence, social event, purchasing food from the same place, or consumption of a specific brand of food. The data can be mapped, graphed, or tabulated to help you gain insight into the geographic extent of the problem.

Analyzing data by place may help you identify the most likely pathogen and may give you insight as to how illness could be spread. If the illness is associated with a particular place, then it's a safe place to start investigating to see whether the risk factors of illness are or were present.

If the students ate a meal together, the place associated with the illness can be determined as well. However, the only information the team may know about the source of illness is that cases are currently occurring over a broad locality such as the state, county, township, or college campus.

Person associations refers the characteristics of ill individuals with similar symptoms or diagnoses of those who were exposed to the agent or suspect agent. Some categories used to group data by person are age group, gender, occupation, immune status, affiliations or group membership, extracurricular activities, or any other unique characteristic. It turns out that personal characteristics sometimes can be used to predict who is at greatest risk of becoming ill from a particular pathogen. You may have to group your data several different ways before deciding on the most appropriate and revealing person-category to use to get a better picture of the situation.

There are times when an outbreak may not be limited to a particular group of people. For example, the students who reported diarrhea may have eaten off campus, and possibly there are other nonstudents out there who may have also become ill. Once the associations for time, place, and person are known, control and prevention methods can be implemented.

Refer the example of the surveillance log. Remember, we had three entries of Hepatitis-A in two weeks with all having some association with the Hillside day-care center. Here, the association by place is easily identified, but sometimes the cases must be interviewed before time, place, and person associations can be identified. The question, "What is the correlation between Hepatitis-A and the day care?" can be answered only with follow-up.

Definition of Foodborne Illness Outbreak

This is the part of the investigation where you need to dust off your people skills. The better your interviewing skills are, the more you'll learn about an event. And the more accurate the information you collect, the better you'll be able to evaluate your hypothesis. An interview is a directed, definite, and purposeful "conversation" that involves more than just the words that are spoken. You must interpret what people say, the way they say it, their gestures, posture, facial expressions, eye contact, and other nonverbal cues. A good interviewer is also a good listener. Before you pick up the phone or walk into an establishment to begin an interview, think about your objectives and what you want to accomplish. Why are you doing this interview? Whom will you be interviewing? What is the person's age, gender, background, and occupation? Are they potentially responsible for the outbreak, angry about becoming ill, or confused about their involvement in an investigation of illness that appears not to involve them? What other characteristics will help you gather information?

For the most part, you'll need to gather information from ills and wells, and those involved with food preparation. One of the tough things about gathering information is it's not a nine-to-five business. You'll probably be working around a lot of different schedules: the owner of an establishment, cooks, kitchen crew, wait staff, people at the function, people who ate the food and those who didn't as well. Trying to find patrons is often very difficult. You may have to call them at home, in the evening or on the weekend. Given the many priorities you must satisfy, schedule interviews for times when they are likely to yield the best results. Whenever possible, make appointments and be punctual. Make it convenient for the person you're interviewing, so they're comfortable and more willing to cooperate. Remember, unlike a facility operator, the public is not obligated to communicate with you. If the time you've set aside is not convenient, try to reschedule.

Your mission is to gather information on the source of the illness and means of transmission. You want the cooperation of the people involved, so begin interviews by letting people know that you're a professional, and your job is to gather information. Remember that the situation may be threatening to them. Conducting the interview in a sensitive manner will help people feel comfortable. If you're not sympathetic to the feelings of others, they may react in a way that limits the free flow of information. Giving someone "the third degree" may not give you the results you need. And privacy is important too. Think about the person's feelings when you select your interview site. You want to avoid distractions, and some of the information may not be for everyone's ears. Most people would prefer not to tell the world about the diarrhea they had, or discuss other symptoms "in public." Perhaps, if you interview a patron or an employee in private, you'll get a more accurate picture of what happened.

Begin your interview with something like "There's been reports of possible foodborne illness and we're investigating to find out what happened." Right away people will want to know more about the investigation. Tell them you're just collecting information right now, talking to those who allegedly were exposed, and those who were involved in preparing the food. Your behavior will show your objectivity. If someone is hesitant about giving you information, tell them the purpose of the interview as clearly as you can and encourage their cooperation.

Many times both patrons and facility employees will also try to solve the case. Remember reputations may be at stake. A restaurant owner may be afraid he was at fault. A victim may want to know how you got their name. Go ahead and tell them. Be as open and honest as possible without divulging confidential information. As you begin the interview, establish rapport with the person. Try to establish mutual confidence and an understanding atmosphere so that the individual won't be afraid to answer openly. Greet the person with respect and introduce yourself, including your name, position, and department. Take time to explain the purpose of the interview and how it relates to the person's needs. If you feel at ease, then they're more likely to feel that way too. Allow a little time for the person to become accustomed to you and the situation. Remember to speak in terms they can understand. If the person can't remember something, ask them to think what else they were doing during the time period, to help them jog their memory.

Questioning is an art. Direct your questions so your goals for the interview will be met. Open-ended, non-directed questions encourage a person to use their own words to describe their experience. This approach, however, may be difficult, especially when you want the answer quickly. So ease into the interview. Start with a few directed questions like: "What is your name? Your address? Occupation?." Then go into open-ended questions. Here are several examples of open-ended questions, followed with close-ended questions. To verify the exposure, ask: "What were the foods that you ate?" Then follow up with: "Did you consume any other foods or drinks?" You might also ask: "After the event, how did you feel? Were you ill?" Let the person answer in their own words, then follow up with directed questions to fill in the gaps. A drawback with using only open-ended questions is that they tend to be more difficult to analyze statistically than yes-no questions. So fill in any voids with direct questions.

Closed-ended questions usually have just one answer option. Some examples are: "How many stools did you have in the last 24 hours? Was the diarrhea watery? What was your temperature? Did you use a thermometer?" If you're in too much of a hurry to complete the interview, you may be tempted to be too direct. Avoid leading statements like these: "I have a list of all the items that were served there that day. I'm going to read this list to you, and you tell me whether you ate these foods: yes or no.". When you interview a person in this manner, they are not likely to volunteer information that could be important.

Closed-ended questions can be restrictive and leading. For the most accurate data, let the interviewee tell their story. Think before you ask a question! Speak at the level of the person you are interviewing. Make eye contact when you're talking. If a response does not address what you're looking for, find another way to ask the question. Repeat the question and answer to be sure you both agree on what was said. Keep in mind that people may not admit to the truth for a variety of reasons and they may be afraid of losing their jobs. If they believe you will protect their identity, they are more likely to be truthful.

We all have a tendency to be so focused that sometimes we don't realize where other people are coming from, how they are going to react, or what's at stake for them. It helps to ask questions while being sensitive to "signals" from your interviewee. Usually, people want to cooperate. They're willing to work with you. But every once in a while someone may be trying to work against you. Learn how to handle these situations as part of your preparation. If you have not had conflict resolution training, make that a part of your professional development plan.

You should know your own style of interviewing. We all have different ways of doing things. Some people come on strong. Others are more passive. Also be aware of preconceived opinions you may be bringing to your interviews. For example, if you begin an interview with the opinion that this outbreak is due to chicken and all your questions are about chicken, then your own bias affects your result.

When you begin, explain the purpose of your interview in terms of your goals. (Editor's note: Be careful to avoid leading people to provide answers they believe you want to hear, rather than accurate and truthful answers.) Without a common point of departure, responses may not be as helpful. Be honest about why you're conducting an investigation. Ask for their help. Tell them you really need their input to make this a complete investigation. How the questions are asked and the order of questions can also affect the answers you get. And watch out for something called recall bias. Memories are not perfect. People may say they ate something that they didn't because they always eat that item. Or they may flat out forget eating a particular food item. Recall bias is almost always a factor in an outbreak investigation.

Be sure you don't influence your interview by the way you ask questions. Avoid leading questions or "double barrel" questions such as, "Did you prepare the chicken on this cutting board?" This question has two parts: Did you prepare the chicken? And did you prepare it on this cutting board? It's a bit difficult for someone to effectively answer two questions with one yes or no response. Another example of a double barrel question is: "Did you cook it and measure the temperature?" That question assumes that they indeed take product temperatures. Instead of assuming, ask how they know when it's done. Remember that the goal is to get the facts as they occurred without implying a particular response. Also use appropriate language, format, and terms for the person with whom you're talking. You may be talking to physicians at some point, cases and controls at another time, and later to kitchen workers.

Your rate of speech and tone may affect the interview. Remember our little role-play at the beginning of this section: "Were you there or not? Did you eat there or not? Did you eat this? The others said you did! What do you mean, you don't remember?" You have to let the interviewee tell their story. As we said before, avoid the third degree. Be alert to verbal and nonverbal feedback. Watch for body language, eye contact, signs of stress in the voice, and body posture. If signs indicate stress or uneasiness, adjust your approach and try to regain a comfort level with the person.

Two-way communication is important. Remember the last time you said something that seemed perfectly clear to you, but the answer you got back was totally off the mark? That's because the other person heard it and processed it differently than you intended. The connection isn't made because each of you is coming from different perspectives. This model shows the many roadblocks that can get in the way of good communication. Each of us filters or encodes-decodes information based on our personality, culture, and experience. As the verbal and nonverbal communication travels through all the noise and distractions, you decode what was said and process the message through your filter and encode a response back. To avoid misunderstandings, recap the conversation. Rephrase what you think they said or repeat what you heard. Say something like: "This is what I understood you to say. Am I right? Correct me if I'm wrong. I want to be accurate. Your information is important."

As you conclude the interview with an ill or well person or food employee, you'll want to review your notes or questionnaire to make sure you have what you need. If for some reason you can't complete the interview, make another appointment and leave your card. Provide your name and phone number and allow people to call you from the security of their own home. Invite the person to contact you if they think of anything else and be sure to thank them for their time.

The following tips for good listening are from: Eastwood Atwater, 1981. I Hear You. Prentice Hall, Englewood Cliff, NJ. In: Dealing with the Press: Newspaper, TV, and Radio. FDA/STB, Handout 12-1-95

Do: Become aware of your own listening habits.

What are your strong points? What are your faults? Do you judge people too quickly? Do you interrupt too often? Better awareness of your listening habits is the first stage in changing them.

Do: Share responsibility for communication.

It takes two to communicate--one to talk and one to listen--with each person alternating as the listener. When you are unclear about what a speaker is saying, it is your responsibility to let the speaker know this. You can ask for clarification or actively repeat what you heard. Then ask to be corrected.

Do: Be physically attentive.

Face the speaker. Maintain appropriate eye contact. Your posture and gestures should show that you are listening. Sit or stand at a comfortable distance to put you and the speaker at ease. Someone who is speaking wants an attentive, animated listener, not a stone wall.

Do: Concentrate on what the speaker is saying.

Don't let your thoughts wander. A physical or verbal response will probably help you concentrate on what the speaker is saying. Listen for the total meaning, feelings as well as content.

Do: Observe the speaker's nonverbal signals.

Watch facial expressions and eye contact with you as the listener. Listen to the speaker's tone of voice and rate of speech. Does the speaker's body language reinforce or contradict what they just said?

Do: Be accepting of the speaker.

An accepting attitude on the listener's part creates a supportive atmosphere for communication. The more they feel accepted, the more they let down their guard and express what they really want to say. Be understanding. Use active listening skills to discover how other people feel, and what they are really trying to say.

Do: Listen to yourself.

If you can recognize your own feelings in reaction to another's message, express those feelings. This "clears the air" and helps you to listen better. And finally:

Do: Take appropriate action.

People speak in order to get something tangible done--to obtain information, to change an opinion, to get something done. You'll show that you're listening by how you respond. Actions speak louder than words!

Don't confuse "not talking" with listening.

People who remain silent aren't necessarily listening. They may be preoccupied with their own thoughts. On the other hand, some people can talk a lot and still listen well.

Don't "fake" listening.

Whenever you try to "fake" listening, your disinterest inevitably shows in your facial expression or body language. Staring, yawning and checking one's watch can be dead giveaways.

Don't interrupt needlessly.

People in positions of power tend to interrupt more often than those not in power... sometimes without realizing it. If you must interrupt someone, try to follow with a "retrieval"--help the speaker to re-establish his or her train of thought.

Don't pass judgment too quickly.

Judgmental remarks invariably put others on the defensive. They serve as barriers to effective communication. After they tell you it tasted and smelled bad and they ate it anyway, don't let the look of astonishment show on your face.

Don't make arguing an "ego-trip."

Even if you argue only "mentally" with what the speaker is saying, you may stop listening and look forward to your turn to talk. If you start to argue verbally, you will be so preoccupied with justifying your own views that you won't hear the other's viewpoint. When you honestly disagree, listen carefully in order to understand what you are disagreeing with. Then state your point-of-view.

Don't ask too many questions.

Too many questions have a way of shifting control of the conversation to the listener, putting the speaker on the defensive.

Don't tell a speaker "I know exactly how you feel."

This remark serves more to justify your own efforts than to convince someone you are really listening. It is difficult to know just how another person feels. Such a remark is likely to distract the speaker from further efforts at self-expression, and may cast doubt on your own listening ability. It is more effective to demonstrate you have heard with an observation like "I sense that you are feeling disappointed," or "I get the impression you are angry about this." Also don't give advice unless it is requested.

Don't overreact to emotional words.

Be careful not to get so caught up in a speaker's feelings that you miss the content of the message. And also realize your own feelings can block hearing something you really need to hear.

We use interviewing and listening skills to collect data about the outbreak, to update the working case definition, to generate a hypothesis and to test the hypothesis. (Editor's note: Different interviewing techniques are appropriate when getting patient information from a physician, when getting hypothesis-generating information from ill persons and food workers, and when getting hypothesis testing information from cases and controls. An unskilled interviewer can eliminate the possibility of correctly understanding the causes of the outbreak.)

A hypothesis is a statement that can be tested and refuted. The initial hypothesis is developed to help organize our thinking about possible time-place-person associations, and to provide direction for the investigation. As the outbreak progresses the hypothesis will be refined periodically to incorporate more of what we have learned about the agent, source, means of transmission, and how the illnesses occurred. The hypothesis will undergo testing to determine how our beliefs and assumptions hold up under the scrutiny of data analysis. If analytic methods testing fail to support the current hypothesis, then generate a new hypothesis. You may need to narrow the focus of your investigation to ask more specific questions on exposure history. Also consider new vehicles and modes of transmission.

Previously we learned that the case definition includes time-place-person and clinical criteria that an individual must meet to be considered ill. As the investigation progresses, the case definition will usually be revised and victims will be re-evaluated to determine if they are still considered within the set of outbreak related ill persons. If ill individuals meet the current case definition, they are outbreak related. Re-evaluation of the case definition should be done prior to organizing data for analysis, so the findings will only contain outbreak related cases and will not be mixed with unrelated illnesses.

Standard case definitions for reporting specific illnesses have been developed by CDC, the Council of State and Territorial Epidemiologists, and the medical community, but the wording of these reporting definitions may not be the best for use during an investigation. A standard case definition focuses on the ailment itself and indicates how certain we are that the subject actually has the illness under study. Time-place-person associations are not included in the standard case definition. The definition only states how sure we are that an individual has a certain disease based on clinical criteria or lab results. The degree of certainty is classified as confirmed, presumptive, and suspect. Thus, every case is diagnosed consistently and without bias. When you have laboratory confirmation of the agent, then the case definition can be stated to reflect how sure you are that each person is a victim of a particular illness. A case can progress from a suspect case to a presumptive case to a confirmed case as laboratory work is initiated and completed.

These classifications are developed to assist us in the comparison of case data. Analysis can be conducted on confirmed cases, on confirmed and probable cases, or on all cases. Comparing results of analyses of confirmed cases with those of suspected cases may indicate the likelihood that suspects are truly part of the outbreak.

A confirmed case has the clinical signs and symptoms of the illness, and there is laboratory confirmation of the agent causing illness.

In general, a probable or presumptive case is defined as a person with clinical signs and symptoms of the illness for whom there is some laboratory evidence that is suggestive of the illness. For some pathogens, the presumptive classification is not relevant because the particular lab test does not yield suggestive results. Lab tests may only confirm the presence or absence of the agent.

A suspect case has the clinical signs and symptoms of the illness in question. It looks like they have the illness, but there is no laboratory confirmation to support a diagnosis.

Now let's talk about case finding, which is the process of locating additional exposed people. Those exposed may be either "ill" or "well." One way to conduct case findings is to contact known victims and ask whom they ate with and if they know the names of others who attended the event. Another way is to obtain a banquet or a reservation list. For facilities that accept credit cards, it may be possible to get the names of the patrons from the credit card receipts. Depending on the circumstances, it may be helpful to contact other health agencies, emergency rooms, and local medical care providers to find additional ills. You can also review surveillance logs for reports that may be similar. New reports may be part of the original outbreak or may be the result of secondary spread.

Don't just focus on the ills. Without information on the wells there can't be any meaningful data analysis to characterize the event. The foods consumed by both the ill and well provide a comparison for possible foods and meals in common. Usually, not everyone that eats the implicated food becomes ill, and others who report that they did not eat that food may become ill. Also people can have medical conditions that resemble foodborne illness or they can have a sympathetic response. For example, someone vomiting may cause others nearby to vomit.

In the early stages of an outbreak, reports may be recorded on a general complaint form. Once enough is known about an outbreak, such as the facility, symptoms, exposure period, and so on, then an outbreak-specific questionnaire can be developed by the outbreak investigation team to help with case finding and organization of information. The outbreak specific questionnaire is customized to focus on the circumstances of a particular event. Many of the questions on this specific questionnaire for the ill and well individuals are direct as compared to open-ended questions you'd be asking employees about food preparation procedures. Getting information by asking detailed questions may be the only way to get accurate information. Once these questions are answered, you can take appropriate public health action.

(Editor's note: This section on collecting information using a questionnaire is related to, but is different from, hypothesis-generation interviews. Once enough is known about an outbreak, such as the facility, symptoms, exposure period, and so on, then an outbreak-specific questionnaire can be developed by the outbreak investigation team to help with case finding and organization of information.)

The usual way to obtain information is to ask a simple, direct question. The direct approach works, but experience shows the question-answer process can become complicated when information about a group is required. Complications result from nonresponse to a question, deceptive answers, lack of knowledge, or refusal to discuss specific subjects. When it comes to conducting a survey, poor memory is one of the biggest problems, and the inability to assign causality is the biggest limitation.

Common Source Outbreaks

Epidemic curves can help show the type of outbreak that is occurring, such as a common source. Many foodborne outbreaks are common source outbreaks. The victims visited a common facility or event to ingest the agent. The source could be a food item that's distributed over multiple states and creates a multitude of outbreaks with a common source. Or, if the contaminated food is only served once at the facility or event and illness occurs, then the common source outbreak is also a single exposure event. If the vehicle was served multiple times, then it would be referred to as a common source, multiple exposure outbreak. This graph is an example of a common source Hepatitis-A outbreak. The peak of the outbreak should occur at approximately the mean or average incubation period for the agent, or in this example for Hepatitis-A, about 28 to 30 days after exposure.

Epidemic curves can help show the type of outbreak that is occurring, such as a common source. Many foodborne outbreaks are common source outbreaks. The victims visited a common facility or event to ingest the agent. The source could be a food item that's distributed over multiple states and creates a multitude of outbreaks with a common source. Or, if the contaminated food is only served once at the facility or event and illness occurs, then the common source outbreak is also a single exposure event. If the vehicle was served multiple times, then it would be referred to as a common source, multiple exposure outbreak. This graph is an example of a common source Hepatitis-A outbreak. The peak of the outbreak should occur at approximately the mean or average incubation period for the agent, or in this example for Hepatitis-A, about 28 to 30 days after exposure.

Propagated Outbreaks

This combination is referred to as a mixed outbreak. From a foodborne perspective, mixed outbreaks originate from pathogens that can be transmitted via food and continue spread as person-to-person. This epidemic curve is a mixed outbreak of shigellosis revolving around a music festival. The shaded portions of the histogram represent the festival staff members and the nonshaded areas represent festival attendees. Remember, the incubation period for Shigella is usually around 24 to 72 hours. The first case occurred on day 1 with additional cases occurring prior to the festival. Approximately two days after the festival, the 16th day of the outbreak, the peak of the outbreak occurred with about 100 cases that day.

Food history from the line list is then transferred to the attack rate table. An attack rate table is used to compare those who ate a specific food with those who did not eat the food. In general people who ate a contaminated food are the ones who become ill and people that didn't eat the contaminated food don't get sick. That's why it's important to interview people who didn't get sick. Consumption and illness data from the line-list is used to calculate the attack rates to determine which foods are most likely to be associated with an illness. A high attack rate among persons who ate a specific food suggests that the food is associated with the illness. A low attack rate among persons who ate the food suggests the food is not associated with the illness.

The difference in attack rates is called the risk difference. The risk difference is the percent ill for those who ate a specific food minus the percent ill of those who did not eat the food. Here is an example:

Here we selected bread from an attack rate table to show you. 74 percent of people eating the bread became ill compared to 56 percent of the people becoming ill that did not the bread. So the risk difference is 74 minus 56 or 18 percent. Usually the risk difference is large for the contaminated food and small for other foods, so, often the largest risk difference identifies the contaminated food.

If the source of the outbreak has not been determined, meal-specific, or facility-specific attack rate tables can be used to focus on possible sources. The attack rate table can be used to test the strength of the association between the suspected food and the illness. Measures of association will be discussed in the drama and in future lecture modules.

Throughout the investigation, control measures should be implemented as appropriate to prevent additional cases. Corrective action may be needed immediately such as embargoing strongly suspected foods, and excluding or restricting ill workers. Even if the agent is not known, personal hygienic practices and critical food handling procedures can be evaluated and corrected as needed. The source could be the foodservice personnel, equipment, preparation procedures, a processor, grower, harvester, flock, farm, growing waters, and so on. Before initiating any control measures, contemplate the effectiveness, timeliness, public acceptance, costs, available resources, personnel requirements, and ramifications of any actions taken.

If humans are potential reservoirs for the suspected agent, and employees are suspected as a source, the appropriate public health response should be planned with the help of a medical epidemiologist Once the agent is confirmed, specific control measures for the agent can be administered. For example, administering IG to those exposed to hepatitis-A within the last two weeks can help prevent additional cases. Also, after the control measures have been implemented, a mechanism must be provided to assess the effectiveness of control measures and adjust them as required.

An individual exposed to the original source would be referred to as a primary contact and a primary contact who becomes ill can be referred to as a primary case. If the pathogen is transmissible person-to-person, like Shigella, and another individual is exposed to a primary case through a plausible means of transmission, then this second individual can be referred to as a secondary contact. If the secondary contact becomes ill, then they can be referred to as a secondary case. Onset for secondary cases is usually one or more incubation periods after initial exposure of primary cases. If a contact is not ill during the interview, then that persons' risk of infection must be assessed based on the individual's susceptibility or immunity to the agent. If a case is susceptible but not ill, consider obtaining specimens to determine whether the person is infected. This would be particularly good where an asymptotic case may occur or where there is a long incubation period.

As we all know, the media can become involved in the investigation. Sometimes it is the press contacting the department for information about an outbreak. Other times it is the department notifying the media to get a message out to the public to aid in an investigation or warn the public of a problem. Either way you want the public to be involved, reasonable, and solution-oriented. You will earn or lose public trust by the way you handle communication. In the example we've been discussing, several hundred customers may have been exposed to the Hepatitis-A virus at the Sandwich Shop. Transmission of the virus can be controlled if the customers who ate at the shop between March 15th and March 31st were informed about the possible exposure and offered immunoglobulin. (Editor's note: Often, control of the situation is not this straight forward.) Alerting the public may prevent further spread of the virus.

It depends on the situation. As a government employee, remember that you work for the public. Accept the public as a legitimate partner. You set the pace to resolve the problem and guarantee accurate information, if you're the source. Inform the public as soon as you can. A delayed information release may result in distortions of the facts or even attacks by the press. People often respond to stress and uncertainty with anger or they may overestimate the risk. So your credibility and competence will be judged by the way you communicate to the public.

Avoid complex scientific arguments and don't try to deny an outbreak's importance. You can't just present numbers, either. It's your job to put those numbers into perspective, with comparisons that people can understand. Risk is seen in different ways, depending on whether someone can control the source of the trouble, how much they understand the risk, and the consequences that might result.

Be accessible to the press, and meet the needs of the news media if you can. Listen to the public's concerns. Don't make assumptions about what people think. Any perceived problem has to be resolved, whether a risk really exists or not. Evaluate the information you have, and then try to put yourself in the place of your audience. Educate your reporters, and tailor your communication to the specific groups and media involved. Be clear about the risks. Evaluate the information you have and be honest if your data is uncertain. Any illness, injury, or death is a tragedy, and this should be reflected in your communications to the public.

As soon as you begin investigating an outbreak, notify your public affairs office. Find out who the best person is to answer questions. A spokesperson is a leader who gets out the information, not just someone who answers questions. The public really wants to know: "Is it safe?" Get the details ready and plan for what people will ask. When you speak to the news media, you're already communicating with the public. So, avoid complex language and jargon. Instead, use concrete images that make the data come alive. Communicate on a personal level with your audience and use comparisons that put risk in perspective. Avoid non-food comparisons because comparing foodborne illness with the risk of physical trauma or other unrelated hazards does not provide useful information. Above all, keep it simple! Whenever possible, it may be best to provide a carefully prepared press release with the tangible and accurate facts to avoid speculation and careless interview statements.

The media is looking for short "sound bites", so avoid long explanations that may be taken out of context. Keep your interview to 30 minutes or less. Be careful what you say around microphones and tape recorders. There is no such thing as "off the record". Any conversation or stray comment is fair game for the media.

You can use a three-prong approach: explain the subject, state the facts, and give the context. "Boil down" the subject, the facts, and the context to 30 seconds or less. That way, the media will get their "sound bite" and you'll get the satisfaction of knowing your message is clear.

Sometimes reporters are more interested in politics than risk. They are looking for a story angle: controversy, emotions, and visuals for television or print news. When public health and the media have very different agendas, the communication gets complicated. Don't be naïve. Give the media your spin, but be careful not to become the story. An unexpected question may catch you off-guard. So think before you answer a question, and if you don't know the answer or aren't sure, say so. If it's not in your field of expertise or the information just isn't available, admit it. Only promise something if you can do it.

Listen to each question, and really focus on what is being asked. If you're not ready to respond to questions, say so, but also say when you will have some answers. Then take the time to really think through those answers in advance.

If you don't understand a question, ask the reporter to rephrase it. Never try to bluff an answer, even if a reporter expects you to know everything. Don't answer "No comment" if you can. This looks to the viewer like you're saying: "I am hiding what I know." Even after the interview is over, if you're not sure you got the facts right, contact the reporter again to clarify.

Make sure you work with all local, state and federal agencies who are investigating the same foodborne outbreak. It is critical to coordinate the release of information to the media. At the outset of a foodborne illness outbreak that goes across state lines, all agencies should get together and discuss a strategy for informing the public immediately.

This is the only way to avoid inaccurate, confusing, or conflicting messages. The agency that issues the press release is considered the lead for media contacts. If more than one agency issues information, determine the main message and keep it consistent.

If more than one agency will be speaking to reporters, you'll need phone numbers, fax numbers, and beeper numbers for each contact including after-hours numbers.

Decide up front what information is going to be included in the press release and what to tell consumers.

Before you issue a press release, share a draft with any other agencies involved, so that all errors can be corrected. After it's issued, keep in contact with each agency to discuss media follow-up.

On completion of this module, participants will be able to:

At this point in the investigation, you're well under way, having gone over surveillance data, determined number of cases, assembled your cross-disciplinary team, and met to discuss possibilities and theories. Now the focus shifts to the food hazard review. This is the "grunt work", the sometimes down and dirty stuff, but it is the key activity for health and food regulatory personnel in outbreak investigations. Within the investigational community, this step has been informally named the food prep review, and you'll see why in just a moment. In any event, it is the environmental leg of our three-legged stool and focuses on the food, methods of preparation, and possible contributing factors at the suspect food-service facility or plant.

An investigation driven by the possibility of a foodborne illness outbreak is not a routine inspection to identify regulatory violations. The food prep review is done on each suspect food or menu item that has been consumed. It concentrates on possible foods for source, means, or mode of possible contamination. You consider the ways a microbial contaminate could survive or grow. You identify actual and potential hazards associated with each ingredient, the processes used, and the way the product was transported as compared to its final use.

Let's be clear about the nature of any food preparation: there are potential hazards associated with each ingredient, the operations used, as well as storage and handling. The hazard analysis of an "epi" investigation considers those potential hazards including: the source of ingredients, the recipes, the processing equipment, food storage, preparations from the back door through the front door and beyond if necessary, time in processes and storage, and expertise and attitudes of the people involved. You're trying to recreate what happened in the past that may have resulted in foodborne illness. You focus on determining what DID happen; not what we think happened. Thus, all foods suspected during an investigation are subject to a hazard analysis.

The food prep review should identify the vehicle that caused the outbreak. In a perfect world, it would be done without knowledge of what food was implicated by the "epi" thought process. It would independently come to the same conclusion. When performing the food prep review, the investigator will go in with some prior information and initial hypotheses. It always helps to have a point of departure; a base of knowledge or ideas. Think of it as having a bag filled with possible hypotheses, and then as you go through the investigation you eliminate unsubstantiated ideas and theories based on what you can prove. If you've done it right, whatever's left in the bag at the end of the investigation is more than likely the vehicle and contributing factors. This may seem as if you're traveling the Catskills with an atlas of the world! But you must consider all possible factors, instead of being tempted to pre-judge the results.

Unfortunately, some hazard analyses are preconceived investigations. What if victims have Salmonella enteriditis infection and the investigator sets out to prove that eggs caused it? Or if an investigator hears E. coli 0157:H7, then does an investigation to prove that hamburger caused it. That's just not good science! We find it best to avoid a biased approach.

A food prep review is quite distinct from other kinds of activities that concern environmental health, in that it looks back. You're trying to identify actual and probable hazards from what happened hours or days or weeks before to the final use of each food substance. After interviewing people, making observations, and collecting samples, you come to the best possible conclusion as to how the vehicle was produced, transported, prepared and served. Then you report the findings and conclusions as to what food was the vehicle and what were the contributing factors. Sometimes there's a tendency to report speculation as findings and conclusions. You'll want to avoid that and strive to report what you actually found. That doesn't necessarily mean you completely disregard your "gut" feelings. There's a place for those personal observations and we'll talk about the comments portion of your final report a bit later.

One type of inspection that could be confused with the food prep review is the Hazard Analysis Critical Control Point system or HACCP. In HACCP, you're trying to prevent problems from happening in the future. So the process includes a review of the procedures going on today, in order to develop a system that will prevent people from becoming ill. A normal HACCP review does not usually result from an outbreak investigation. But if the hazard analysis during an epi investigation reveals significant food safety breakdowns, one of the recommended control measures may be a HACCP system for the facility in the future. Because HACCP is a preventive program, it's really quite different from a food prep review, which is after the fact.

When an investigation implicates a food and suggests that contamination may have occurred sometime prior to its preparation in the facility, it's likely that a trace back investigation would occur. But first, it's very important that you've been able to establish that other likely causes of this kind of illness did not occur at the point of service. Using our previous example of Salmonella enteriditis, shell eggs have been identified as a source of that infection. But food workers and improperly raised poultry could also be the culprit. If during the investigation you systematically and scientifically eliminate other possibilities, leaving only the eggs, then a trace back investigation is warranted to identify the source flock.

What do you do with the information collected in an outbreak hazard analysis?

Many regulators view the food preparation review in the legal context, in light of violations. Although it's true that we can find violations, the hazard analysis findings and conclusions have other significant meanings. By identifying the contributing factors, we make sure that the establishment does not repeat these problems in the future. Information from many investigations can be summarized to look for trends or most common contributing factors over time. Determine your priorities for food safety programs and the education programs provided. The information collected from one investigation can have significance that goes far beyond that investigation when pooled with the findings from previous investigations.

Contributing factors include contamination, the survival of pathogens, the persistence of toxins, and multiplication of pathogens. For bacterial hazards, two or more of these factors usually occur sequentially before an outbreak occurs. Here's a list of contributory factors derived from foodborne illness outbreak data from several countries.

Significant factors that contribute to contamination of foods are:

The length of the lag phase and the slope of the log phase are affected by the following environmental factors.

Source: FDA/CFSAN, Foodborne Pathogenic Microorganisms and Natural Toxins

Potentially hazardous food does not include (other conditions also apply):

The first step in planning the investigation may take place in the office before you leave for the site. Consider and identify where to find the records you need for the investigation. The type and availability of records will vary with type of facility. For example, a retail food establishment and a food processor would have different records available. If these records do not exist in your files, they may still be available from the firm.

Review your files for:

One way to interpret food prep data is to plot out a time- temperature curve. Time-temperature curves allow you to evaluate if sufficient time and temperature occurred in a food to permit the growth of microbial pathogens, specifically bacteria. It's not relevant for viruses or parasites. To plot a time-temperature curve, you need to account for the time when the food was in storage, when it was being prepared, when it was being cooked, when it was being cooled, reheated, and held for service. You want to know whether there was an opportunity for bacterial growth, survival, or destruction in all those steps.

Now, let's follow the preparation, cooking, cooling and service of a roast beef. You have a preliminary flow diagram from your talk with the manager. You'll be filling it out as you talk to each employee and note their specific tasks in roasting the beef. Feel free to draw your own flow diagram as we go through this example.

Here's a raw roast beef that was held in the refrigerator at 39 degrees for three days. Some of the pathogens that might be found on a piece of beef include Salmonella, Clostridium perfringens, and E. coli O157: H7. A piece of raw beef would be expected to have a water activity, pH and oxidation-reduction or redox potential that would be conducive to the growth of bacterial pathogens. The refrigeration temperature and competition from non-pathogenic bacteria would inhibit the growth of pathogens. Most of the microbial contamination would be expected to be on the surface of the meat. But, there could be internal contamination due to piercing of the meat by knives, forks, and those tenderizers that turn a tough piece of meat into something a little more chewable!

First the meat is seasoned and placed in the roasting pan. It's then cooked in an oven at 350 degrees for several hours until the internal temperature reaches 145 degrees for 3 minutes. Ignore the post oven heat rise when factoring cooking time and temperature. This cooking process would be expected to destroy vegetative microorganisms on the outside and inside of the roast, but any spores would survive the heating process.

After the meat cooks, it's going in two different directions. After it is reduced in size, one part is going into the walk-in for a rapid cool-down. Tomorrow we'll reheat, slice and serve it. Now the other half can be sliced and served; we'll enjoy it when we're finished up here. Contamination from utensils or worker hands or both would survive and multiply if given the opportunity here. Re-heating to the appropriate time/temperature parameters will destroy vegetative bacteria and viruses on the surface of the meat.

Some key temperature guidelines:

Make every effort to avoid contaminating a sample by your actions during the collection process. Consider all samples contaminated and potentially infectious.

Hands should be washed before and after collecting samples. Wash your hands under running water with liquid soap for a minimum of twenty seconds; dry with a paper towel and use a paper towel to turn-off the faucet.

Wear gloves during sample collection if you have cuts, chapped hands, or dermatitis.

Ensure your personal safety. Avoid eating, drinking, or smoking in areas where there's a reasonable chance of exposure. When there is a risk of occupational exposure wear appropriate protective equipment to maintain a barrier for protection. Remember some pathogens can become airborne in an aerosol.

Wear a lab-coat, face-mask, and protective eyewear when splashes are likely. In general, wear a lab-coat, hair restraint, and cleanable boots to protect yourself and ensure that samples do not get contaminated. After equipment is used, treat the equipment as contaminated.

Caution: Unless you've received specific training in rectal swab collection, do not collect rectal swabs as it is possible to injure the patient by damaging the lining of the rectum. The same is true for using a saline enema to obtain specimens. Since nurses are probably the only ones in the audience with training in the collection of swabs and enemas the following discussion is targeted to the nurses on the outbreak team.

Most bacterial pathogens can be cultured from appropriately acquired rectal swab specimens. Collect at least two rectal swabs from each patient and place swabs in refrigerated (i.e., chilled 1-2 hours before use) Cary-Blair transport medium. When obtaining swabs from a patient, first moisten each rectal swab in the holding medium, insert the moistened swab into the rectum 1 to 1-1/2 inches, rotate the swab gently, and then return the swab to the same tube of holding medium. Try to ensure that visible fecal material is present on each swab. The handling and storage of the rectal swab is the same as for swabs of fresh stool discussed above.

Only qualified personnel should collect blood samples. For patients thought to have an illness caused by viruses or bacteria it may be appropriate to submit an acute-phase specimen and a convalescent-phase specimen. Obtain the acute-serum specimen as close to the time of onset of illness as possible; at most, within a week after onset of illness. Obtain the convalescent-serum specimen 3-4weeks after the onset of illness. Antibodies to viruses usually begin to rise the first week after onset of illness, peak by the fourth week, and can fall by the sixth week.

Unless you've received specific training in blood collection; do not collect blood. Only physicians, nurses, phlebotomists or specially trained health personnel may draw blood. Universal precautions need to be followed for the collection of all samples of blood and body fluids. Latex gloves should be worn for the collection process. Syringes and vacutainers must be sterile and used only once. The tourniquet must be used carefully, and properly to avoid ruining the sample and causing injury. Blood is obtained by venipuncture (the amount is determined by the type of test; 5-10 ml is usually adequate). Some tests require whole blood, some plasma, and some serum. Some tests cannot be done on hemolized blood. Serum or plasma for analysis should be separated from the cells within one hour after collection of the specimen. Refrigeration will slow down microbiological processes but will not stop them. Blood samples should be refrigerated.

Submit two serum specimens, an acute-phase specimen and a convalescent-phase specimen, for each patient thought to have illness caused by viruses or bacteria. Obtain the acute-phase serum specimen as close to the time of onset of illness as possible (at most, within a week after onset of illness) and the convalescent-phase serum specimen 3-4weeks after the onset of illness. If a viral agent is suspected, for optimal test results, specimens should be collected within 6 weeks after onset of illness.

If possible, obtain paired serum specimens from the same 10 patients from whom stool samples were obtained. Ten paired serum specimens obtained from well persons can serve as control specimens in certain studies.

Collect blood specimens from adults (15 ml) and from children (3 ml) in tubes that do not contain anticoagulants (usually red-top tubes). Centrifuge the blood and send only the serum for analysis. If no centrifuge is available, store the blood specimens in a refrigerator until a clot has formed; then remove the serum and pipette it into an empty sterile tube (using a Pasteur pipette). Refrigerate the tubes of spun serum until they are shipped. Refrigerate, but do not freeze, tubes containing unspun serum.

Ship serum specimens either refrigerated or frozen. If the clotting technique described above is used to obtain the serum, ship the specimens refrigerated so that they can be centrifuged before they are frozen. Specimens can be refrigerated by placing them in an insulated box with ice or frozen refrigerant packs. Frozen specimens can be kept frozen by shipping them on dry ice. Batch the specimens and send by overnight mail, scheduled to arrive at the laboratory during business hours on a weekday, if possible.

Antibodies to viruses usually begin to rise the first week after onset of illness, peak by the fourth week, and can fall by the sixth week. Acute-phase serum specimens should be collected in the first week of illness, and convalescent-phase serum specimens from the third to the fourth week after onset of illness but not later than the sixth week.