Epi Bio exam 2

Epidemiology

study of something that afflicts (affects) a population

Stages of disease

Pre-disease stage: before pathologic process begins.

Latent Stage: disease process has started but still asymptomatic.

Symptomatic stage: disease manifestation evident

Primary Prevention

preventing the disease process from starting (Pre-disease stage)

Secondary Prevention

screening and appropriate treatment may prevent progression to symptomatic disease (latent stage)

Tertiary prevention

intervention may be slow, arrest, or reverse progression of disease. (symptomatic stage)

Triad of Factors that cause disease

Host, Agent, Environment
(Vector 4th factor)

**BEINGS MODEL**

major categories of risk factors

B - Biological, Behavioral
E - Environmental
I - Immunological
N - Nutritional
G - Genetic
S - Services, Social, Spiritual

Biological, behavioural (examples)

unprotected Sex -> HIV/AIDS
Excessive alcohol intake-> pancreatitis, cirrhosis
Drug abuse -> overdose

Enivronmental examples

infectious agent in air-conditioning system (Legionella pneumophilia)

Immunologic Factors examples

immunodeficiency

ex given Smallpox eradicated from globe due to vaccines

Nutritional Factors examples

Diet effects disease prevalence
lack of fiber intake (US) constipation

Genetic Factors examples

BrCA1 & BrCA2 genes for breast cancer risk.

Services, Social and Spiritual Factors examples

LDS members (Mormons) lower risk from respiratory disorders due to not tobacco use.

degree of immunity necessary to eliminate a virus from a population depends on:

Type of virus
Time of year
density & social patterns

Why do some vaccines need boosters?

When diseases were more prevalent those immunized would be "re-infected" by disease building a natural booster effect. With diseases less common, a 2nd dose (booster) now required to reach same level of immunity.

Most important factor in reducing infant mortality rate

Sanitary revolution

Iceberg phenomenon

Asymptomatic infections may be uncovered by finding elevated antibody titers, in otherwise clinically well people. (or culture organism from them)

Frequency (Epidemiologic)

frequency of a disease, injury, or death can be measured in different ways.
Can be related to different denominators.

Incidence

The Frequency of NEW occurrences of a disease, injury, death during the time of study.

Prevalence

population who have a specified disease or condition at a single point in time.
Incidence x (average) duration
AKA point prevalence

Period of Prevalence

# of cases during specified time

Cumulative incidence

total number of cases of an epidemic disease reported over time

What can cause an increase in prevalence of a disease?

increased length of survival
medical advances

Risk

proportion of persons who are unaffected at the beginning of a study period, but undergo risk event during study period

Cohort**

persons at risk for developing event

Case fatality Ratio**

proportion of clinically ill persons who DIE from condition
Higher the fatality ratio, more virulent the infection

Pathogenicity of an organism

Proportion of INFECTED persons critically ill

Infectiousness of an organism

Proportion of EXPOSED person who become infected

Which measure is more precises? Rate or Risk?

Rate

*RATE*

frequency of events that occur in a defined time period, divided by the average number people at risk during the study period

Mid period population

Good estimate of the average number of people at risk for outcome during a specific time period
often used as the denominator in rate.

why is a constant multiplier used in Rate ratios?

To make it easier to put the data into perspective.
ex. death rate of .0086 per year ->
becomes 8.6 deaths out of 1000 (easier to understand)

Rate is a good estimate of risk only when:

event in numerator occurs only once (example death)
proportion of population affected by event is small
Time interval is relatively short.

if time interval long, or % people die is large, rate will be larger than risk.

validity of rates

all events in numerator must occur in persons included in denominator.
(death of person in US population. Death must be a US person)

All persons counted for in Denominator must be at risk for events in numerator.
(cervical cancer/ female population ) (men can't develop cervical cancer)

Incidence Rate**

number of incident cases over a defined study period, divided by the MIDPOINT population of the study period.

Incidence Density**

frequency (density) of new event per person-time (person/month) (person/year)

useful when event can occur more than once in same person. (colds, otitis media, etc)

Types of Comparisions for Rates & RIsks

Comparison of observed rate w/ Target rate.

Comparison of Two different populations @ same time- (MOST COMMON) ex. death rates of 2 different countries

Comparison of same population at different time (must account for trends of change of a population)

3 categories of rate

Crude: entire population

Specific: specific groups in a population

Standardized: rates adjusted for characterization to compare

crude rates

rates that apply to entire population, w/o reference to any characteristics of individuals in it.

specific rates

Population divided into subgroups based on a particular characteristic of interest.
(age, sex, race, risk factors, comorbidity)

What is the simplest way to control for age bias in crude rates?

Age Specific Death Rate (ASDR)

Age Specific Death Rate***

No of deaths in particular age group (defined place & time) /mid period population. (same age group & Time) x1000

Standardized Rates

AKA adjusted rates.

Crude rates that have been adjusted to control for effects of age or other characteristics to allow for valid comparisons of rates.

Crude Death Rate**

sum of the ASDR in each age group weighted by relative size of each age group.

Direct standarization

Most used method to remove biasing effect of differing age structures.
ASDR of 2 populations compared are applied to a SINGLE standard population.

How to do direct standardization of ASDR

multiply each ASDR population by the total combined population of each ASDR.
ex. Pop A: 1000 people (ages 30-50)
Pop B: 4000 people (ages 30-50)

ASDR for each pop the multiplied by 5000 to get standardized rate.

early fetal death

aka miscarriage. dead fetus delivered w/in 1st 20 weeks gestation

intermediate fetal death

Dead fetus delivered between 20-28 weeks of gestation

infant death

death of live born infant before Infants 1st birthday

Neonatal death

death of live-born infant before completing of 28th day of life (less than 1 month old)

Post neonatal Death

death of infant after 28th day of life, but before 1st birthday

Crude BIRTH rate**

# of live births / midpoint population

Infant Mortality Rate (IMR)

overall index of health status of a nation.
# of infant deaths (<1yr old) / # of live births.

Maternal Mortality Rate (MMR)

useful measure of progress of a nation in providing adequate nutrition & medical care for pregnant women
# pregnancy-related deaths/ # of live births.

First Responder to disease outbreak

CDC

Active Surveillance*

requires periodic telephone calls/personal visits to reporting individuals & institutions to obtain required data.

more labor intensive & costly

passive surveillance*

most of the surveillance done on a routine basis.

physicians, clinics, laboratories, & hospitals required to report all cases of reportable diseases that come to their attention

Patterns of diseases studied by:

Time & geographic locations
Characteristics of persons involved

continued surveillance

allows epidemiologists to detect deviations from usual pattern of data

Time trends

implications of long-term trends of disease usually different from those of outbreaks/epidemics.

seasonal variation examples

incidence of Flu increases during the winter.
Drowning Rates go up in the summer.
Diphtheria rates rise early autumn.

Epidemic**

aka disease outbreak, occurrence of a disease at an unusual (unexpected) frequency

How are flu epidemics determined

Compare current flu rates to flu rates from previous years to compare reported percentages.

Epidemic Threshold

epidemic examples

single case of smallpox
paralytic poliomyelitis western hemisphere
Food poisoning

Attack Rate**

number of new cases/number of persons exposed x100

Endemic

disease in a population occurs regularly & at a constant level

Investigating Epidemics steps**

1. establish diagnosis
2. establish epidemiologic case definition
3.characterize: time, place, person

Epidemiologic Case definition**

list of specific criteria used to decide whether or not a person has the disease of concern. (not the same as clinical diagnosis)

Epidemic by time, place, person

cant start data collection before case definition established.
cant count people w/ food poisoning from a restaurant if they didn't eat there.

geographic clustering

mapping geographic locations of cases. Spots on a map will show where affected individuals, live, work, or go to school.
example. Cholera mapping deaths around water pump. (1855)

propagated pattern

Occurs when the infection "propagates itself" by spreading directly from person to person over an extended period

mixed pattern

Persons acquire a disease through a common source and spread it to family members or others (secondary cases) by personal contact

4 common types intervention to control outbreak

Sanitation: modification of environment.
Prophylaxis: barrier to the infection w/in susceptible host
Diagnosis & Tx: prevent spread
Control Vectors: (ex condom use)

Sanitation Control measures

modification of environment

Removing pathogenic agent from sources of infection.

(water/food)

Removing human source of infection

(quarantine)

Preventing contact w/ source.

(removing susceptible people from environment)

Written & Oral Communication:

enable other agencies to assist after outbreak
adds available information regarding prevention

Follow-Up Surveillance:

active surveillance needed after outbreak
Sound surveillance will detect subsequent outbreaks & evaluate effective control measures.

**Standardized Mortality Ratio**

observed deaths in a population/expected deaths in a population x100

Sufficient Cause*

if factor (cause) is present, the effect (disease) will ALWAYS occur.
example: Genetic abnormality causing Tay-Sachs

Necessary cause*

the cause must be present for disease to occur. However, cause can be present without the disease occurring.
example: M. tuberculosis needed to cause TB but not every person develops TB

Risk Factor*

if factor is present, the PROBABILITY of the effect will occur increases.
example: Smoking increases risk of developing lung cancer

Directly causal association*

The factor exerts its effect in the absence of intermediary factors. (intervening variables)
"severe blow to the head will causes brain damage"

Indirectly causal association*

The factors exert its effect via intermediary factors

"Poverty doesn't cause disease/death but prevents adequate nutrition/med. care/housing can lead to ill health."

Noncausal association*

If a statistically significant association is found b/w two variables, BUT some other factor causes the presumed cause & effect.

Non-causal association example

Baldness and Risk of coronary artery disease.
Both are functions of other factors; (Age, gender, levels dihydrotestosterone) (confounders)

For causation to be identified:

Presumed risk factor MUST be present significantly more in persons w/ particular disease of interest than those w/o disease

Mill's Canons*

Criteria that increase the probability that a statistical association is causal.

Mill's canons criteria**

Strength of the association

The consistency of the association

The specificity of the association

The Biological Plausibility of the association

The presence of a dose-response relationship

Mills: Strength of association

the difference is large (stronger association stronger cause)

Mills: The consistency of the association

Is it always found

Mills: The specificity of the association

no effect if cause not present

Mills: The biological Plausibility of the association

It makes sense based on current knowledge

Mills: The presence of a dose-response relationship

The greater the causal factor, the greater the effect

Assembly Bias*

may take the form of selection bias or allocation bias.
Characteristics of the intervention group & those of the control group are not comparable at the start.

Selection Bias*

subjects are allowed to select the study group they want to be in.
Almost any nonrandom method of allocation of subjects to study group may produce bias.

Allocation Bias*

May occur if investigators choose a nonrandom method of assigning subjects to study group

Detection Bias*

may be the result of a failure to detect:
A case of disease, A possible casual factor, An outcome of interest

Measurement Bias*

may occur while collecting baseline data or follow up data
(measure of height of patient w/ shoes on)

Recall Bias*

may occur when subject has experienced adverse event more likely to recall previous risk factors than subjects who never experienced the event.

Late Look Bias (Neyman)**

severe/rapidly fatal diseases are less likely to be found by a survey.

tend to find less aggressive cases b/c they live longer and more likely available to be found at screening.

Random Error

produce findings too high or too low in approximately equal amounts.
ordinarily less serious than bias, less likely to distort data.