EPID 555 - Lecture 1.10.23 Steps to an Outbreak Investigation

Ch. 22 - Wedding punch outbreak (sodium nitrite) - cash cab

-Name of disease with which they became ill - methemoglobinemia -What color does blood turn? Chocolate brown -What feature of the epi suggested chemical? Rapid onset (3 hrs) and same time -Who was the first person to get sick? Photographer - got there early -How did HD first hear of the outbreak? News reporter called and asked about ambulances -Investigators needed to survey guests - what challenge did they face with this? Name at least 2 of 3 - many on invitation list hadn't attended and some signatures were not legible - neither source had phone numbers -There are outbreaks of reportable diseases (ex. Salmonellosis) - since there is no surveillance for methemoglobinemia, what does the law say about the reporting of an outbreak like this? All outbreaks are reportable regardless of the disease according to the state's rules and regulations -Case def had suspect and confirmed - main difference? Suspect had clinical criteria (cyanosis, headache, etc.) but confirmed had lab evidence of methemoglobin level 1+% -What chemical was substituted for citric acid in the punch? Sodium nitrite -Concern that outbreak was caused by foul play - what info determined it wasn't? Sodium nitrite levels were similar to what citric acid levels should have been in the punch and there may have been some language barriers with workers -More they drank = more ill - dose response -11 Blue Men ate what? Oatmeal

Ch. 3 Leptospirosis at the Bubbles summary

-Objective - what caused the outbreak? -Think about biological plausibility for hypotheses - St. Louis encephalitis (mosquito-borne illness) which is an arbovirus (mosquito-borne) - SLE can be worse in older adults - LaCrosse encephalitis can affect younger people but no fever - equine more serious in infants -See notes for this - couldn't really hear him -Meets with director of HD - this is an important step and is very common -Verify the outbreak - confirmed number of cases was much higher than expected -Perform descriptive epi - looking at onset back to June and making conclusion about common source (does end up being common source) - not point source, which means people came together around the same TIME - kids came many times to the source -Clue about 80% boys - helps to eliminate certain things as less likely - infectious disease wouldn't skew towards a certain gender -Can't confirm the diagnosis yet (step 2) because he doesn't know what he's dealing with -Skips to step 8 - develop hypothesis - so that he has some idea of what to do next - performs interviews of some of the typical cases - open-ended - like doing a focus group or in-depth interview before generating a questionnaire -Case suggests he check out the Bubbles - helps generate hypothesis -Bubbles - concrete structure connected to irrigation system with bubbling water -Contacts other cases to ask if they also went to the Bubbles - yes they did - has a sound hypothesis now -Talks about coliform found in the water - 240,000 colonies per ml of coliforms - can mean there is fecal contamination in the water -Environmental investigation - takes photos of the river to see what may have contaminated the water - sees a herd of cattle upstream - gets blood and urine specimens from the herd for testing - blood results show 21/25 positive for antibodies for leptospirosis - doesn't mean they've been ill recently but it's a clue - urine shows 9/43 shedding culture positive/living leptospires -Cattle are related to the outbreak -Also tests water from the canal for leptospira - negative -Also tests standing water in the field - positive -Human studies - blood samples from those who were ill - sends for antibody testing and tests for blood cultures - finds evidence of antibodies in the typical cases but blood cultures are negative (by the time he drew blood cultures, they would have recovered - so it makes sense that they were negative - not a problem - he arrived pretty late) -Assembled a long list of evidence that it was a leptospirosis outbreak -Biologic plausibility for exposure - swam at the Bubbles and got abrasions -Cows upstream that tested positive in urine - that urine is in the field in standing water -High seropositivity in these cattle compared to other cattle -The herd was new to the area - explains why they hadn't seen an outbreak before this -Children who were ill tested positive for leptospirosis antibodies - wouldn't normally expect that -Their illness was compatible with leptospirosis -Behaviors of the kids made sense - abrasions - going underwater (eyes) -Incubation period - onset 7-10 days after exposure to Bubbles - onset matched incubation period for leptospirosis -Number of cases increased 10 days after warmest days of June/July - we would expect more people at the Bubbles on the hot days - more cases 1 incubation period out -What did he do for case ascertainment? -HD had reports -Asked cases for contacts who went swimming -Went to the site and identified kids who signed the walls of the Bubbles -Did a questionnaire to local HS students -Put info in newspapers/media outlets -Syndromic review looking for fever illness from hospital and clinic records -Prevention/control measures didn't wait until analytic confirmation -Warned public not to swim there anymore - told people and posted warnings -Restricted cattle access to canal -Stopped using rill irrigation - was causing standing water -High risk groups for leptospirosis - veterinarians, farmers, people who work with fish, sugarcane workers, lab workers -Common thread with moisture and water - exposed to urine of animals like rodents or cattle

Pathogen characteristics and external factors that influence outbreak characteristics (likelihood that someone becomes a case)

-Exposure to disease influenced by many factors -Might not be exposed at all -Age (infant vs. 20 vs. 80), gender, occupation, living/working conditions, close and prolonged interactions

Descriptive epi - time intervals of interest

-Exposure to onset of illness (incubation period) -Diagnosis to reporting (delay of reporting - want notification to happen ASAP) -Duration of hospitalization (provides morbidity figure - can calculate cost indirectly which can help get resources in the future)

Steps: ASCERTAIN/find cases

-Find them -Interview them w/ case report form or questionnaire designed for the outbreak (might not need a new one - ex. Pertussis is already reportable so you can use the report form and add details as needed) -Enter data into a computer

Epidemic

-Greater than expected but more widespread, like nationally or state-wide *Media might call something an epidemic when it's not

Outbreak

-Greater than expected number of cases during defined time and place *Rabies and Botulism can be single case outbreak

Special considerations

-How does the status of a case (confirmed vs. probable or primary vs. secondary) inpact analyzing the data to make decisions? -Ex. dataset w/ confirmed and probable - risk factor analysis = might only look at confirmed if sample size is large enough

Benefits of outbreak investigation (RMPET LMLSN)

-ID populations at RISK for disease -Characterize MODE of transmission -PROVIDE info to PREVENT further transmission (control the outbreak) -Opportunity to EVALUATE EFFECTIVENESS of public health programs -Opportunity to TRAIN public health staff (ex. local/rural HD might not ever see outbreaks) -Fulfill LEGAL obligations and duty of care (state/HD responsible for controlling health) -Opportunity to educate public through MEDIA outlets (front page news) -Opportunity to LEARN more about a disease -SHARE knowledge and findings w/ other healthcare professionals -Learn something NEW

Ch. 33 - Hard to reach populations w/ Dr. Wayne Wiebel

-Indigenous leader outreach model (ILOM) - to reduce HIV infection among IV drug users -Combined approach of research and treatment to address heroin addiction - drug abuse as an epidemic - Hughes -Hired indigenous research assistants - ex. Former heroin dealer from south side of Chicago to engage the population - initiate and maintain relationships with target population because the drug users would go to the same places every day -Found that they could introduce methadone treatment to these people -ID outbreaks as they were unfolding - go into the neighborhood, introduce methadone treatment, and turn the outbreak around -Thought these methods might work for other populations too as HIV/AIDS arose - MSM outreach was good, but didn't know what to do with IV drug users - used indigenous leader outreach model to reach injectors -Male prostitutes who were sharing injection equipment - one was intentionally infecting others with HIV/AIDS -Former heroin addict that was highly respected on the NW side - knowledgeable about the shooting galleries - San Francisco distributing bleach instead of clean needles to prevent spread of HIV - wanted to try that in Chicago - demonstrated that it was possible to work with this population -Symbolic interaction - theoretical underpinning of ILOM - apply theory to interventions - based on the fact that meanings for individuals are socially constructed - there is no inherent meaning to anything - things have no meaning until people give meaning to them -Behaviors are socially learned and constructed - they are amenable to change based on what people think and believe - can't understand them until we see behaviors through the eyes of those with whom we are trying to intervene -Applied this to needle sharing - refine strategies to change behavior and prevent spread of HIV -Hired many indigenous leaders - had to convince university to hire former addicts - help them understand that they were experts - developed 3 major field stations -Replicated the approach in SE Asia too

Communication with the media

-Is important

Steps: Perform SUPPLEMENTAL lab or environmental investigation

-Lab - ex. need to know which Salmonella strain -Environmental - molecular testing, facility inspection (send inspector to restaurant)

Steps: CONTINUE ascertaining additional cases

-Ongoing surveillance -During and after conclusion of investigation

Steps: Assess ADEQUACY of control measures

-Look at surveillance data to see how many more cases are coming in -Ex. We expect cases to stop after restaurant is closed - if we know incubation period, we can say we shouldn't see more cases after 2 more incubation periods

Ch. 7 HIV/AIDS

-MMWR June 1981 - didn't realize they were dealing with a new disease right away - thought maybe it was a kind of CMV - ex. CMV caused pneumonia → death -We know now that severe immunosuppression = opportunity for other viruses to come out -4 questions they needed answered early on -Verify that an outbreak is really occurring - is this a new disease? -What is the extent of the outbreak? Is it happening anywhere else? Related to ascertaining cases, descriptive epi - got info from San Fran and NYC and established surveillance early on -Why MSM? Related to performing additional studies -Other unusual diseases occurring among MSM? Related to descriptive epi -Case-control - cases were MSM with PCP/KS - matched on age, race, sex, MSM -Wanted random sample - why was this difficult? There is no list -Where did they find controls? 3 control groups - from clinics, etc. -Controls didn't have KS or PCP - could have had low CD4 counts and been infected with HIV because they had no test -Checked them for immune suppression to deal with this - found that some did have immune suppression - potential limitation, but didn't affect overall conclusions -Hypotheses - CMV, poppers (related to supplemental lab investigation - sent samples to test for contaminants - didn't find anything related to immune suppression) -Case-control results - having a large number of male sex partners related with KS/OI -Patient O - flight attendant - helped w/ hypothesis that sexual activity was important for transmission - he was highly mobile

Ch. 12 - Shigellosis outbreak in daycare centers in KY

-Main point of this outbreak/chapter is the community task force as a method of social mobilization to change behavior of a large group of people to control a disease -Shigellosis -Bacterial diarrheal disease -Fecal-oral transmission -Hand hygiene is important for control -Treatment with antibiotics limits duration of shedding -Symptoms -Acute abdominal pain, fever, watery diarrhea, bloody mucus in stool -Might not have all of these because someone might have partial immunity from being ill in the past -Can fool providers into thinking it's something else - differential diagnosis - ex. Some people with abdominal pain have been taken to the operating room for appendicitis -Lexington KY 1991 - 138 culture-confirmed cases Jan-May - most in people who attended or worked at childcare facilities or elementary schools (young children or workers who are in direct contact with young children - also parents might be cases) -Cases = lab evidence of infection with Shigella (very specific case definition - best to be specific in this case because it will help you avoid picking up other diarrheal illnesses which can be very vague - specific is also good if an intervention would be very expensive or toxic - sensitive would be good if it's a very severe or infectious disease like Pertussis which is a baby killer) -Standard HD activities before CDC arrived (lots of good work) -Visited schools and childcare centers with kids with diagnosis and tested classmates (requires lots of resources but is useful - can also detect those who are just mildly ill to prevent further spread) -Home visits -Collection of stool from close contacts -Instructions about hand washing before and after meals, diapering -Sent info to parents about the outbreaks -School exclusion if they were sick -Why did the outbreak still persist? Lack of behavior change - this is why community partners are so important because behavior change is difficult -Needed social mobilization -CDC comes in - goes to daycare centers - sinks were too high, children were helping children wash their hands - lots of touching surfaces (play equipment - bad because Shigellosis has a very low infectious dose - hand to mouth) - children walking around with just diapers - training toilets near play equipment with no sink nearby, no toilet paper -Did community task force -Visited facilities to ensure compliance with control measures -Increased emphasis on hand hygiene -Stopped collecting stool samples - frees up time for staff and money for other resources (focus on control - hand hygiene) -Increased awareness of worker practices by daycare directors - discordance between what workers were saying/doing vs. what daycare directors thought they were doing -Diaper changers don't prepare food -Community task force -What stakeholders will you invite? Daycare directors, staff, environmental health team members (input about cleaning solutions), hospital lab workers (give us a sense of any surges/declines with sample processing, advice about sending samples), elementary school principals (ensure school nurses, teachers are enforcing hand hygiene), parents (can spread word to other parents), media (HD press release, local press SOCO prevention message) -Case-control study - risk factors for outbreaks - used centers, not individuals -Cases: center with at least 3 cases - Controls: center with no cases (excluding centers with 1-2 cases, but this is fine because we're focusing on problem centers - sporadic cases will always happen) -Controlled for diapers and low SES (matched) -Epi curve - community wide intervention in June = steep decline - didn't totally wipe it out right away -What step does this chapter illustrate? Perform new control measures, ensure compliance (like pseudo outbreak illustrated confirm the diagnosis)

Cluster

-May or may not be more than expected *Ex. 3 people on same block w/ brain cancer - it is a cluster but there may be no cause other than chance - not an outbreak

Steps: Develop HYPOTHESES

-Might have developed earlier with the help of descriptive epi steps

Transmission basics (5) - what we need to know

-Modes of transmission of pathogens -What influences transmission? (ex. impact of environmental conditions like UV light) -Amount of a pathogen (ex. IV infusion w/ contaminated bottle - small vs. large amount of contamination) -Human behavior (ex. only w/ a case for 30 secs vs. 1 hour) -Risk factors (talk to cases to determine this)

Steps: Introduce PRELIMINARY control measures

-Ex. close the restaurant, tell people to cover their cough, ID close contacts to provide prophylaxis

Steps: COMMUNICATE to those who need to know the info

-

Steps: DRAW conclusions and formalize recommendations

-

*****Ch. 23 - Ammonia chicken outbreak

-42 kids, 2 adults to 5 hospitals within 1 hour of lunch -Verify the outbreak - common sense (vs. more than we would expect in surveillance data) -Confirm diagnosis - was performed later but no "confirm" step -Classic meaning is that samples get tested at state lab and they confirm that hospital lab got it right -Assemble the team - EIS, student, local HD, etc. -Case definition - person, place, time -Person - student or teacher who developed symptoms -Place - 2 schools that served food from Laraway kitchen -Time - onset within 180 minutes of lunch -Initial hypotheses -Chemical -Bacterial - causes toxin to form -Earlier in the day? Ex. meal from previous day -Bioterrorism event -Count cases - questionnaire for students and teachers - interviewed students and teachers before the end of the school day - cohort approach - ID who is a case and who is not a case -What food had they eaten, quantity (dose response), odor -Epi analysis -81% had onset within 60 minutes - 157 became ill (AR 50%), 91% reported odor -Field investigation -Laraway kitchen - lunch arrived in packed boxes and stored -Students and teachers reported odor of cleaning products -HD identified brands of all foods and their pathway from packaging to delivery -Epi curve - not incubation period for chemical - just histogram -Epi findings - chicken tenders w/ smell associated w/ being a case RR9.2 -Why different ARs for schools? Oak Valley maybe just got lucky and not all chicken was tainted at the same amount - could also be that the longer that the chicken was out of the freezer, the ammonia dissipated over time -Ammonia levels 2,500 ppm in uncooked chicken tenders -Contaminated during warehouse leak - packaging was a bag inside a box - liquid ammonia becomes a gas very quickly which can penetrate cardboard very easily - it can also penetrate plastic bag so once it gets through that, it binds to the ice crystals on the chicken - once it does that, there is no odor - fools people into thinking the product is safe - doesn't smell again until it is cooked -Warehouse manager got 1 year in prison

HIV medication adherence in India

-A Public Health Approach to Problems -Surveillance → Risk Factor ID → Intervention Evaluation → Implementation -A Request -Asked by Indian physician in Chicago to do research in India - HIV medication adherence - helps decrease viral load to undetectable levels - personal and public health benefits -Unexpected findings from an ART adherence study in India -SHARE India - NGO clinic and lab -25 person study - 40% had missed at least 1 dose in the past 4 days - pretty high -Other Studies Attracted My Attention -Did a rural village survey to learn about childhood immunization practices -Epi of diarrheal disease at a cholera hospital - what were the actual germs? Had lots of crypto - should test for HIV too -Restaurant food handlers study - understand what they know to prevent foodborne disease -Analysis of syphilis coinfection data -Years passed - back to Study of Adherence -Sample size bigger this time - 211 - recruited 9 facilities - estimate prevalence of nonadherence - ID risk groups or risk factors for intervention -Designed a questionnaire to ask patients about risk factors of poor adherence (ex. MSM, sex worker, drug abuse, depression, motivation to take meds), reasons for missing doses - piloted it before using -Findings - ⅓ had primary school or less - low literacy population - 20% not adherent to ART - older patients (45+) more likely to not adhere compared to 31-45 - moderate to severe depression also = elevated prevalence of nonadherence - female sex workers also higher -Prioritized sex workers because they would have the most impact in terms of spreading HIV - improving adherence will reduce spread -Next Steps -Cross-sectional survey among a larger number of FSWs to confirm prevalence of nonadherence and what are the risk factors for nonadherence - needed to find locals to reach the population - NGO helped and he collaborated with Wayne -Created survey based on lit review -Wayne said he needed to understand the context first - need to make sure you're asking the right questions and in the right way - see the world through their eyes first - can't assume sex workers across the world have the same risk factors -Worked with NGOs to get answers about types of sex workers - brothel-based, street-based, highway-based, lodge-based, etc. -Then went to India and CMM group would bring women to the NGO so he could talk to them and ask open-ended questions - most were arrested, illiterate, sold by husbands/boyfriends -ART centers - people would get irritated with the FSWs, doctors shouted - shame, stigma -Biggest concern for the women was food - new study with hypothesis that food insecurity increases risk - associated with nonadherence in the study - highly prevalent - arrest and stigma also associated -Mahila-Life Approach - combo intervention - increase health literacy for the women, raise understanding about ART meds, improve nutrition and cooking, increase knowledge of social entitlement (ex. Free food programs, health insurance programs), vocational skills, meet with police and ART center workers to make sure they understand how they are contributing -Lessons Learned - importance of qualitative data, combo interventions are more complicated but more effective - partnerships are important

Ch. 24 summary - When your food grows blue (aflatoxin in wet maize stores)

-Aflatoxin - toxin produced by a fungus - mainly impacts the liver - can become large and inflamed - can cause death - not like a bacteria, can't cook it off -Pg. 288 3 questions -Was the outbreak real? This is the 1st step - determining if there is an outbreak (common sense) -Was the outbreak caused by aflatoxicosis? Confirm the diagnosis and/or generate hypothesis -If there was a mass intoxication, what was the source of aflatoxins? Generate hypothesis, perform additional analysis to ID risk factors, perform environmental investigation -Confirm diagnosis - preliminary findings from Kenyan lab, then CDC lab confirmation -Preliminary study, then more thorough study -Preliminary work - looked at preliminary data that had come up negative (not traditional foodborne outbreak) over the phone -2nd phase - 3 parts - went to hospital to interview cases, go to patient's field (hypothesis affirmation - open ended questioning to inform questionnaire), UV light to check maize - glows blue (environmental investigation - confirms hypothesis) -3 efforts: Test the food with commercial test kits to ID aflatoxin, case-control study to ID risk factors, test ill persons for aflatoxin in serum (blood) -Case definition - acute jaundice of unknown origin leading to hospitalization - no history of cirrhosis or obstructive liver disease - acute jaundice is a syndromic terminology - unknown origin is trying to make it specific (ex. Excluding acute Hep B patients) - needs to be severe enough that it led to hospitalization - selecting from same time period (height of epidemic) -No confirmed vs. probable -Controls - same village (same opportunities to be/not be a case - share similar soil, climate, farming practices, etc.) and at least 2-3 huts down from the case - no jaundice -40 cases, 80 controls was what was planned - 28 and 43 in the end - hard to find case-patient homes in rural Kenya - also had quality control issues - some case-patients didn't have blood samples, others were missing their randomly selected controls -Found controls by spinning a bottle and walking in that direction -Findings - cases had significantly higher aflatoxin levels than controls -Assessment bias - by lab techs could have happened if samples were labeled case/control (might lean towards positive for aflatoxin if the results are unclear and you see that the sample is from a case) - didn't happen here b/c lab techs couldn't see if samples were from cases or controls -Proper crop selection, irrigation, pest control, post-harvest processing of grains and pulses are crucial for preventing this -Challenges are poverty and behavior change

Other factors that influence severity of disease

-Age at time of infection -Immune system (ex. HIV/AIDS, pregnant) -Antibiotic use (ex. diarrheal illness - antibiotics kill good germs = easier for germs to make you sick) -Genetic factors -Duration of exposure to organism -Amount of orgnism ingested -Characteristics of organism

Ch. 5 Legionnaires outbreak summary - July 1976

-Background - attention due to swine flu - concerned that this was going to be the next big pandemic (didn't happen) - also a presidential election year (concerned it may be a terrorism event targeting Veterans) - also bicentennial - also pope was going to visit -Backlash against CDC - public felt vax caused neurologic complications -Figures - show that the issue wasn't a Philly issue - there were more cases than expected, but if it was a Philly issue there would have been a much larger peak - trying to determine the extent of the outbreak -1) Confirm the outbreak is real - compare current numbers to past years - use surveillance data if you can OR use common sense (didn't know what this disease was, but wouldn't expect as many pneumonia deaths) -2) Confirm the diagnosis - ex. We are being told that there's an increase of Salmonella - sent samples to state lab where they will test to confirm they really are Salmonella -Couldn't do that in this outbreak b/c unknown cause - couldn't test for it -Set this step aside and explain why -3) Assemble the investigation team - CDC got involved and sent out EIS officers - there were already epis, statisticians, toxicologists, environmental engineers, etc. on the team - got lots of resources due the context -4) Tentative case definition - person, place, time -Person - fever, chest x-ray evidence of pneumonia, fever/cough at convention, at hotel -Place - in Philly or connection to the hotel -Time - July 1 - August 18 - American Legion Convention too -Sensitivity vs. specificity balance - catch as many cases as possible (sensitive) but want to make sure they are true cases (specificity) -Highly sensitivity good for high mortality -Can't be more specific with inclusion of testing positive for Legionnaires disease - doesn't exist - not a fully specific case def -5) Count/ascertain cases - find them and interview them to collect data - line list - spreadsheet that includes clinical and lab info - each line is a case and contains info -List of basic info on cases so far - use for hypothesis generation -Ex. Lots of children based on age column - what kinds of things do children get? -Go to hospitals/clinics to interview cases and collect data - important because no surveillance on this yet -6) Descriptive epi - could ID if someone visited the hotel, what they consumed, any similarities between cases - found that the hotel was really important - even those that didn't stay in the hotel became cases -Trend was older men - higher AR - also delegates of convention were more frequent cases - mortality was 14% which they felt was high -Talked about outlier cases - ex. Pilot who slept there for a bit, woman who just went to the bathroom and then got sick and maybe died later -Helpful to look at outliers for hypothesis generation - ex. Could be airborne/respiratory -7) Perform supplementary investigation - lab or environmental - ex. Go to the restaurant, do molecular testing -8) Hypothesis generation -Thought about pigeons outside the hotel - bird diseases - histoplasmosis and psittacosis - both cause pneumonia -Influenza -Chemicals -9) Set up preliminary control measures -Not illustrated in this chapter because they aren't even sure it's respiratory or where it's coming from - can't really say turn off this machine, block off this room, etc. - they did highlight wearing masks while visiting patients -10) Address whether control measures are working -11) Decide about whether or not to do a case-control study -They do this - primary hypothesis doesn't exist - just some ideas -Didn't find any smoking gun or obvious exposure -12) Continue monitoring surveillance data to determine if outbreak is continuing **They eventually figured it out - went back and investigated some of the cultures and noticed something else growing - identified the pathogen that he thought was just dismissed as a contaminant originally -Tested specimens from past similar outbreaks and confirmed the same pathogen - Pontiac fever (without the pneumonia) - waterborne just like Legionnaires

Cyclosporiasis outbreak exercise 4/4/23

-Background - exotic parasite sickens Canadian businessmen - no person to person spread - incubation period 7 days - attended meeting in TX May 9th-10th - symptoms include diarrhea, weight loss, etc. -Question 1: -Incubation period for cyclosporiasis - 7 days -How will this be used in the investigation? Timeline for questionnaires, business meeting in TX matches time frame from when cases were reported (May 16-18) -Question 2: -On what sources of infection should PH officials focus for the three cases? Fecal oral transmission route, not person to person (food or water - but primarily foodborne) - focus on food and beverages - but oocysts are inactivated at less than 28 degrees Celsius or 140+ degrees F -Is it possible one of them was the source of infection for the others? No - wouldn't make sense w/ timeline b/c they all got sick at the same time (incubation period 7 days) -Do you think it's likely that they became infected in TX? Yes -Part II - Outbreaks in TX -What do we know? Private meeting in TX, 28 attendees from 3 US states and Canada -Question 3 -2 most common types of epi studies used to investigate an outbreak - case control and cohort - use cohort if you have defined group of people (small group like a wedding) - if statewide outbreak, case control is better -Which would you use? Cohort because it's a small group of 28 and we have their contact info already -What next? TDH initiated retrospective cohort study - interviewed 27 of 28 - created case definition - case = diarrhea 3+ days in someone who attended the meeting - lab confirmation not required - sensitive because lab confirmation took a long time at that time - they also want to capture as many cases as possible - point source epi curve tells us it was only one event (business meeting) -Question 4 -Why question people who didn't become ill? Need to have a comparison group to figure out which is the correct exposure -Results of cohort study - 12 of 13 attendees who ate berry dessert became ill - only 1 of 11 who didn't eat it got ill - RR 10.2 with p-value <0.0001 -Question 5 -Interpret results - risk of being a case is 10x higher for those who ate berries - high AR for those who did eat it, very low AR for those who didn't - might have forgotten they ate it, food might have touched the berry dessert - p-value very small which is important since it's such a small sample size (hard to get small p-value with small sample size) -Question 6 -What problems in study design or execution should you consider when reviewing results? Recall bias, wells might not remember, haven't confirmed what foods were consumed (not sure other ingredients in berry dessert), selection bias (might have misclassification bias, 4 people didn't provide any info) -Hypothesis - strawberries are the source of the outbreak - RR 8.0 for eating strawberries -Question 7 -What additional studies might confirm the hypothesis? No cyclo lab tests available at this time so we couldn't test the strawberries - could look for more cases in the community and do another study - could also do a trace back study (contaminated groundwater) -Strawberry info - found that they were grown in CA - 3rd outbreak in Toronto - encouraged physicians to report cases to local HDs -Part 3 - outbreaks in other states -End of June 1996 = 800+ cases reported to CDC from 23 states - some pointed to raspberries instead of strawberries - national news, strawberry recall - no individual farm has been named as the source, so the whole industry in CA is taking a hit - multiple berries now? Ends up being raspberries - $40 billion lost, thousands lost jobs -Question 9 -Would you undertake CC or cohort to investigate source of cases in NJ? Cases in NJ are not linked to single event or well defined area, so case-control -Case def in NJ - case is patient w/ lab confirmed case and history of diarrhea - slightly more specific because there are many more cases now -Question 10 -How might you ID cases for the study? Hospital data/lab reports since it has to be lab confirmed (otherwise you could do RDD, could put out PSA in the news for people to call in) - they would have gone to doc -Who would you use as controls? Family controls so they have the same risk profile and are eating similar things, sharing the same fridge -CC results - did RDD to HHs in same community - unclear if they did neighborhoods, towns, or census tracts - 2:1 ratio - controls couldn't have loose stools in previous 30 days - 30 cases and 60 controls interviewed by phone to ask about exposures - found that 21/30 cases and 4/60 controls had eaten raspberries = OR 32.7, p-value <0.00001 - odds of exposure to raspberries 32x higher for cases than controls -Other studies - supported raspberry hypothesis -Question 12 -Would you alert the public to this PH threat? Pros: avoid raspberries = avoid cases, prevents additional cases and morbidity/mortality and costs associated - important for people who are very young/old/immunocompromised - Cons: people will avoid all raspberries on the market = financial/economic issue - no perfect answer -Part 4 - traceback and environmental investigations - ID'd distributors and importers of raspberries for the 54 events - found 29/54 - 21 came from Guatemala - other 8 could have originated from Guatemala but unsure -Question 13 -Does this info support raspberries as the source of the outbreak? Yes, especially because only 4-20% of fresh raspberries came from Guatemala during that time -Question 14 -On what cultivating or harvesting practices would you focus? Irrigation, fertilization, where do they get the water from? Proximity of fields to human fecal matter, substances added like fertilizer, insecticides, etc. - harvesting equipment, human handling, storage containers, etc. - focused on human feces during this outbreak because they didn't know if animals could be infected with cyclo -Environmental investigation results - 6 farms supplying exporter A in same region on Guatemala - 5 pulled water from wells that varied, 6th pulled river water - exporter B was 25km from A and also used well water - all 7 used ground level drip irrigation (drips only on the soil so irrigation water wasn't contacting berries) - used well/river water to spray insecticides onto the berries and all were picked by hand -All water samples showed fecal coliform but no cyclo -Question 15 -Cyclo not found in raspberries or water samples in Guatemala - if the Guatemalan raspberries were the source, list plausible explanations - life cycle of cyclo, testing methods might have been insensitive - water and raspberries were collected after the outbreak so they may not have been contaminated at the time of the testing -Hypothesized that berries became contaminated from insecticides mixed with contaminated water - but also improperly constructed wells near sewage could be possible, especially in rainy season - they are also difficult to wash properly because of how fragile they are - CDC and FDA determined that raspberries from Guatemala were the most likely source -Part 5 - control and prevention -Question 16 -What specific measures would you suggest? Construct wells correctly, divert latrine wells away from water source - water should be potable for insecticides - give farmers info about hand hygiene -What happened? Growers improved hygiene and water quality - implemented systems to ID potential contamination sources - FDA eventually banned imports of Guatemalan raspberries due to another outbreak -Question 17 -Do you think raspberries were the source? Yes -Criteria for causality (strength of association, bio plausibility, consistency with other studies, exposure precedes disease, dose-response) - almost all - just don't have anything on dose response -How would you convince others of the validity of findings? -Conclusion - US ban lasted 1998 growing season - no outbreaks in US during that time - Canada didn't block, and they had an outbreak linked to Guatemalan raspberries -Question 18 -What criteria would cause you to implement control measures before you were certain of the source? High mortality, high morbidity which can be expensive, what kind of population is getting sick (ex. Only children), impact on business/industry (try to avoid if possible)

Steps: PLAN for field work

-Be organized, determine leader, phone/fax/email key stakeholders/team members -Review disease info and lit -Discuss w/ lab -Internal memorandum

Ch. 11 Botulism in Cairo summary

-Botulism incubation period 18-36 hours - can be up to 10 days though -Importance of anaerobic conditions - spores are fine unless they are given certain anaerobic conditions -Besides home canned goods, this shows importance of fish - also onions or potatoes - germ is found in soil so it makes sense -Classic symptoms - double/blurred visions, drooping eyelids, paralysis, slurred speech, difficulty swallowing, muscle weakness (starts in the head and descends - can impact respiratory muscles which can lead to death) - cranial nerves are important -Can be put on a ventilator - there is also an antitoxin - blocks progression - doesn't cause the toxin to be released - prevents paralysis from getting worse -If someone's lungs are already paralyzed and they are on a ventilator, antitoxin won't get them off the ventilator - it will reduce how long they will be on the ventilator -Also wound botulism and infant botulism -Step 1 - verify the outbreak is real - this was the first time botulism was reported in Egypt so they knew it was an outbreak - larger than expected because expected is 0 - no surveillance data to use in this case so they used common sense -Step 2 - confirm the diagnosis - not until end of chapter - takes samples of fish, stomach contents, and stool samples back to CDC - tested positive -Determine extent of the outbreak - partners will change - might just be investigator and HD or it might be multiple HDs or multistate outbreak -Called other health officials in other cities to determine that there are no cases outside of Cairo -Assemble the team - no description of a team in this case -Develop case definition - suspected case = illness in a hospitalized patient whose physician suspected botulism and who had at least one of the following symptoms: dyspnea (shortness of breath), blurred vision, or ptosis (drooping eyelids) -Everyone has to be hospitalized - severe illness required -Physician suspected = clinically compatible -Sensitive vs. specific - this one is pretty sensitive - can have any 1 of the 3 symptoms listed - also no lab requirement - also doesn't need to be physician confirmed, just suspected - makes sense because it's early on and deadly -Ascertain cases - called hospital administrators and told them to report cases (active surveillance) - there was also publicity which likely stimulated some reporting -Epi analysis - Christians were a majority of cases even though they were a minority of the Egyptian population - 91 reported with cases, 18 died - 61% of cases were interviewed (some died or were too sick) - age range was 8-85 - 60% were male which isn't very skewed - onset of illness was within 24 hours of a holiday (Sham-el-Nessim) which indicated it was likely linked to the holiday gatherings -Coptic Christians - mainly in Egypt - have their own pope - Orthodox branch -Preliminary control measures - can't do much because they don't know what food caused it and the feast is over so they can't stop that from happening -Hypothesis generation - must be with holiday but not sure what food -Supplemental lab/environmental lab investigation - can't do this yet either -Observe or perform additional study - need to do additional study to determine cause - case-control -Case-control - cases from hospitals - changed case definition - need something more specific as we have learned more - at least 3 symptoms instead of just 1 -Limitations that they faced - some family members had died so they had to use a surrogate family member - couldn't get a very specific list of foods like in a restaurant so had to ask open-ended questions about food consumption -Results - all 16 cases ate fish compared to 10 controls (6.6 OR) -Why aren't we blaming olive oil and lemon if they are also associated? It's a very common seasoning - ex. Burger with E. coli and onions - onions would have E. coli too because it was on the burger - doesn't mean it caused the outbreak -Fish was purchased from the same store - owner was arrested - told police he never made the fish in his life -Biologic plausibility of fish - fish sucks up soil so if there were spores in the soil and is in the fish intestines, humans would consume that when they eat the fish - they also don't chill the fish which favors growth of germs - gut environment is also anaerobic which favors growth too - would also put fish in a covered barrel which creates anaerobic environment - they also used a salting process which wouldn't penetrate the gut in high enough concentrations to kill germs within the fish -Why are outbreaks more prevalent around holidays? Sometimes because more people are gathering (close contact - ex. COVID) - bringing large number of people together around a food source, so more chances to get sick - also increased volume of food production around holidays (food handlers have to cook for more people and may cut corners) - also homemade errors are more common when people are rushing - also there may be extreme temperatures (ex. BBQ meat exposed to high temps while waiting to grill it = more likelihood for pathogens to grow - if meat isn't cooked properly the germs won't cook off)

Example of early swine flu case definition (confirmed vs. suspected, close contact)

-Confirmed was someone w/ acute respiratory illness (this is general/sensitive) w/ lab confirmed swine flu at CDC (very specific) by one or more of the following tests: PCR, viral culture, four-fold rise -Super specific - only wanted to announce cases they were certain about because it was a very big deal in the media - high stakes -Later this definition became more sensitive -Suspected case (no probable) - person w/ acute respiratory illness who was a close contact to a case while the case was ill OR acutely ill w/ recent history of contact w/ animal that had swine flu -No info about lab testing -Later this changed -Close contact defined as within 6 ft of a person w/ confirmed swine flu

Outbreak investigation is ... (2 things)

-Control (stop it) -Prevention (prevent future cases) *In real time -Organism, host, and environmental factors

Ch. 27 - DEG from cough syrup in Panama

-DEG - Patent medicine outbreak in 1937 - wanted it to taste better for kids so they added diethylene glycol which tasted sweet - over 100 died, many children -Thought it might be this because patients didn't have fever which pointed to toxin exposure - renal failure and neurotoxicity = DEG poisoning? -How does the outbreak get recognized by the ministry of health in Panama? Physician said there was an unusual amount of patients with renal failure and neurologic problems -Early descriptive epi from cases - 21 cases, mean age 62 so older population, case fatality rate 57% which is very high -Verify outbreak - no surveillance because this is a syndrome - common sense -Confirm diagnosis - can't send a lab test to state lab, so they just need to make a diagnosis instead - did some bacterial and viral testing to try to make a diagnosis but those were negative, then they called in CDC -3 features of this outbreak that warranted CDC involvement despite Panama's significant epi capacity: large, high fatality, rapidly growing -Observe vs. perform additional studies (which is what CDC does) -Need to have a sound hypothesis, high morbidity or mortality, high visibility of the outbreak with media attention, enthusiasm by those affected, novelty of the pathogen and its mode of transmission or its clinical manifestations (learn something new), availability of personnel/resources, whether it's already controlled, recall limitation -Key clinical clue that helped predict that this syndrome was not infectious - no fever -Objectives of the study/investigation from the CDC - identify the route of transmission so as to stop it, determine cause of unexplained renal failure - helps with control measures -Assemble team - multidisciplinary at the CDC - toxicology, epi, lab, etc. -Case ascertainment - not explicit about that here - implied that they have been finding them -Hypothesis generation - CDC's early hypothesis was DEG related because of clinical presentation - kidney and neurological problems -Early descriptive epi - 21 cases, 57% fatality rate, mean age 62, mostly males, etc. -Demographics don't match DEG - usually children and no gender differences -Eventually they get more cases though and the gender ratio evens out - but still older so need to figure out why this is -No family clustering - clue that it's not infectious because foodborne things would spread in families -Medication that they considered as the cause - physicians suggested lisinopril (hypertension) - recently added at the hospital - not a great hypothesis initially because it was only really a temporal coincidence - but it was good because many cases took it -Minister of health recalls lisinopril - political move to show it was being taken seriously - also gives opportunity to see if cases decrease after recall -How did they definitively determine if lisinopril was at fault? Sent it to a lab for testing and found that it tested negative - not the cause -Needed to perform additional studies other than testing lisinopril - sent samples to CDC and FDA to test for DEG in cough syrup that was obtained from the hospital -Also did case-control study - set it up in a hospital setting because all cases w/ severe illness required in-patient care and the majority had only received care from that individual hospital -Case definition - patients admitted to CSS hospital on or after August 16th which was 1 week before the outbreak - acute renal failure of unknown etiology OR acute worsening of pre-existing chronic renal disease -Controls - patients admitted for any other cause at CSS -Matched on age and date of admission - want patients to have had similar opportunities for exposures - 5 controls per case which is a lot - more than 4 there is no benefit statistically -Discovered that sugar free cough syrup was implicated - biologic plausibility is because there are previous outbreaks - explains lack of child cases (sugar free = older people, those with diabetes) - lisinopril causes dry cough - local hospital also made it which explains geographic limitation -Captured cough syrup with 2 questions - did you have it Y/N - open-ended question also - both showed elevated OR but they differed quite a bit - Y/N = 13.1 vs. open-ended 37.3 -Why? Recall limitation - both groups equally likely to not remember

Steps: DESCRIPTIVE epi

-Demographics (ex. gender, age) give us clues about where clusters are (ex. 75% of cases are school-aged) -Helps with hypothesis generation -Ex. Epi curve for onset of symptoms for cases - can do confirmed and probable in different colors -If shape of epi curve is normal (bell curve), suggests normal distribution of incubation period which suggests single exposure = POINT SOURCE OUTBREAK (people went to same place, got exposed at same time, etc.) -Vs. Propagated (person to person), clearly defined secondary cases - might not follow normal distribution

What is epidemiology and how does outbreak investigation relate?

-Distribution (person, place, time) and determinants (risk factors) of disease -Outbreak investigation is the real-world application of epidemiology

*****Ch. 26 - EMS with L-Tryptophan use in MN (Showa Denko manufacturer)

-Eosinophil is a type of white blood cell - important for parasites - someone who has a parasite would likely also have high white blood cell counts -Mention of toxic oil outbreak - 20,000 sick, 12,000 hospitalized in Spain - 315 died - victims all had EMS with fever, some had pneumonia-like illness - chemical used in industry that was added to rapeseed oil to be considered an industrial oil - dyed it blue and then used a high-temp refining process which makes the chemical very toxic -1) Verify the outbreak - EMS expert in MN recognizes cluster in New Mexico - more than what we would expect - common sense -2) Jump to hypothesis generation when they confirm that cases took L-tryptophan -3) Confirm diagnosis - couldn't do that here because we're working with a syndrome - there is no germ or lab test for this - can test for white blood cells, but that's to MAKE the diagnosis -4) Assemble the team - CDC multi-state investigation - Mike Osterholm from MDH State Epidemiologist took control -5) Case definition in MN - eosinophil count greater than 1,000 cells per microliter, muscle weakness, etc. - not quite the same as CDC case definition - MN included muscle weakness as well as myalgia - more sensitive -6) Case ascertainment - contacted rheumatologists (who might see patients with aching muscles and joints) to find cases - also there was a statewide pathology conference and they made an announcement there - also held press conference and had active surveillance (mailed to rheumatologists, dermatologists, etc.) - gender and age matched to avoid confounding because they had a lot of cases in women - used 1:1 matching ratio (higher ratio generally increases statistical power but more time consuming - diminishing returns after 4) -Controls - dialed number above and below case to find one -What did they ask on the questionnaire? L-tryptophan use month before onset, other medications (meds can cause EMS), uncooked pork and bear meat (parasites), asthma, eczema (causes EMS too), over the counter supplements -L-tryptophan is an amino acid (taken for insomnia) - maybe it has been contaminated in some way - not uncontaminated L-tryptophan on its own because amino acids are fine - unless you take a huge amount of it - can get L-tryptophan from eating turkey (this is why you get sleepy after), chickpeas, etc. (metabolized to produce serotonin) -7) Descriptive epi - main finding of the first case-control study - L-tryptophan was associated and all 12 cases took it, all 12 controls did not - OR undefined because denominator is 0 -Epi curve - small cluster in 1988 but peak is in August of 1989 - might have been a small contamination event in 1988, bigger one later -Median age 45 - 87% female, why? They are more likely to take L-tryptophan - or is there some reason women are more susceptible? Need to consider these things -8) Preliminary control - avoid L-tryptophan - said this at press conference - L-tryptophan users should contact their physician and MDH regardless of whether you got symptoms -Thinking ahead to find controls (took it but didn't get sick) -Asked L-tryptophan users if they had any left, what brand they used, and where they purchased it -Misclassification bias - when you put subjects in the wrong group, like cases in controls or vice versa - they tested blood of asymptomatic drug users (that would be potentially assigned as controls mistakenly) to determine white blood cell levels - found that none of the 18 were asymptomatic EMS - they were true controls -Selection bias - cases and controls were self-selected - they called MDH after press release - could they be different from a general sample? Might be SES differences, education differences, gender differences, etc. - those differences might matter -How could you deal with this in the design phase? True random sample or you could study the whole population (not realistic because we're working with a state outbreak/issue) -They couldn't minimize selection bias in this case because random sample wasn't possible -Community survey - wanted to see if prevalence of L-tryptophan use had increased - random digit dialing, 93% ended up participating, 4% of HHs had L-tryptophan user - showed increase in 80s and differed by gender -Trace back investigation - showed that particular manufacturer was involved (Showa Denko) - also tested for contaminants (supplemental analysis) - also found Peak E -2nd case-control study - 1st objective was to determine what was associated with being a case (hypothesis was L-tryptophan) - population was general population of cases and controls - 2nd case-control study population was only those who took L-tryptophan -Among those who have this, what can I determine that is more precise? In this case, what manufacturer is associated? -5 pragmatic criteria most useful for distinguishing causal and noncausal associations -Strength of the association - determined by magnitude of measure of association -Specificity in cause and effect - ideally, a given effect has a unique cause and vice versa - ex. Only L-tryptophan was associated -Consistency - persistence of association in multiple assessments - multiple studies should show similar results -Predictive performance - ability to predict unknown fact from the causal hypothesis - ex. If we recall, will it stop the outbreak? -Coherence (biologic plausibility) - extent to which a proposed causal association is consistent with existing knowledge -Biologic, factual, theoretic, and statistical coherence

Ch. 34 - Amish immunizations

-Outbreak of Pertussis a few years previous near Rock Island (NW Illinois, rural) - reporting suddenly stopped - person working w/ Amish may have said/did something they didn't like, so they stopped cooperating with the health department - left unanswered questions -Objective was to understand vaccination practices, prevalence, etc. - collecting data to inform possible interventions to increase vaccination - value of qualitative data to fill in the gaps where quantitative data is missing - also wanted to understand what was happening in this specific community to avoid making assumptions - Amish is a very large community -Hypothesis was that they don't want immunizations - turned out to be wrong -How did they get to the population? Personal connection through a relative to a local nurse -Medical committee - wasn't consistent with hypothesis of anti-vax - helped Amish access the healthcare system -Started designing a field study - brought to medical committee to propose - cross-sectional survey/questionnaire -Things to consider when creating a questionnaire - AEIOUCUP (Appropriate, Ethical, Intelligible, Omnicompetent, Unambiguous, Appropriately coded, Unbiased, Piloted) -Couldn't pilot this because the community was already wary - didn't want to offend anyone and lose the connection -Asked about vaccination status, reasons for why they weren't vaccinated - didn't ask names, addresses, etc. - anonymous - met with committee to go over the survey - they were concerned about information being published (agreed to discuss before) and if this was part of a government plan to get them to change vaccination practices in the community -Agreed to distribute it for them and that it would be anonymous -Found that they were not anti-vax - trusted healthcare provider was important to them - if they recommended it, they were likely to trust it -Douglas County Health Department didn't show much interest - 5 years later, Pertussis outbreak in the community - needed to improve prevention efforts -Talked to the community, communicated using their newspaper, set up a clinic for vaccinations, etc. - had good turnout - very successful in increasing vaccination and decreasing number of cases - used lessons learned from initial outbreak

History of regulating medicines

-Patent medicines in 1800s-1900s -Domac WWI German medic - no antibiotics for war wound infections -Farbenindustrie in Germany 1930s - medical/medicine research - found combo of chemical and dye (sulfa) -1930s Oklahoma physician - outbreak causing renal failure/death - Exlixir contained antifreeze --> Congress passed Food & Drug Act (modern FDA)

Steps: Decide to observe vs. perform ADDITIONAL studies

-Perform case control or cohort studies or additional lab testing

Steps: Develop a tentative CASE definition (revise as we learn)

-Person, place, time elements -Subcategories indicate status of a case and specificity -Confirmed (lab evidence) vs. probable (epi linked or meets case def except for no lab info - ex. fever, diarrhea but never got tested) -Primary vs. secondary (ex. Salmonella at a wedding buffet - guest or someone who eats the food is primary - exposed directly - vs. son who ate sandwich prepared by guest is secondary - fecal oral contamination)

Ch. 25 summary - What do people eat when they have no food (gaghra shak)?

-Preliminary investigation - case counting was difficult because of rural nature of villages -Study design that was logistically most efficient was enrolling all HHs in 2 key outbreak villages (assumed that whatever exposures led to cases would be the same everywhere) -Choosing right control group is difficult -Hypotheses that they tested in the epi study mainly came from the anthropological investigation -Info from patients or their families for those who had died -ID commonalities among all cases -Compiled list of possible causes - Pesticide use, packages foods, medicines, consumption of uncultivated plants (eating due to floods washing away crops) -Eating a number of foods was associated with a case in unadjusted analysis -Adjusted - gaghra shak was left - also low quality lentils -Which of the 2 was implicated? Turned to GS -Known to be toxic - toxin in seeds - saw in lit that it is associated with outbreaks in livestock -Larger proportion of cases reporting eating GS compared to lentils -Interpretation difficulty - only ½ of cases reported that they ate GS before becoming sick -Different from anthropological study -Confident that it was GS - Eating seedlings instead of mature plants (toxins are in seeds) - people were also eating GS as a main dish instead of as a seasoning - more toxic, especially for those with low bodyweight -Gabriella question - why isn't this a retrospective cohort study? Cohort by definition is the study of people over time - prospective is one where you ID exposures in the study population and then you follow people over time to see who develops the outcome - then you can calculate differences and rates between exposed and unexposed groups -Retrospective - same except that you aren't able to assess exposure and then follow over time chronologically - but you do have enough records where you can reconstruct who was exposed and unexposed historically, and then observe patterns of disease incidence today - didn't measure exposure before outcome chronologically, but it was measured elsewhere -Cohort is gold standard because you know the exposure happened before the outcome (causality) - no recall bias about if exposures were in the past because they are already documented/measured -So they didn't have any records - they didn't measure exposures and follow over time -They measured exposure and outcome at the same time - traditionally called cross-sectional - closer to CS, but those are typically a random sample of a population -Gabriella question - thought about case-control vs. cohort - mentioned time pressure and severity of disease - if the situation was different (more time, less severe), is there anything you would have done differently? -Cohort still wouldn't work because the outbreak was over - couldn't observe people over time (plants were growing so they weren't toxic anymore) -Case-control could have worked - likely would have had multiple control groups -Might want controls that are age and sex matched (we know that there were more women and children cases - their behaviors could be different from others in the population due to their age and sex) - match on important confounders -Might also want random set of controls - make comparisons - maybe more women and children cases simply because this is what the population looks like (not very likely) - then among all cases, how are they different compared to age and sex matched group from the same community

Music example at a club - what modes of transmission are enhanced?

-Respiratory droplets - have to shout -Close proximity -Fomite - contact surfaces (doorknobs) -Airborne - likely not good ventilation -Physical contact/sexual transmission -Foodborne - hand hygiene

Steps: VERIFY that there is really an outbreak

-Review existing data -Compare # of cases to historical data from surveillance databases if available -If it's a disease we don't know much about or something like diarrheal disease (no surveillance), use common sense and make a judgment call about if it's an outbreak

Waterborne Disease Outbreak Investigation - Perform Additional Studies? Selection of Controls?

-Scenario - Aug. 13th - IDPH notified of cluster of 3 cases of diarrheal illness - visited local water park in central IL - crypto identified in stool samples -Need to think about different factors of the water park and behaviors - Lazy river that ran around the entire water park - also had a splash pad for very young children with very shallow water - also pool for laps - also a big bucket that dumps water on kids below -Cases are ascertained - descriptive epi of 100+ cases thanks to things like media alerts, notifying local hospitals -Are additional studies indicated? What should be considered in making this decision? -Sound hypothesis, public interest/political (ex. Kids at risk), severity, whether it is ongoing or will further investigation control it (in this case it is because ascertainment is ongoing), magnitude (high in this case, already 100+ cases), is it novel or unique, is there public willingness to continue/apathy (ex. Is everyone better already?), are there resources (in this case, resources came from IDPH rapid response team), will recall bias be a problem? -If you pursue additional studies, what might you perform? -Hypothesis is that the outbreak is caused by the water park - then 2nd study would be looking at what individual activity/behavior is the cause within the water park -To determine the source of crypto in the community, do a case control study in the community - hypothesis that Splash Town is the source -Who would controls be? Cases are people who are sick with crypto (don't want to include water park because we need to compare it in analysis) - controls are people who aren't sick living in or visiting Central Illinois between 7/1-8/31 -Ex. Peoria County (water park) Taswell County (where investigation is, where hospital is with cases) -Should there be matching? Yes - water park is important for age - could do exact age match (ex. Case is 5, control is 5) or age range (ex. Case age +/- 5 years) - wouldn't make sense to match on sex - might want to match on neighborhood (way around people maybe having similar incomes in case park charges for a fee) -Additional epi study to perform? Once you confirm park is associated, you would do a case control study at the actual park -What specific risk factors might be examined among water park attendees? Lazy river (not likely to be a source because you probably aren't ingesting water), splash tower, food stand at the water park, any accidental swallowing of water, putting their head underwater, any symptoms from before visiting, contact with animals (unlikely because it is fenced in), diapering infants, infant splash pad -Results of study: -Pool water in mouth was 89% for cases, 63% of controls, OR 6.0 (1.3-26.8) so statistically significant -Swallowed water OR 4.5 significant -Head under water OR 3.3 borderline significant -Swimming 2.5 but not significant -If you performed the above additional CC study examining risk factors in the park, how would you define controls? Park attendees who are not ill - had lab confirmed and clinical case for their case definition -Let's say you have 180 cases for this study - how many controls should we select? Since there are so many cases, we might be able to use 0.5 controls per case for feasibility reasons - but if we have the resources, we might as well - but we shouldn't include sporadic cases from before the peak of the epi curve - should we include those from after the peak of the epi curve? Secondary transmission is possible, so no

Outbreak investigation - a high stakes game

-Scrabble: more lit you know about organisms/outbreaks beforehand, the more prepared you will be when it happens - good to go to lit before going into the field to see what they learned and did -Monopoly: unexpected things happen - be flexible -Craps: Want to have a good sense of odds and risk - don't bet on long shots (hypothesis-driven) -Chess: Think ahead, strategize - what are the consequences of your moves? -Ex. restuarant outbreak - shut down restuarant --> when will we re-open? need plan to respond

Steps: Confirm the DIAGNOSIS

-Send samples to state lab to confirm (ex. Salmonella) - double check that diagnosis from a hospital, for example, is correct -Culture, species, serotype, or strain -Serology -Toxin identification **This doesn't always happen - ex. syndromic, can't test for it - might just MAKE a diagnosis - might not send samples to a state lab or CDC

Ch. 4 Cholera in Portugal summary

-Summary of 4 investigations (all case-control) -1st) Lisbon case-control at the tail end of the epi curve - focused on most recent cases instead of the peak to avoid potential recall limitation (everyone struggles equally with recall) - didn't find anything significant - big failure, why? Tail end means that any primary sources are probably not left - also no questions about bottled water on the questionnaire from the government which means he launched a case-control study without a sound hypothesis -2nd) Southern district of Faro - Tavira - matched case-control study - this time early cases - found association with drinking water from Fonte do Bispo - pipe emerging from the side of a hill from a concrete structure - locals liked the taste of the water - construction blast could have ruptured sewer line - limestone is porous - biologic plausibility is that a river nearby that could have carried sewage from military training base (cholera imported from Angola) -3rd) Same district, but shifts to a different city - case-control - this time at the peak - raw or semi-cooked cockles associated - biologic plausibility is that cockles are filter feeders -4th) Back to Lisbon - case-control again - bottled water (Agua do Vimeiro) was associated - also showed that earlier data collected proved this too, but they didn't analyze that at the time (1st study) - biologic plausibility for bottled water causing cholera outbreak is that 2 springs north of Lisbon were polluted (limestone can have fissures so dirty water can make its way into the supply and into the bottling plant) - river tested positive for cholera -1st step - verify that an outbreak is happening - not illustrated in this chapter because he came in after - retrospectively, they confirmed because there was no cholera happening in Portugal before that -2nd step - confirm the diagnosis - sending sample to state lab to double check that diagnosis is correct (ex. Hospital tests positive for Salmonella, send sample to state lab to make sure) - this isn't illustrated here because he came after it would have been done -3rd step - assemble the team - Director General of Health in Portugal, CDC Mark Rosenberg, local nurses, sanitarian -4th step - tentative case definition - Portugal would have their own case definition they would have used for cholera (ex. From WHO) - this would have already existed - he may have used a more specific case definition for each case-control study he did (ex. Any culture-confirmed case from Lisbon diagnosed on or after 9/13) -5th step - ascertain cases - implied in this step that we would have created a survey tool or questionnaire - went door to door after talking to cases to get controls -6th step - descriptive epi -Figure 4.1 - geography/spread of the outbreak - hospitalized cholera patients April-October - hospitalized is usually not very representative because cholera usually doesn't hospitalize people - but this is a new organism for this population, so it makes sense that more people would become more ill (vs. somewhere where cholera is endemic) - figure shows most cases in Faro and Porto - makes sense that he went to Faro - Porto is in 3rd highest, could have gone there next -Figure 4.2 - epi curve - patients by date of hospitalization - peak was in late August/early September - important to consider that he arrived in early September, after the peak - makes it more difficult to figure out -7th step - supplemental lab/environmental investigation - something beyond just interviewing cases - ex. Testing water samples, serotyping to ID strain or get molecular fingerprint - wasn't indicated in this chapter -8th step - develop hypothesis - often happens earlier - 1st Lisbon questionnaire included possible risk factors like snails, lettuce, beans, watercress, raw shellfish, well water, antacids (cholera is sensitive to acid, so those who have lower stomach ph are more likely to get sick in general and more likely to get more seriously ill), gastric surgery (can remove part of stomach that creates stomach acid which would make someone more susceptible to cholera) -Asks cases about exposures 5 days prior to onset of illness - why 5 days? Incubation period for cholera is 2-5 days - uses higher end to make sure he doesn't miss anything which is sensitive, not specific -Control selection - went to cases and used schematic to go door to door - also age and gender matched - controls didn't get sick, so we can't ask about exposures since they got sick - his approach was 5 days before the interview - less problematic because we're looking at recent cases, but potentially an issue because time periods for cases and controls will probably be different -Ends up changing things in the questionnaire as he goes - can be a problem with analysis because not all questionnaires are asking the same thing - need to subset the data -Bias concerns - one of the nurses was over-enthusiastic and leading cases to answer questions a certain way - also neighbor controls were a problem - initial Portuguese studies used controls who were conveniently accessible instead of using something systematic → selection bias - this is why he used systematic way of selecting controls

Steps: Assemble investigation TEAM

-Team members will depend on type of outbreak

Steps: REVIEW rules and regs for the disease being investigated

-There are for Salmonella, aren't for diarrhea -Ex. Salmonellosis - need to work w/ food handlers to prohibit them from working again until 2 consecutive stool samples are negative -Putting out the fire on the outbreak and also testing food handlers

Primary activities/steps during an investigation (16: VDT CADS HPA RAC DCP)

-VERIFY that there is really an outbreak -Confirm the DIAGNOSIS -Assemble investigation TEAM -Develop a tentative CASE definition -ASCERTAIN/find cases -DESCRIPTIVE epi analysis -Perform SUPPLEMENTAL lab or environmental investigation -Develop HYPOTHESES -Introduce PRELIMINARY control measures -Assess ADEQUACY of control measures -REVIEW rules and regs for disease being investigated -Decide to observe vs. perform ADDITIONAL studies -CONTINUE ascertaining additional cases during and after conclusion of investigation -DRAW conclusions and formalize recs -COMMUNICATE to those who need to know the info -PLAN for field work

Ch. 14 - crypto outbreak in MKE

-Verify outbreak - heard from EDs that there were unusually large numbers of diarrheal disease cases - also heard from individuals in the community - also labs were being asked to do lots of testing suddenly for bacterial culture of stool specimens *No surveillance for crypto at this time - used common sense - surveillance for crypto emerged as a result of this -Then they respond by doing some active surveillance - calls to multiple ERs and the state to determine the geographic extent of the outbreak - helps to determine resources and for assembling the team/determining who needs to be in charge - mainly southern region (used for hypothesis generation) -Confirm the diagnosis - early on it was just diarrheal disease so they can't really confirm anything - bacterial cultures were negative so no common bacterial cause - normally would test for a virus next (ruled out because of how widespread it was) -Hypothesis generation - waterborne transmission is their initial hypothesis because it is so widespread (consumption of water is widespread) and there is a water plant in the south side - thinking about foodborne, could be milk but that wouldn't make sense because of the geographic clustering -Assemble the team - DOH, MHD - 9 teams because the outbreak was so large -Case definition - no surveillance = no existing case definition that can be used/adjusted for the outbreak - created diarrhea surveillance -Clinical cases (presence of watery diarrhea - March 1-April 9 when plant was closed, then extended to April 28) and lab confirmed (crypto identified in stool) cases because they're confident by now that they're dealing with crypto - clinical very sensitive -Case ascertainment - focused on nursing homes and EDs - smaller part of the surveillance to not overwhelm the system - picked nursing homes because they were geographically fixed and not likely to get water from other sources - susceptible population that doesn't have a ton of exposures - EDs focus on more severe cases and collect lots of data - set this up in 4 counties to monitor rise and fall of cases and determine if any control measures are successful -Preliminary control measures - they do this even before they have their case definition - boil water advisory - other option was treating water with chlorine, but crypto is highly chlorine resistant so the amount needed would be much higher which would be unsafe for drinking -Water boiling can be unsafe, increases energy consumption, requires education -Also political - asked mayor if he would drink the water and he said no - went with boil advisory ultimately -Potential environmental sources - 3 groups: industry (meat packing), agriculture (cows), and wildlife (infected deer - feces could be carried away by rains) -33% tested positive vs. 29% pre-outbreak - why? Denominator is much different during the outbreak, everyone was going to get tested -Did a lot of supplemental lab and environmental testing - water plants, ice for ice sculptures -HIV case control study - patients weren't more likely to get crypto, but they were more likely to have severe disease - amount of crypto needed to make you sick is pretty small but once you get it, if you have another illness like HIV/AIDS you are more likely to get very ill -Delphi survey - want to get expert opinion about when to re-open the south water plant -Recalled some foods where water might have been used in preparation - ex. Pre-washed lettuce

Ch. 19 - Whooping cough

-Verifying the outbreak - used surveillance data to confirm -Confirm the diagnosis - making a diagnosis in the first place is hard (can do with PCR, serology, clinically, culture - but each has issues - culture is gold standard but it often doesn't grow - clinical is difficult because symptoms are similar to other illnesses like common cold - serology hasn't been well standardized - PCR is sensitive = false outbreaks) -Cases are usually dismissed in early stages - common cold -Clinical diagnosis usually not until whooping stage but not everyone whoops -Lab confirmation has highest yield early, before significant clinical symptoms begin -Case definition - clinical vs. lab - clinical: cough 2+ weeks w/ one of the following - lab: isolations of pertussis or positive PCR -Outbreak case definition - 2 weeks of cough without attributable cause -Probable vs. confirmed -Team - rapid response team at IDPH - vet, sanitarian, microbiologist, infection control nurse, data systems expert, state epi, etc. - lots of political pressure so the governor funded it -Objectives of a community outbreak of a disease like pertussis -Identify cases - find as many as we can to stop the spread to treat them and their close contacts = control the outbreak - surveillance is information for action -Introduce preliminary control measures -Case ascertainment - how was surveillance increased? -Increase reporting with list of stakeholders (local labs, hospitals, day care centers, etc.) -Memo - descriptive epi with age range, number of cases - gives guidance about who should seek treatment - medication adherence is important because otherwise it will continue spreading -Also did personal visits to talk about importance of surveillance -Lab reporting - talked to lab workers -Chart review to try to find more cases and chains of transmission -Surveyed local pharmacies to ask about increased sales of cough medicines -Implement control measures - 4 main measures -Prophylaxis for close contacts -Isolation, cover your cough, wear a mask for those who are sick -Encourage vaccination -Treatment of cases -Descriptive epi - older people whoop too - area had waning immunity among adults -Age range was very wide - not just a baby disease

THings to consider when deciding whether to continue performing studies

-Whether it's already controlled or if further investigation will lead to control (if we're still seeing cases come in, continue) -Resources - do we have enough? -Is there something novel/unique about the outbreak? If there is something to learn, continue - advances knowledge/science -Public or political influence -Willingness of cases to continue/apathy - ex. might not want to continue participating (questionnaire) -Severity of illness -Whether or not you have a sound hypothesis -Will recall bias be too hard to overcome? Ex. foodborne outbreak hard to remember after much time has passed