VOLUME 3, ISSUE 2

Case-Control Studies for Outbreak Investigations

CONTRIBUTORS
Authors:
Amy Nelson, PhD, MPH
Kim Brunette, MPH
FOCUS Workgroup*
Reviewers:
FOCUS Workgroup*
Production Editors:
Tara P. Rybka, MPH
Lorraine Alexander, DrPH
Rachel A. Wilfert, MD, MPH
Editor in chief:
Pia D.M. MacDonald, PhD, MPH

* All members of the FOCUS Workgroup are named on the last page of
this issue.

The North Carolina Center for Public Health Preparedness is funded by Grant/Cooperative Agreement Number U90/CCU424255 from the Centers for Disease
Control and Prevention. The contents of this publication
are solely the responsibility of the authors and do not
necessarily represent the views of the CDC.

If you have never been to Western

Michigan in the winter, the odds are
you have never seen how lake-effect
snow can keep falling and falling.
And if you have never been rollerblading, the odds are you do not
know the joy of rolling freely down a
path or the terror of not being able to
stop at an intersection.
People make statements like these
all the time. They use odds as a way
of describing the likelihood of a particular event in relation to some
more general experience. Maybe you
do rollerblade; then you probably
have had the experience of yelling
How do I STOP?! But if you have
never been rollerblading, you probably have not experienced the terror
of having wheels on your feet and
not being able to stop.
Lets think about this example in
terms of exposure and outcome. The
exposure is equivalent to the experience of rollerblading, while the outcome is the terror of not being able
to stop. And the odds reflect the likelihood of the exposure-outcome relationship. In this issue of FOCUS, we
discuss the case-control study design, which uses the idea of odds to
describe exposure-disease relationships.
As noted in earlier FOCUS issues on
study design, analytic studies are
conducted to answer the question,
What is the relationship between
exposure and disease? The casecontrol design is an efficient method
of exploring this relationship. It is

often conducted with relatively few

diseased individuals (so it is efficient),
and it is particularly useful for studying a rare disease or investigating an
outbreak.
Case Selection
Determining who will be a case in the
study depends on how a case is defined. A case definition is a set of
standard criteria for deciding whether
an individual should be classified as
having the health condition of interest. (1) The definition usually includes clinical criteria and is restricted
to a certain time and place, and certain personal characteristics. The case
definition must be clear, objective,
and consistently applied.
After the case definition is decided on,
case-patients need to be recruited
into the study. There are many ways to
identify case-patients, including medical records, laboratory results, surveillance systems, registries, and mass
screening programs. Case-patients
can also be asked to identify other
persons they know who have a similar
illness.

For example, in August 2001, the

Illinois Department of Health was
notified of a cluster of cases of
diarrheal illness associated with
exposure to a recreational waterpark in central Illinois. To identify
additional case-patients, local
media and community networks
were asked to encourage ill persons to contact the local health
department. In addition, casepatients were asked if there were

North Carolina Center for Public Health PreparednessThe North Carolina Institute for Public Health

FOCUS ON FIELD EPIDEMIOLOGY

Page 2

lected students in grades K2 of the primary school

with no history of current or previous varicella. One
control was recruited for each case-patient. (5)

any other ill persons in their household or if anyone

attending the waterpark with them was ill. (2)
Control Selection

The most difficult part of a case-control study is choosing

the control group. It has even been said that, This is one
of the most difficult problems in epidemiology. The challenge is this: If we conduct a case-control study and find
more exposure in the cases than in the controls, we would
like to be able to conclude that there is an association
between the exposure and the disease in question. The
way the controls are selected is a major determinant of
whether such a conclusion is valid. (3)
Controls are persons who do not have the disease in question. Controls should be representative of the population
from which the cases arose [known as the source population], so that if a control had developed the disease, he or
she would have been included as a case in the study.
Controls should also provide a good estimate of the level
of exposure one would expect in that population. (1)
There are several sources of controls for case-control studies. They can be selected from the same health-care institutions or providers as the cases, the same institution or
organization as the cases (e.g., schools, workplaces), from
relatives, friends, or neighbors of the cases, or randomly
from the source population. (1) Investigators may even
use multiple methods of control selection. Additionally,
investigators may choose to select multiple controls per
case to increase the likelihood of identifying significant
associations (usually no more than 3 controls per case).
Keep in mind, however, that the sources for your controls
will depend on the scope of the outbreak.

In August 2000, an increase was noted in Salmonella

serotype Thompson isolates from Southern California
patients with onset of illness in July. The preliminary
interviews confirmed that many of the patients with
S. Thompson infection had eaten at a Chain A restaurant in the 5 days before illness onset. Therefore a
case-control study was conducted to evaluate specific
food and drink exposures at Chain A restaurants.Controls were well friends or family members
who shared meals with cases at Chain A during the
exposure period. (6)

Members of the same institution or organization

In a 2004 outbreak of varicella in a primary school in a
suburb of Beijing, China, a case-control study was conducted to identify factors contributing to the high rate
of transmission and to assess the effectiveness of
control measures. Controls included randomly se-

Random sample of the source population

During January - June 2004, an aflatoxicosis outbreak
in eastern Kenya resulted in 317 cases and 125
deaths. A case-control study was conducted to
identify risk factors for contamination of implicated
maize.[Investigators] randomly selected 2 controls
from each case-patients village.To choose each control, [investigators] spun a bottle in front of the village
elders home and walked to the fifth house in the direction indicated by the bottle (or to the third house in
sparsely populated areas). At the selected household, a random number list was used to select one
household member. (7)

Multiple methods of control selection

In the waterpark outbreak of diarrheal illness in Illinois
mentioned above, investigators recruited 1 control per
case using 3 methods. First, case-patients were asked
to identify a healthy person. Second, investigators
used the local reverse-telephone directory based on
residential address of case-patients. Lastly, investigators canvassed local schools and community groups
for controls. (2)

Persons served by the same health-care institution or

providers as the cases
In August 2001, a cluster of Ralstonia pickettii bacteremia occurred among neonatal intensive care unit
(NICU) infants at a California hospital.A case-control
study was conducted to identify risk factors for infection.Controls were NICU infants who: (1) had blood
cultures taken during either cluster period (July 30August 3 and August 19-30); (2) had blood cultures
that did not yield R. pickettii; and (3) had been in the
hospital for at least 72 hours. Investigators attempted to recruit 2 controls per case-patient. (4)

Relatives, friends, or neighbors

Selection Bias
Bias is a distortion of the relationship between exposure
and disease. If there are systematic differences in the way
you select your controls and the way you select your cases,
you could introduce bias. In epidemiology, we refer to bias
related to the way cases or controls are chosen as
selection bias.

For example, suppose that most of your case-patients

work on lower floors of an office building and employees on the lower floors are more likely to leave the
building to go out for lunch. If your control population

North Carolina Center for Public Health PreparednessThe North Carolina Institute for Public Health

Page 3

VOLUME 3, ISSUE 2

ness. Investigators selected 2 controls for each case,

matched for age, sex, and place of residence, and
identified through the computerized Swedish National
Population Register, which stores the name, date of
birth, personal identifying number, and address of all
citizens and residents. (8)

is mostly employees from upper floors, the comparison may lead you to conclude that there is a real difference between cases and controls associated with
eating at a local deli. But in fact the difference is due
to where they worked in the building, which resulted in
how often they ate in restaurants.

Lets say that you are investigating an outbreak at a

gym and a majority of the case-patients are females. A
majority of the controls you recruit are male. You conclude that there is an association between the illness
and participation in an aerobics class. In reality, however, the outbreak was caused by the steam in the
sauna in the womens locker room. The appearance of
a relationship between the illness and the aerobics
class was simply due to the fact that women are more
likely to take an aerobics class than men.

Matching
Since the validity of case-control studies is dependent on
the similarity of cases and controls in all respects except
exposure, investigators frequently match cases and controls on characteristics like age and gender. Matching factors should be important in the development of the disease, but not in the exposure under investigation. Since
the matching variable will not be associated with either
case or control status, it cannot confound, or distort the
exposure-disease association.
There are two ways to match cases and controls: individual matching and group matching. In individual matching,
or the use of matched pairs, each case is matched with a
control that has specific characteristics in common with
the case. In group matching, also known as frequency
matching and category matching, the proportion of controls with certain characteristics must be identical to the
proportion of cases with these characteristics. This
method requires that all cases be selected first so that the
investigator knows the proportions to which the controls
should be matched. For example, if investigators wanted
to match on gender and 30% of the cases were male, then
investigators would select controls so that 30% of controls
would be male.

To illustrate individual matching, in an outbreak of

tularemia in Sweden in 2000, investigators conducted
a case-control study to identify risk factors for the ill-

To illustrate group matching, in an outbreak of Escherichia coli associated with a petting zoo at the
2004 North Carolina State Fair, investigators recruited
3 controls for each case. Controls were groupmatched by age groups (1-5 years, 6-17 years, and 18
years and older), meaning that the proportion of controls in each age group was identical to the proportion
of cases in each age group. The controls were identified from a list provided by fair officials of 23,972
persons who purchased tickets to the fair online, at
kiosks, or in malls. (9)

Matching can be time efficient and cost effective, and improve statistical power. However, the more variables chosen as matching characteristics, the more difficult it is to
find a suitable control to match to the case. It is important
to remember that once a variable is used for matching,
there can be no discernible relationship between this variable and the disease. So do not match on anything you
think might be a risk factor for disease. Remember, when
a matched study is carried out, data must be analyzed
using methods consistent with matched data.
Conducting the Investigation
The next step is to gather demographic information and
exposure histories from cases and controls. (Information
on questionnaire development and interviewing techniques is available in past issues of FOCUS.) After you
have collected the data you need, you can begin the analysis and calculate measures of association.
Analyzing the Data
In a case-control study, an odds ratio is calculated to
measure the association between an exposure and the
occurrence of a disease.
Calculating Odds and Odds Ratios
What are odds? How are odds different from probability or

Useful resources for case-control studies:

Case-Control Studies. ERIC Notebook. September 1999, Issue 5.

Rothman KJ. Epidemiology: An Introduction. New York, Oxford University Press; 2002.

North Carolina Center for Public Health PreparednessThe North Carolina Institute for Public Health

FOCUS ON FIELD EPIDEMIOLOGY

risk? Lets say there is a bag containing 20 poker chips: 4

red chips and 16 blue chips.
Probability is the number of times something occurs divided by the total number of possible occurrences. The
probability of getting a red chip is 4/20 (or 1/5 or 20%).
The probability of getting a blue chip is 16/20 (or 4/5 or
80%).
Odds are the number of times something occurs divided by
the number of times something does not occur. The odds
of getting red are 4/16 (or 1/4) and the odds of picking a
blue chip are 16/4 (or 4/1). People may refer to the odds
of getting a blue chip as 4 to 1 against getting a red chip.
Odds = probability/(1-probability)
If the probability of picking a red chip is 20%, then the
odds are 0.20/(1-0.20) or 1/4.
Probability = odds/(1+odds)
If the odds of picking a red chip are 1/4, then the probability of getting a red chip is 0.25/(1+0.25)=0.20.
Thus, while odds are a measure related to probability, they
measure the occurrence of an event as compared to the
non-occurrence of the same event. Variables with two levels, called binary variables, are used to calculate odds.
Binary variables are those with yes/no responses, such as
disease/no disease, or exposed/not exposed.

Page 4

An odds ratio (OR) is the odds of exposure among cases

(a/c) divided by the odds of exposure among controls
(b/d). Exposure among cases is compared to exposure
among controls to assess whether and how exposure levels differ between cases and controls. This OR is numerically the same as that obtained by multiplying diagonally
across the 2x2 table and dividing the products (ad/bc);
hence the name
Odds Ratio = (a/c) (b/d) = ad/bc
cross-products
ratio.
To interpret the odds ratio, we compare the value of the
OR to 1:
If the odds ratio = 1; the odds of exposure are the same
for cases and controls (no association between disease
and exposure).
If the odds ratio > 1; the odds of exposure are greater
among cases than among controls (a positive association
between disease and exposure).
If the odds ratio < 1; the odds of exposure are less among
cases than among controls (a negative, or protective, association between disease and exposure).
Note: Accurate interpretation of the odds ratio requires the
use of confidence intervals and other statistical methods
that will be discussed in an upcoming issue of FOCUS.

The odds of exposure among cases are calculated by dividing the number of exposed cases by the number of unexposed cases. Similarly, the odds of exposure among controls are calculated by dividing the number of exposed controls by the number of unexposed controls.
Odds for cases =

Exposed cases
Unexposed cases

Odds for controls = Exposed controls

Unexposed controls

For example, in an investigation of an outbreak of

Hepatitis A among patrons of a Pennsylvania restaurant, a case-control study was conducted to identify
food associated with the illness. A total of 240 casepatients and 134 controls were identified. Data obtained from the case-patients and controls found that
218 case-patients and 45 controls had consumed
mild salsa, as shown in the table below. (10)

2x2 table showing exposure to mild salsa among casepatients and controls

Ate mild salsa (exposed)

A 2x2 table shows the distribution of cases and controls:

Did not eat mild salsa (not exposed)

Odds: 2x2 table

Case

Control

Exposed

Not exposed

Odds of exposure

a/c

b/d

CASE

CONTROL

218

45

22

89

OR = (218/22) = (218x89) = 19.6

(45/89)
(45x22)

In this example, the odds ratio of 19.6 means that

among patrons who became ill the odds of having
eaten the mild salsa were 19.6 times the odds of patrons not who did not get ill. To put it simply, cases
were 19.6 times more likely than controls to have
eaten mild salsa. (10)

North Carolina Center for Public Health PreparednessThe North Carolina Institute for Public Health

Page 5

VOLUME 3, ISSUE 2

Odds vs. Risk

On occasion, you may have an exposure variable that has
more than two levels (e.g., age group, race, or serving size).
You can calculate an odds ratio for each level relative to a
reference group. For example, suppose you have several
serving sizes: 0 servings, <1, 1, and 2 or more servings. The
reference group is those who had 0 servings. You calculate
the odds ratios as follows:
Case

Control

Odds Ratio

2 servings

a3

b3

a3d/b3c

1 serving

a2

b2

a2d/b2c

<1 serving

a1

b1

a1d/b1c

0 servings

reference

Matched Analysis
If you decide to match controls to cases using individual
matching (rather than group matching), the 2x2 table needs
to be set up differently: you examine pairs in the table, so
you have cases along one side and controls along the other,
and each cell in the table contains pairs. The generic set-up
of the table is:
Controls

Cases

Exposed

Not Exposed

Total

Exposed

e+f

Not Exposed

g+h

Total

e+g

f+h

Cell e contains the number of matched case-control pairs in

which both the case and the control were exposed. This is a
concordant cell because the case and the control have the
same exposure status. Cell h is also a concordant cell.
Cell f contains the number of matched case-control pairs in
which the cases were exposed but the controls were not exposed. This is a discordant cell (as is cell g) because the
case and the control have a different exposure status.
Because you want to contrast the exposure between cases
and controls, only the discordant cells (f and g) provide useful data. In individually matched analysis, the matched odds
ratio is calculated as cell f
Matched Odds Ratio = f/g
divided by cell g.
With group matching, stratified analysis should be used. This
type of analysis will be discussed in an upcoming issue of
FOCUS.

Odds differ qualitatively from the risk calculated in a cohort study, and it is important to understand this distinction. Case-control studies select participants based on
disease status and then measure exposure among the
participants; therefore they can only approximate the risk
of disease given exposure. The values needed to calculate
risk are not available from a case-control study because
the study does not include the entire population at risk.
Although you might be able to include all or most of the
cases in the study, finding all of those who did not get sick
would be difficult or impossible. A case-control study thus
uses only a subset of many potential controls, and you
calculate the odds ratio as an estimate of the risk.
Examples of Case-Control Studies
E. coli associated with a fast-food restaurant chain
In November 1999, a childrens hospital notified the
Fresno County (California) Health Department of 5 cases
of E. coli O157 infections during a 2-week period (6). Initial
interviews revealed that all case-patients had eaten at
popular fast-food restaurant chain (chain A) in the 7-day
period before the onset of illness. Local health officials
and clinicians throughout California were asked to enhance surveillance for E. coli O157 infections. Additionally,
states bordering California were asked to review medical
histories of persons with recent E. coli O157 infections
and arrange for PFGE [pulse-field gel electrophoresis] subtyping of recent E. coli O157 isolates. To identify risk factors for infection, 2 sequential case-control studies were
conducted in early December 1999. (11)
The first case-control study was conducted to determine
the restaurant associated with the outbreak. For this
study, a case was defined as a patient with (1) an infection with the PFGE-defined outbreak strain of E. coli
O157:H7..., a diarrheal illness with 3 loose stools during
a 24-hour period, and/or an HUS [hemolytic uremic syndrome] during the first 2 weeks of November 1999; or (2)
an illness clinically compatible with E. coli O157:H7 infection, without laboratory confirmation but with epidemiologic connection to the outbreak. A control was defined
as a person without a diarrheal illness or HUS during the
first 2 weeks of November 1999. Controls were agematched and systematically identified using computerassisted telephone interviewing of residents in the same
telephone exchange area as case-patients. [Investigators
attempted] to obtain 2 controls per case. Case-patients
and controls were queried using a standardized questionnaire to determine whether they had eaten at a number of
national fast-food restaurant chains in the week before
illness onset. Investigators enrolled 10 cases and 19
matched controls. Of the 9 restaurants, only chain A

North Carolina Center for Public Health PreparednessThe North Carolina Institute for Public Health

FOCUS ON FIELD EPIDEMIOLOGY

showed a statistically significant association with illness among cases and controls.
(11)
[Based on these results], a second casecontrol study involving patrons of chain A
restaurants was conducted to determine
the specific menu item or ingredient associated with illness. For this study, a case was
defined as above but restricted to those
who had eaten at chain A, and only those
who could be matched with meal companion-controls. Cases and controls were
asked about consumption of specific foods
and beverages that appeared on the chain
A restaurant menu.Eight cases and 16
meal companion-controls were enrolled in
this study. By calculating the matched
odds ratio, consumption of a beef taco was
found to be significantly associated with
illness. A traceback investigation implicated
an upstream supplier of beef, but a farm
investigation was not possible. (11)
Listeriosis associated with deli meat
[In] July and August 2002, there were 22
cases of listeriosis in Pennsylvania, a nearly
3-fold increase over baseline. PFGE subtyping identified a cluster of cases caused
by a single Liseteria monocytogenes strain.
The CDC [Centers for Disease Control and
Prevention] asked health departments in
the northeast United States to conduct active case finding, prompt reporting of listeriosis cases, and retrieval of clinical isolates
for rapid PFGE testing. Investigators
conducted a case-control study to identify
the source of the outbreak. (12)
A case-patient was defined as a person
with culture-confirmed listeriosis between 1
July and 30 November 2002, whose infection was caused by the outbreak strain. A
control-patient was defined as a person
with culture-confirmed listeriosis between 1
July and 30 November 2002, whose infection was caused by any other non-outbreak
strain of L. monocytogenes, and who was
from a state with at least 1 case-patient.
Case-patients and control-patients were
interviewed with a standard questionnaire
[including more than 70 specific food
items]...to gather medical and food histories during the 4 weeks preceding culture
for L. monocytogenes. (12)

[The study obtained data] from 38 casepatients and 53 controls. By calculating

the odds ratio, investigators found that
infection with the outbreak strain was
strongly associated with consumption of
precooked turkey breast products sliced at
the deli counter of groceries and restaurants.No other single food item was significantly associated with outbreak strain
infection, except for lettuce, which was
protective. (12)
Based on the results of a traceback investigation, 4 turkey processing plants were
investigated. The outbreak strain of L.
monocytogenes was found in the environment of plant A and in turkey breast products from plant B. Both plants suspended
production and together recalled more
than 30 million pounds of products, resulting in one of the largest meat recalls in US
history. (12)
Conclusion
In using the case-control study method,
considering underlying characteristics of
the population that gave rise to the cases
will help to select appropriate controls.
Improper selection of controls can introduce bias and result in a spurious association between an exposure and illness. If
controls are representative of the source
population, case-control studies are an
efficient way to conduct an analytic study
to determine the relationship between exposure and disease.

Page 6

Glossary
Matching: The process of
making cases and
controls comparable with
respect to extraneous
factors.
Odds: The ratio of the
probability of occurrence
of an event to that of
nonoccurrence, or the
ratio of the probability
that something is so to
the probability that it is
not so.
Odds Ratio: The ratio of
the odds of exposure
among cases to the odds
of exposure among
controls.

From: Last, JM. A Dictionary of Epidemiology. 4th ed.

New York, NY: Oxford
University Press; 2001.

North Carolina Center for Public Health PreparednessThe North Carolina Institute for Public Health

REFERENCES:
1.

Gregg MB. Field Epidemiology. 2nd ed. New York, NY: Oxford University Press;
2002.

2.

Causer LM, Handzel T, Welch P, et al. An outbreak of Cryptosporidium

hominis infection at an Illinois recreational waterpark. Epidemiol Infect.
2006;134(1):147-156.

3.

Gordis L. Epidemiology. 2nd ed. Philadelphia, PA: WB Saunders Company;

2000.

Chapel Hill, NC 27599-8165

4.

Kimura AC, Calvet H, Higa JI, et al. Outbreak of Ralstonia pickettii bacteremia
in a neonatal intensive care unit. Pediatr Infect Dis J. 2005;24:1099-1103.

Phone: 919-843-5561

5.

Ma H, Fontaine R. Varicella outbreak among primary school students--Beijing,

China, 2004. MMWR Morb Mortal Wkly Rep. 2006;55(suppl):39-43.

6.

Kimura AC, Palumbo MS, Meyers H, Abbott S, Rodriguez R, Werner SB. A

multi-state outbreak of Salmonella serotype Thompson infection from commercially distributed bread contaminated by an ill food handler. Epidemiol
Infect. 2005;133(5):823-828.

7.

Azziz-Baumgartner E, Lindblade K, Gieseker K, et al and the Aflatoxin Investigative Group. Case-control study of an acute aflatoxicosis outbreak, Kenya,
2004. Environ Health Perspect. 2005;113:1779-1783.

8.

Eliasson H, Lindbck J, Nuorti JP, et al. The 2000 tularemia outbreak: a casecontrol study of risk factors in disease-endemic and emergent areas, Sweden. Emerg Infect Dis. 2002;8:956-960.

9.

Goode B, OReilly C. Outbreak of Shiga toxin producing E. coli (STEC) infections associated with a petting zoo at the North Carolina State Fair Raleigh,
North Carolina, November 2004. Raleigh, NC: NC Dept of Health and Human
Services; June 29, 2005. Available at:
www.epi.state.nc.us/epi/gcdc/ecoli/EColiReportFinal062905.pdf. Accessed
September 6, 2006.

CONTACT US:
The North Carolina Center for Public Health
Preparedness
The University of North Carolina at Chapel Hill
Campus Box 8165

Fax: 919-843-5563
Email: [email protected]

FOCUS Workgroup:
Lorraine Alexander, DrPH
Kim Brunette, MPH
Anjum Hajat, MPH
Pia D.M. MacDonald, PhD, MPH
Gloria C. Mejia, DDS, MPH
Sandi McCoy, MPH
Amy Nelson, PhD, MPH
Tara P. Rybka, MPH
Michelle Torok, MPH
Rachel A. Wilfert, MD, MPH

10. Wheeler C, Vogt TM, Armstrong GL, et al. An outbreak of hepatitis A associated with green onions. N Engl J Med. 2005;353:890-897.
11. Jay MT, Garrett V, Mohle-Boetani JC, et al. A multistate outbreak of Escherichia coli O157:H7 infection linked to consumption of beef tacos at a
fast-food restaurant chain. Clin Infect Dis. 2004;39:1-7.

If you would like to receive electronic copies of FOCUS on Field Epidemiology, please
fill out the form below:

12. Gottlieb SL, Newbern EC, Griffin PM, et al and the Listeriosis Working Group.
Multistate outbreak of listeriosis linked to turkey deli meat and subsequent
changes in US regulatory policy. Clin Infect Dis. 2006;42:29-36.

NAME: ___________________________________

DEGREE (S): ______________________________

AFFILIATION: ______________________________

E-MAIL ADDRESS: __________________________

May we e-mail any of your colleagues? If so, please include their

e-mail addresses here:

UPCOMING TOPICS!

Conducting Traceback
Investigations

Conducting Environmental
Health Assessments

_____________________________________________________
_____________________________________________________

Basics of Data Analysis

_____________________________________________________

Advanced Data Analysis

Please fax to: (919) 919-843-5563

or mail to: North Carolina Center for Public Health Preparedness
The University of North Carolina at Chapel Hill
Campus Box 8165
Chapel Hill, NC 27599-8165

We are on the web!

http://www.sph.unc.edu/nccphp

Or go online: http://www.sph.unc.edu/nccphp/focus/

North Carolina Center for Public Health PreparednessThe North Carolina Institute for Public Health