Managing Outbreaks: The 10 Essential Steps

1
Managing OutbreaksThe 10 Essential Steps
Dick Zoutman, MD, FRCPC
Hosted by Paul Webber paul_at_webbertraining.com
www.webbertraining.com
2
Lecturing is.

• Dr. Dick Zoutman
• Professor Chair
• Divisions of Medical Microbiology and of
  Infectious Diseases
• Queens University at Kingston
• Medical Director of Infection Prevention
  Control, Kingston Hospitals
• zoutmand_at_kgh.kari.net

3
Goals

• The Goals of the Outbreak Module are to
• Equip the student with the basics to undertake an
  outbreak investigation
• Provide key definitions used in describing
  outbreaks
• Demonstrate and review use of biostatistical
  methods helpful in managing outbreaks
• Provide some real life examples of outbreaks and
  approaches used in solving them

4
Specific Objectives

• To define and describe the 10 essential steps in
  outbreak management
• 1. Preparation in advance for outbreak management
• 2. Confirm the presence of an outbreak
• 3. Verify the diagnosis
• 4. Identify Cases
• A. Establish a case definition
• B. Case finding
• 5. Perform descriptive epidemiology
• 6. Develop hypotheses
• 7. Test hypotheses
• 8. Refine hypotheses and do additional studies
• 9. Implement control prevention measures
• 10. Communicate findings and conclusions

5
Definition of an Outbreak

• Occurrence of more cases of disease than expected
• Nosocomial outbreak
• any group of illnesses of common etiology
  occurring in patients of a medical care facility
  acquired by exposure of those patients to the
  disease agent while confined in such a facility

6
Reasons to Investigate

• Control/prevention
• Research opportunities
• Training
• Public, political, or legal reasons
• Program considerations

7
The Biggest Pitfall!

• Communication
• Communication
• Communication
• Communication
• Communication
• Communication
• Communication
• Communication

8
10 Steps in an Outbreak Investigation
9
1. Preparation

• Advanced preparation
• Have protocol for managing outbreaks developed
  BEFORE
• Clearly defined command and control policies
• Communication protocol
• Notification and fan outs
• Scientific knowledge
• Review literature /or regulations before you
  embark
• Consult experts
• Sample questionnaires ready to adapt
• Supplies
• Consult with laboratory about samples
• Equipment
• Laptop for data entry and notes, digital tape
  recorder for meetings

10
Preparation cont.

• Administration to assure resources, staffing,
  funding
• Travel arrangements as needed
• Approvals for travel, cross coverage
• Personnel matters, extra staffing needs
• Consultation-make sure you know your role and its
  parameters
• Lead investigator or just lending a hand?
• Know who to contact at investigation site or an
  ward/care unit

11
2. Establish Existence of an Outbreak

• Is an outbreak truly occurring?
• True outbreak vs
• A pseudo-outbreak?
• Sporadic and unrelated cases of same disease
• Unrelated cases of similar unrelated disease
• Determine the expected number of cases before
  deciding whether the observed number exceeds the
  expected number!

12
Establish Existence of an Outbreak, cont.

• Comparing observed with expected
• through surveillance records for notifiable
  diseases
• hospital discharge data, registries, mortality
  statistics
• Infection control data
• data from other facilities, province, surveys of
  health care providers
• community survey

13
3. Verify the Diagnosis

• Ensure proper diagnosis and rule out lab error as
  the bias for increased diagnosis
• Review clinical findings, lab results with
  knowledgeable expert
• Summarize clinical findings with frequency
  distributions
• Characterize spectrum of disease
• Verify diagnosis
• Develop case definition

14
Verify the Diagnosis, cont.

• See and talk with patients if at all possible
• Better understand clinical features
• Mental image of disease and the patients affected
• Gather critical information
• Source of exposure
• What they think caused illness
• Knowledge of others with similar illness
• Common denominators
• Helpful in generating ideas for hypothesis about
  etiology and spread
• Keep notes!

15
Process Control Charts

• Many different kinds in the statistical process
  control literature
• Since much of infection control is binomial only
  2 possible outcomes, Yes or No
• We can use binomial mean and standard deviation
  calculations

16
The Binomial Distribution

• Not all data are normally distributed
• Binomial data has the following characteristics
• There are n identical cases or trials.
• Each case or trail has only two possible outcomes
  
• (eg infection or no infection, boy or girl,
  yes or no etc).
• The probabilities of the two outcomes remain
  constant.
• The cases at risk (or trials in statistical
  language) are independent of each other.

17
Mean and Standard Deviation of the Binomial
Distribution

• The mean proportion and standard deviation of a
  binomial distribution are
• where n is the total number of trails (cases),
• p is the proportion of occurrences (in our case
  the item of interest ie an infection)

18
Upper Lower Control Limits
3 standard deviations
19
Sample SSI Data
20
Calculate Mean Standard Deviation
21
UCL and LCL
22
P-Chart
UCL
p
LCL
23
Control Limits for Control Charts

• A process is said to be out of control if
• One point falls above upper control limit(gt3?) or
  below lower control limit (lt3?)
• Two of three consecutive points gt2? but lt3? on
  one side of the mean
• Four of five consecutive points gt1? but lt2? on
  one side of the mean
• Nine consecutive points on one side of the mean
• Six consecutive points increasing or decreasing
• Fourteen consecutive points alternating up or
  down
• Fifteen consecutive points within 1? above or
  below the mean

24
Process Control Chart
25
Process Control Charts Out of Control
26
4a. Establish a Case Definition

• Case definition
• Standard set of criteria for deciding whether an
  individual should be classified as having the
  health condition of interest
• Includes clinical criteria and restrictions by
  time, place and person
• Must be applied consistently and without bias to
  all persons under investigation
• Must not contain an exposure or risk factor you
  want to test

27
4a. Establish a Case Definition, cont.

• Classification
• Definite (confirmed)
• Laboratory confirmed
• Probable
• Typical clinical features without lab
  confirmation
• Possible (suspected)
• Fewer of the typical clinical features
• Some times use Definite one other
  classification

28
4a. Establish a Case Definition, cont.

• Early in investigation may use a loose case
  definition
• Better to collect more information than necessary
  so you dont need to make repeat visits
• Identify extent of problem and population
  affected
• Generating hypotheses
• Later when hypotheses are sharpened investigator
  will tighten case definition

29
4b. Identify and Count Cases

• Target health care facilities/units where
  diagnosis likely to be made
• Enhanced passive surveillance e.g. letter
  describing situation and asking for reports
• Active surveillance e.g. phone or visit
  facility/unit to collect information
• Best always go in person!
• Alerting the public
• Media alert to avoid contaminated food product
  and seek medical attention if symptoms arise
• Hospital or facility public relations to develop
  media release

30
4b. Identify and count cases, cont.

• As far as possible identify entire population at
  risk and attempt to survey as many as possible
• Survey entire population if it is manageable
• Case finding will continue throughout the
  outbreak
• Always ask case-patients if they know of any
  others ill with the same symptoms
• Eg friends, relatives, other contacts
• In hospital perform contact tracing using
  admission and registration data

31
4b. Identify and count cases, cont.

• Information to be collected about every case
• Identifying information
• Re-contact if additional questions come up
• Notification of lab results and outcomes of
  investigation
• Check for duplicate records (health record or SIN
  helpful)
• Map geographic extent
• Demographics
• Provide person characteristics for defining
  population at risk

32
4b. Identify and Count Cases, cont.

• Information to be collected about every case
  cont.
• Clinical findings
• Verify case definition met
• Chart time course
• Supplemental dates e.g. deaths
• Risk factor information
• Tailored to specific disease in question
• Reporter information
• Id of person making report can be useful

33
4b. Identify and Count Cases, cont.

• Collection forms
• Standard case report form
• Questionnaire
• Data abstraction form
• For common events have these prepared ready in
  advance (from past events)
• Line listing
• Abstraction of selected critical items from above
  forms
• Keep it up to date several times a day
• Contains key information only
• Spread sheets useful for this purpose

34
5. Perform Descriptive Epidemiology

• After collection of data characterize the
  outbreak by
• Time
• Place
• Person

35
Time

• Epidemic curve
• Histogram of the number of cases by their date of
  onset
• Visual display of the outbreaks magnitude and
  time trend
• Where you are in the time course of the outbreak
• Future course?
• Probable time period of exposure
• Helps in development of questionnaire focusing on
  that time period
• Common source vs. Propagated differentiation

36
Epidemic Curve
37
Place

• Geographic extent of problem
• Clusters or patterns providing important
  etiologic clues
• Spot maps
• Where cases live, work or may have been exposed

38
Person

• Determine what population at risk
• Usually define population by host characteristics
  or exposure
• Use rates to identify high-risk groups
• Numerator number of case
• Denominator number of people at risk

39
Epidemic Curves

• How to set it up
• What it tells you
• Mode of transmission
• Propagated
• Common source
• Timing of exposure
• Course of exposure

40
Epidemic Curves Useful

• Determine whether the source of infection was
  common, propagated (continuing), or both
• Identify the probable time of exposure of the
  cases to the source(s) of infection
• Identify the probable incubation period
• Determine if the problem is ongoing

41
Common Source Outbreaks

• A common source means that all cases have the
  same origin
• The same person or vehicle is identified as the
  primary reservoir or means of transmission
• One source infects a generation of contacts
• With a common source outbreak the epidemic curve
  approximates a normal distribution curve
• if there is a sufficient number of cases and if
  cases are limited to a short exposure with
  maximum incubation of a few days or less (point
  source)
• Exposure may be continuous or intermittent (water
  vs contaminated air handling system)
• Intermittent exposure to a common source produces
  a curve with irregularly spaced peaks

42
Creating An Epi CurveExcel or PowerPoint
43
Epidemic Curves
Common source point exposure (e.g., salmonella)
Source of salmonella
Note short time frame (days)
44
Epidemic Curves
Common source Intermittent exposure (e.g.,
contaminated blood product)
45
Propagated Source

• A propagated source means that infections are
  transmitted from person to person
• cases identified cannot be attributed to agent(s)
  transmitted from a single source
• One person gives to many others who give to yet
  more people (eg. SARS, influenza)
• Propagated (continuing) source cases occur over a
  longer period than in common source transmission
• Explosive epidemics due to person-to-person
  transmission may occur (i.e., chickenpox)
• If secondary and tertiary cases occur, intervals
  between peaks usually approximate average
  incubation period

46
Epidemic Curves
Propagated source single exposure, no secondary
cases (e.g., measles)
This unvaccinated cohort all exposed together.
Disease did not go further due to vaccinated
population
47
Epidemic Curves
Propagated source secondary and tertiary cases
(e.g., hepatitis A)
Two Waves of infection
Mean Incubation Period
48
Estimating Date of Exposure Rubella Outbreak of
Unknown Source
Peak
Rubella incubation 18 days
Probable time of exposure
49
Estimating date of exposure Rubella Outbreak
Maximum incubation 21 days
Probable time of exposure
Minimum incubation 14 days
50
6. Develop Hypotheses

• Hypotheses should address
• Source of the agent
• Mode of transmission
• Vector or vehicle
• Exposure that caused disease

51
Develop Hypotheses

• Generating the hypothesis
• What do you know about the disease?
• Reservoir, transmission, common vehicles and
  known risk factors
• Talk to several case-patients
• Use open ended questions
• Ask lots of questions
• Talk to local health department staff or
  hospital/facility staff
• Use descriptive epidemiology e.g. epi curve

52
7. Test Hypotheses

• Evaluate the credibility of your hypotheses
• Compare with established facts
• When clinical, lab, environmental and/or epi data
  undoubtedly support hypothesis
• Use analytic epidemiology to quantify
  relationships and explore the role of chance
• Cohort studies
• Case control studies

53
Analytic Studies

• Two types used in outbreak investigations
• Cohort
• Case-control

54
Definition of a Cohort

• In epidemiology, Any designated group of
  individuals who are followed or traced over a
  period of time.
• - Last, JM. A Dictionary of Epidemiology, 3rd
  ed. New York Oxford University Press, 1995

55
Cohort Studies
Study Population
Exposure is self selected
Non-exposed
Exposed
Follow through time
Disease
No Disease
No Disease
Disease
56
Cohort StudiesProspective vs. Retrospective
57
Cohort Study

• Preferred study design when
• Members of cohort are easily identifiable
• Members of a cohort are easily accessible
• Exposure is rare
• There may be multiple diseases involved

58
Cohort studies start with an exposure and go
forward to diseases.
Drawing by Nick Thorkelson
59
7. Test Hypotheses cont.

• Cohort
• Small, well defined population of exposed
  individuals
• Contact each attendee and ask a series of
  questions
• Ill Vs not ill
• Look for source exposure
• Attack rate is high among those exposed
• Attack rate is low among those not exposed
• Most of the cases were exposed, so that the
  exposure could explain most, if not all, of the
  cases
• Relative Risk measure of association between
  exposure and disease (Risk of disease)

60
Case-control studies start with a disease and go
back to exposures.
Drawing by Nick Thorkelson
61
7. Test Hypotheses cont.

• Case-control
• Population not well defined
• Case patients and control group questioned about
  exposure(s)
• Compute measure of association Odds Ratio (odds
  of being exposed)
• Quantify relationship between exposure and
  disease

62
8. Refine hypotheses and do additional studies

• Epidemiologic
• When analytical epidemiology unrevealing need to
  reconsider your hypotheses
• Go back and gather more information
• Conduct different studies
• Laboratory
• Additional tests
• Environmental studies

63
9. Implement Control /Prevention Measures

• Implement control measures as soon as possible
• Do not wait to second last step!
• May be aimed at agent (eg vaccines), source
  (sterilization), or reservoir (isolation)
• Short or long term measures

64
10. Communicate The Findings

• Orally within facility/community
• Local health authorities and persons responsible
  for implementation of control and prevention
  measures
• Written reports (consider publication) for
  planning, record of performance, legal issues,
  reference, adding to knowledge base

65
10 Steps in Outbreak Management

• 1. Preparation in advance for outbreak management
• 2. Confirm the presence of an outbreak
• 3. Verify the diagnosis
• 4. Identify Cases
• A. Establish a case definition
• B. Case finding
• 5. Perform descriptive epidemiology
• 6. Develop hypotheses
• 7. Test hypotheses
• 8. Refine hypotheses and do additional studies
• 9. Implement control prevention measures
• 10. Communicate findings and conclusions
• These 3 steps should happen throughout the
  outbreak

66
GI Outbreak Among Staff in a Hospital
67
Objectives

• To describe
• An outbreak investigation in a hospital
• An example of a cohort study
• 10 steps in outbreak investigation
• Epidemiology of GI infections

68
December 15 Illness in 28 of 60 staff on a
Medical In-Patient Unit

• Predominant symptom vomiting
• Onset early AM of December 15
• Physicians, nurses, residents
• Staff had eaten three meals in common
• Dec 12 catered meal of Heavenly Ham
• Dec 13 Christmas pot-luck
• Dec 14 food from local sandwich take out

69
Step 1 Advanced Preparation

• Lets hope!
• Investigation
• Appropriate scientific knowledge, supplies,
  equipment

70
Step 2 Establish the Existence of an Outbreak

• Occurrence of more cases of disease than expected
  in a given area or among a specific group of
  people over a particular period of time

71
Vomiting as a Chief Complaint

• Possibilities
• Viral gastroenteritis
• Rotavirus (infant)
• Norovirus (older child / adult)
• Food poisoning due to pre-formed toxin
• Staphylococcus aureus
• Bacillus cereus
• Non-infectious (Sb, As, Cd, Cu, Fl, Zn, etc.)

72
Incubation Periods for Suspect Infectious Agents
73
Step 1 Prepare for Fieldwork

• Administration
• Make coverage arrangements
• Put your phone on call forward
• Put batteries in your pager
• Consultation (roles)
• Collaboration on all steps
• Occupational Health, Local Public Health, Union

74
Step 3 Verify the Diagnosis

• Through effort of the Medical Microbiologist
• Routine stool cultures submitted through the
  hospital
• Stool for Norovirus submitted to the public
  health lab/virology lab for PCR
• C. difficile toxin assay?

75
Step 4 Establish a Case Definition, Identify
and Count Cases

• 4a) Establish a case definition
• Initial case definition
• persons employed by or assigned to the Medical
  Unit who called in sick on December 15
• 4b) Identify and count cases
• Twenty-eight individuals were identified
• Determined that NO patients involved!

76
Step 3, 5 and 6

• Verify diagnosis.
• Do descriptive epidemiology and develop
  hypotheses.

77
Open-Ended Interviews N10 Persons Who Called in
Sick

• Verify diagnosis
• Symptoms
• Sudden onset of profuse vomiting and diarrhea
• Systemic symptoms, including headache,
  arthralgias, myalgias, weakness
• Recovery (or near recovery) 12 hours

78
Open-Ended Interviews N10 Persons Who Called in
Sick

• Descriptive Epidemiology
• Onset late on December 14 early AM and morning
  of December 15

79
Open-Ended Interviews N10 Persons Who Called in
Sick

• Hypothesis generation
• No common events outside of work
• Attendance at
• Monday luncheon (Dec 12) 3 (30)
• Christmas pot luck (Dec 13) 10 (100)
• Wednesday lunch (Dec 14) 7 (70)

80
Step 7 Evaluate hypotheses

• You obtained the menu for the Christmas luncheon
• Questionnaire has been constructed
• Interview of a convenience sample
• Recovered / well individuals on-site
• Fellow ICPs and/or supervisors helping with this
• Ill individuals contacted by phone
• Occupational Health

81
Back to Step 4

• Case individual in attendance at the Christmas
  party luncheon (Dec 15) with illness
  characterized by vomiting or two or more episodes
  of diarrhea, and onset on or after December 14
• Control individual in attendance at the
  Christmas party luncheon with no symptoms of
  illness the week of December 14-20

82
And Back to Steps 3 and 5 Verify the diagnosis
and perform descriptive epidemiology

• Interviews allow refinement of
• Descriptive epidemiology outbreak curve (time)
• Diagnosis

83
Epidemic Curve
84
Study Population

• 39 interviews
• Exclusions
• 1 person ill, but did not meet the case
  definition
• 1 did not attend the dinner
• 3 had onset prior to December 14
• Final population N 34
• 16 cases
• 18 controls

85
Characteristics of Illness (N16)

• Aches
• 10 (62)
• Chills
• 10 (62)
• Cramps
• 12 (75)
• Diarrhea
• 13 (72)
• Avg. 5.25 episodes

• Headache
• 11 (69)
• Nausea
• 13 (81)
• Vomiting
• 13 (83)
• Avg. 5.3 episodes
• Fever
• 4 (25)

86
Step 7 RR of illness for the exposure candied
sweet potatoes
RR 0.69 95CI 0.13 to 3.56 p1.0
87
Using the EpiMax Table Calculator

• http//www.healthstrategy.com/epiperl/epiperl.htm
• On line multi-purpose statistical calculator
  (free!)
• Enter in 4 values of a 2 x 2 table and it
  calculates Chi Square p value, 95 confidence
  intervals
• Among many other things

88
Risk vs Odds

• Risk
• Refers to the risk of getting the disease or
  infection
• Odds
• Refers to the odds of being exposed to a
  particular risk factor

89
Measures of Effect
Risk Ratio Rate Ratio Compares the
incidence of disease (risk) among the exposed
with the incidence of disease (risk) among the
non-exposed (reference) by means of a
ratio. The reference group assumes a value of
1.0 (the null value)
Relative Risk (RR)
90
Relative Risk
If the relative risk estimate is gt 1.0, the
exposure appears to be a risk factor for
disease. If the relative risk estimate is lt
1.0, the exposure appears to be protective of
disease occurrence.
91
Relative Risk
Hypothesis Prostitutes who also use injection
drugs are more likely to have HIV
infection Results Of 2,240 prostitutes not
using injection drugs, 16 have HIV. Of 840
prostitutes using injection drugs, 10 have HIV.
Note that the row and column headings have been
arbitrarily switched from the prior example.
92
Relative Risk
Hypothesis Prostitutes who also use injection
drugs are more likely to have HIV
infection Results Of 2,240 prostitutes not
using injection drugs, 16 have HIV. Of 840
prostitutes using injection drugs, 10 have HIV.
RR IE / IE-
RR (10 / 840) / (16 / 2,240)
RR 0.0119 / 0.0071 1.68
93
How Did He Do That?
http//www.healthstrategy.com/cgi-bin/epiperl.pl
94
Step 7 RR of illness for the exposure chocolate
cake
RR0.97 95 CI 0.46 to 2.03 p0.78
95
Step 7 RR of illness for the exposure seafood
jambalaya
RR1.45 95 CI 0.70 to 2.98 p0.50
96
Step 7 RR of illness for the exposure Christmas
punch
RR4.9 95 CI 1.32 to 18.25 p0.004
97
Step 8 Refine the hypothesis

• How could the punch have become contaminated?

98
Other Data in Speaking to Medical Officer of
Health

• Local Nursing home outbreak (same week)
• Onsets consistent with person-to-person spread
• Background illness
• Community
• Family Practice Center

99
How was Christmas punch made?

• Bottled grape juice
• Unsweetened canned pineapple juice
• Sprite
• Homemade ice rings
• Water
• Sliced fruit
• Sliced fruit
• Mixed in bowl found on top of refrigerator

100
Homemade Ice Rings

• Ice ring from household A
• City water
• Person who made it had GI distress the day of
  the event

• Ice ring from household B
• Well water
• All members of family of this household
  sequentially had similar illness over the
  previous month!

101
Step 9,10 Control Measures / Communication

• Contacted Medical Director of IC Program and
  Medical Officer of Health
• Preliminary results of analysis suggest Christmas
  Punch is the most likely culprit
• No evidence for contamination of commercial food
  products
• Recommend exclusion of ill persons and good
  hand-washing

102
Timeline

• Friday, December 16
• Notification approximately 300 PM
• Open-ended interviews
• Study design
• Saturday, December 17
• Interviews using a standard questionnaire
• Data entry
• Analysis completed 1030 PM
• Sunday, December 18
• Phone interview of persons who made the punch
  800 AM
• Preliminary results shared with the Medical
  Officer of Health and IC Med Director 1030 AM

103
Step 8 The lab gets the last word

• Environmental Specimens via Health Unit
• Water sample from kitchen tap of household B
• () total coliforms
• () E coli
• Eueww!

104
Step 8 The lab gets the last word

• Human Specimens
• 12 stool specimens
• Negative for Salmonella, Shigella, Yersinia and
  Campylobacter in the clinical laboratory
• 10 stool specimens
• PCR positive for NLV at public health lab
• Identical nucleotide sequence (months later)

105
Conclusion

• Christmas punch was the source of an outbreak
  affecting approximately half the staff of a
  medicine in patient unit
• Contamination likely introduced by
• Fecally-contaminated well water, OR
• Hands of one of the people who prepared the
  punch OR
• (possibly) residual environmental contamination
  in household B.

106
Step 9 Implement control and prevention measures

• The well was taken out of service.

107
Step 10 Communicate findings

• Written outbreak report distributed with
  laboratory results approximately one month later
  to
• Public Health
• Infection Control Medical Director
• Occupational Health
• Hospital Nursing VP and Chief of Staff
• Public Relations

108
Conclusions

• NLV outbreaks are
• Good practice
• Important to investigate because of the total
  burden of disease
• Highly disruptive to hospitals and LTCFs
• Cause of significant disability and death,
  especially in vulnerable populations
• Challenging to investigate because laboratory
  diagnosis is not readily available

109
Conclusions

• 10 steps of outbreak investigation
• Conceptual
• Provide a logical progression for the
  investigation
• Can should be taken out of order (with caution)

110
Would Next Years Christmas Punch Benefit from A
Pharmacological Concentration of Ethanol???
111
The Next Few Teleclasses
November 29 Effective Infection Prevention in 3-5
Steps with Allen Soden, Deb Ltd. December
6 Infection Control in the Living and the Dead -
The Angola Marburg Outbreak with Prof.
Adriano Duse, U. of Witwatersrand, South
Africa December 13 Water Quality Issues
Pertaining to Medical Device Reprocessing
with Dr. Michelle Alfa, St. Boniface Hospital,
Winnipeg
For the full teleclass schedule
www.webbertraining.com For registration
information www.webbertraining.com/howtoc8.php