HLTH 207 Week 2 Asynchronous Work (ONE)

Descriptive epidemiology

studies distribution of health-related outcomes

3 descriptive epidemiology questions

1. Who is affected (demographic characteristics)?

2. Where is the health outcome occurring (geography)?

3. When is the health outcome occurring (changes over time)?

Analytic epidemiology

studies determinants of health-related outcomes

Analytic epidemiology question

Why is the health outcome occurring?

Difference in populations for descriptive vs. analytic studies

Descriptive studies provide information about the experience of a single group, while analytic studies compare the experience of two (or more) groups

Can descriptive studies imply association? Why or why not?

Descriptive studies cannot imply associations because there is not a comparison group

Case definition

set of clinical and/or laboratory criteria used to identify individual how have the outcome of interest

How should case definitions be used?

These should be used consistently to ensure all individuals with the outcome are identified

International Classification of Diseases (ICD)

commonly used system which allows for standardized classifying and reporting of many health-related outcomes

How can the case definition reflect uncertainty in the classification of a health outcome?

by including categories with varying degrees of certainty (suspect, probable, or confirmed)

Symptoms typically fall into the ____________ certainty category of a case definition, while diagnostic tests typically fall into the _______________ category

suspect, confirmed

3 pitfalls of case definitions

1. Insensitivity

2. Specificity

3. Poorly defined

What occurs if a case definition is insensitive?

it may miss some of the individuals who have the outcome

What occurs if a case definition is not specific?

it may include individual who do not have the outcome along with those who do have the outcome

What occurs if a case definition is poorly defined?

individuals with the outcome may be erroneously excluded (false negatives) and/or individuals who do not have the outcome may be erroneously classified as cases (false positives)

False positive

individuals who do not have the outcome who are classified as having the outcome

False negative

individuals with the outcome who are classified as not having the outcome

5 considerations of a case definition

1. How sensitive and specific is the case definition?

2. Can the case definition be applied rigorously and consistently in your study?

3. Does a national or international standard case definition exist?

4. Is there an existing case definition that is often used in other studies?

5. Has the case definition changed over time?

Case report

detailed description of a single person with the outcome of interest

What guidelines should a case report follow?

Should follow CAse REport (CARE) guidelines to ensure consistency and allow for comparison across case reports

Case series

collection of case reports for multiple individuals with the same outcome

Case series function

Case series help paint a better picture of the signs and symptoms of a new outcome

Neither case reports nor case series provide information about what 2 factors?

1. The population from which the cases came

2. The frequency of the outcome

Public health surveillance

ongoing, systematic collection of health-related data used to describe and understand the burden of disease, health-related outcomes, and/or exposures

What should be done with public health surveillance data after it is collected?

These data should be disseminated to relevant audiences in a timely manner to allow for public health action

What can be done with public health surveillance data when the size of the population is known?

When the size of the population is known, data from public health surveillance can be used to estimate outcome frequency

3 types of public health surveillance

1. Passive surveillance

2. Active surveillance

3. Sentinel surveillance

Passive surveillance

reporting sites submit data to health agencies or select health-related outcomes

What kind of health outcomes are required to be reported to relevant local and international health agencies?

Notable conditions like anthrax, cholera, hepatitis, etc.

How does the CDC track notable conditions like anthrax, cholera, hepatitis, etc.?

From the National Notifiable Diseases Surveillance System (NNDSS)

Passive surveillance use

Used for routine, long-term activities

Active surveillance

 health agencies engage in proactive data collection efforts, where they maintain regular contact with healthcare providers to identify cases of a particular health outcome

Active surveillance use

Short term, useful for outbreaks and seasonal diseases

2 benefits of active surveillance over passive surveillance

1. Yields better data completeness and quality

2. Is more expensive and resource intensive

Sentinel surveillance

Is a form of active surveillance

Where and when does sentinel surveillance occur?

a limited number of reporting sites are selected for proactive contact by health agencies

Sentinel surveillance functions

• useful for short term, high frequency outcomes

• could also be used to identify hard to reach populations

Measures of frequency

quantify the occurrence of a health outcome in a defined population and provide context about the population using counts and relevant denominators

3 key measures of frequency

1. Risk/cumulative incidence

2. Rate/incidence density

3. Prevalence

What do risk and rate measure?

measures of the incidence of health outcomes and provide information about the frequency of new cases

What does prevalence measure?

measure of prevalent health outcomes and provides information about the occurrence of existing cases

Risk calculation

number of newly occurring outcomes (I) during a follow up period/number of at-risk people observed (N) during a follow up period = I/M

Risk

Proportion of at-risk individuals who develop a newly occurring outcome during a given follow-up period

At-risk meaning in the context of risk calculation

those who are included in the calculation are eligible to develop the outcome

How is risk presented?

Present as a decimal or percentage and include the length of the follow-up period

When can risk not be calculated?

Risk cannot be calculated when the outcome status for everyone at the end of the follow-up period is not known

What is a common reason for not knowing the outcome status for every participant at the end of a follow-up period?

loss to follow-up, which occurs when individuals cannot be contacted or no longer want to be in the study

Rate calculation

number of newly occurring outcomes (I) during a follow-up period/total person-time contributed by at-risk individuals = I/PT

Rate

A measures of how quickly new outcomes occur in at-risk pearson during a follow-up period

How is rate presented?

Cases per person-time

Rate vs. risk calculation denominator

• Risk sums the total number of at-risk people during the same follow-up period

• Rate sums the time that at-risk people contribute, can differ between people and allows for inclusion of those who are lost to follow-up

Prevalence calculation

number of existing outcomes (P) at a single point in time/number of people in the population (N) at a single point in time = P/N

Prevalence

Proportion of individuals in a specified population with an existing outcome at a single point in time

How is prevalence presented?

Present as a decimal or or percentage

3 factors that increase prevalence

1. More people get sick

2. Duration of illness increases

3. Immigration of those with the illness

3 factors that decrease prevalence

1. Mortality

2. Recovery happens more quickly

3. Emigration of those with the illness

Person characteristics

Demographic factors including age, biological sex, gender, race, ethnicity, SES, health-related behaviors, genetic factors, etc.

Measures of frequency implications in person characteristics

Measures of frequency can be used to describe the distribution of outcomes in different populations

Description of the distributions of outcomes by person can be used to do what 4 things?

1. Understand how outcomes vary across groups

2. Consider different treatment or prevention needs for different groups

3. Inform public health policy

4. Generate hypotheses about potential causes of the outcomes

Place characteristics

geographic factors including country, state, city, and neighborhood, can be visualized using tables or maps

Measures of frequency implications in place characteristics

Measures of frequency can be used to describe the distribution of outcomes in different places

Descriptions of the distributions of outcome by place can be used to do what 3 things?

1. Understand how outcomes vary by geography

2. Inform the public health policy

3. Generate hypotheses about potential causes of the outcomes

Time characteristics

trends in health outcomes change over time due to factors such as new treatments, screening tools, and emerging pathogens

Measures of frequency implications in time characteristics

Measures of frequency can be used to describe the distribution of outcomes across time

Descriptions of the distributions of outcomes by time can be used to do what 3 things?

1. Evaluate success of treatment or prevention programs

2. Generate hypotheses about potential causes of the outcomes

3. Predict future health outcomes

Epidemic curves

 plots of the number of new health outcomes during a specified time period, allowing for review of time trends

Epidemic curves function

Rather than just plotting counts, plotting measures of frequency over time can provide additional context about the occurrence of health outcomes

Time scale requirement for epidemic curve

Time scale should be appropriate for the health outcome

Secular trends

changes in the frequency of an outcome over a longer period of time

2 ways that investigators create epidemiological data

1. May use existing data sources

2. Design a new data collection process

What must investigators consider when sourcing epidemiological data?

• Must consider strengths and limitations for each sources, especially in terms of data quality

• Pay attention to the population for a given measure of frequency

8 common sources of epidemiologic data

1. Surveillance systems

2. National death indices (deaths and causes of death)

3. Birth certificates

4. Reportable disease databases

5. Medical record abstraction

6. Surveys/Questionnaires (self-reported data)

7. Medical exams

8. Laboratory and diagnostic tests

What measures do epidemiological studies often use when factors are hard to measure?

Proxy measures that act as stand-ins for those factors

2 examples of proxy measures

1. Race as a proxy for the effect of racism on health

2. Infant deaths as a proxy for maternal and child health