epi 221: midterm exam

significant results

RR?
CI?
p-value?

RR: includes 1

CI: 95%, narrow

p-value: <0.05

environmental epi

effect on human health of physical, biologic, and chemical factors in environment
similar to nutritional epi

ecologic studies

at the population level, large sample size
observational/analytical

measured at group level (all other studies → individual level)
e.g., measuring exposure as % malnourished vs malnourished/not malnourished
e.g., measuring outcome as infant mortality rate vs infant deaths

nutritional epidemiology

relations between diet and health

defined by exposures

life course theory

what?
five principles?
three concepts?
three epi models?

explains health & disease patterns across populations over time
biological, behavioral, social, economic, environmental → disease

1. lifespan development: natural consequences

2. human agency: personal control, behavior, and social contexts (!!)

3. timing: when? how long? in what order? (e.g., critical/sensitive windows)

4. linked lives: social networks

5. historical time & place: period, cohort (e.g., great depression)

1. trajectories: substantial part of lifespan (e.g., alcohol consumption)

2. transitions: brief part of lifespan (e.g., parenting)
   maintains/is within trajectory(s)

3. turning point: redirection in situation/behavior (e.g., parenthood → decreased alcohol)
   change in trajectory

1. latency: biological risk
   exposure that manifests later in life (e.g., fetal exposure → adult disease)

2. cumulative: accumulation risk
   additive exposures that accumulate throughout life → outcome
   sequence does not matter, can be any factor (social, biological, etc.)

3. pathway: social risk
   one exposure can result in multiple/different outcomes (e.g., school prep → poor vs good → dumb vs smart)
   sequence matters

confounding

confounder: third variable, precedes exposure, related to outcome

confounding control:

• design: randomization

• analysis: stratification

infant mortality

death of a live-born child before the age of 1

maternal mortality

death of a woman while pregnant OR up to 6 weeks after birth

3 factors of MCH historical foundation

4 phases in US health policy for MCH

policy development for MCH: Phase 1

Characteristic:

Motivation:

Event:

policy development for MCH: Phase 2

Characteristic:

Motivation:

Event:

policy development for MCH: Phase 3

Characteristic:

Motivation:

Event:

policy development for MCH: Phase 4

Characteristic:

Motivation:

Event:

title V

Main source of MCH support from the government

Social Security Act

When: 1935

What:

• Title IV:

• Title V:

• Title XVIII:

• Title XIX:

OBRA

When: 1989

4 Initiatives

Healthy People 2020

What are the 4 Goals?

When: 2010 (by DHHS)

4 Goals

Healthy People 2030

What are the 5 objectives?

When: 2020

5 Objectives

public health

science of protecting and improving health of communities through health promotion, research, and [prevention, detection, & control of disease]

epidemiology

one of public health’s disciplines

study of distribution of determinants of health in the population & application to control health problems

distribution

what?
hypothesis?
studies?

descriptive epidemiology; describes time/person/place

no hypothesis/not tested

case/cross-sectional studies

determinants

what?
hypothesis?
studies?

analytical epidemiology; determines risk factor → outcome (hypothesis)

hypothesis/tested

observational/experiemntal studies

application

what?
4 intervention targets?

prevention epidemiology

when/where intervention is targeting

i.e., primordial, primary, secondary, tertiary

primordial

application → prevention → intervention

targets risk factor
e.g., exercise to prevent obesity in people who are not obese

primary

application → prevention → intervention

targets early disease
e.g., exercise to prevent early GDM in people who are obsese

secondary

application → prevention → intervention

targets late disease
e.g., screening/glucose monitoring to prevent late GDM in people who have early GDM

tertiary

application → prevention → intervention

targets disability
e.g., to prevent disability/complications of GDM in people who have late GDM

distribution study design

analytical study design

qualitative design

distribution

survey design

distribution

experimental

analytical

observational

analytical

case report

case series

qualitative (distribution)

cross-sectional

survey (distribution)

parallel

cross over

factorial

experimental (analytical)

cohort

case control

cross-sectional

observational (analytical)

incidence proportion (attach rate/risk)

# of new cases of disease (specified time)

/

population (at risk) at start of specified time

secondary attack rate

# of new cases among contacts

/

total # of contacts

incidence (person-time) rate

# new cases of disease (during time interval)

/

avg population during time interval

point prevalence

# current cases (present & past) (at point in time)

/

population (at same point in time)

period prevelance

# current cases (past & present) (over period of time)

/

avg population

prevalence

incidence rate # new cases/start population) x avg duration of disease

relative risk

relative change ratio in incidence (associated with presence of exposure)

risk of disease in exposed → risk of disease in unexposed

includes 1 = strong association

RR (relative risk equation)

RR <1?
RR >1?

= incidence of disease in exposed/unexposed

RR <1 = % exposed less than unexposed

RR >1 = % exposed greater than exposed

odds ratio

measure of association in which study type?
when interpreted as RR?

odds of disease exposed → odds of disease unexposed

measure of association in case-control studies

can be interpreted as RR if disease is rare (<10%)

inference by association

Inferred or observed?

observed! (not inferred!)

causal inference

Inferred or observed?
causal relationship: exposure as direct factor? exposure determines outcome?
what is required for causal relationship?

inferred! (not observed!)

exposure does not have to be a direct factor for causal relationship

exposure cannot determine an outcome as a causal relationship
Must use…

• internal validity (NOT external)

• specificity: E → O

• temporality: E occurs before O

• biologic gradient: proportion E increases compared to O

• biological plausibility: amount E can be related to O

• coherence: design similar to other studies

• consistency: findings similar to other studies

• analogy: outcome is similar to other studies

• strength of associations

• considers alt explanations

• experimentation

necessary vs sufficient exposure

necessary?
sufficient?
both?
neither?

necessary: required for outcome, cannot cause on its own

sufficient: not required for outcome, can cause on its own

both: is required for outcome, and can cause on its own

neither: not required for outcome, cannot cause on its own

internal validity

what is it?
3 factors?

established causal relationship

1. bias

   • selection bias: individuals not representative of target population

   • information bias: key factors inaccurately measured, reported, classified

2. confounding: third factor relating exposure & outcome

3. chance: i.e., statistics

information bias

misclassification?
prevented by?
types?

misclassification

• differential: dependent on outcome/exposure, effect unknown

  • systematic, usually human error

  • effect unknown (worse T^T)

  • different between the two group

• non-differential: not dependent on outcome/exposure, attenuation of associations

  • random, usually machine error

  • weakens associations

  • same between the two groups

→ prevented by: blinding!!

• interviewer bias

• recall bias

• observer bias

• laboratory error

chance

null hypothesis?
alt hypothesis?
p-value?
confidence interval?

null hypothesis (Ho): association due to chance
true = no association

alt hypothesis (Ha): association not due to chance

p-value: if Ho true, probability of strength of data against Ho
p-value <0.05 = strong evidence against Ho

confidence interval: interval that true value is likely found (95%)
→ more informative than p-values

• wider: greater variability and/or small sample size

• narrower: smaller variability and/or large sample size

association testing: errors

type I?
type II?

type I: false positive
→ i.e., multiple testing

• rejects Ho (wrong) → no association

• accepts Ha (wrong) → association found

type II: false negative
→ i.e., small sample size, difference size, & significance level

• accepts Ho (wrong) → there is association

• rejects Ha (wrong) → association not found

external validity

what?

generalizability: inferences from one population can be extended to another

reproductive epidemiology

study of distribution and determinants of reproductive processes

defined by outcomes (fecundity & fertility)

critical vs sensitive windows

critical: leads to structural defects

• occurs in early pregnancy

• increasing metabolic capabilities

• cellular proliferation & development

sensitive: leads to functional defects

• occurs in mid-late pregnancy

• susceptible to non-structural defects

fecundity

what?
male vs female?

biological capacity of males and females for reproduction

male outcomes

• puberty, libido/hormones

• semen quality

• urologic health

• andropause

female outcomes

• puberty, libido/hormones

• menses & ovulation

• gynecologic health

• contraception

• menopause

fertility

relation to fecundity?
primary vs secondary?
clinical classifications?

demonstrated fecundity
fecundity = necessary, insufficient → fertility

primary: no prior conceptions
secondary: at least one prior conception

fertile: <6 months
sub-fertile: 6-12 months
infertile: >2 months

general fertility rate

fertility

# live births per 1000 women 

total (completed) fertility rate

fertility

# of births per woman

sex ratios (measuring fertility)

primary vs secondary?

fertility

primary: # conceptions, endogenous cases
secondary: # live births, exogenous cases

teratology

teratogens?

study of congenital malformations (structural birth defects)

teratogens: genetic

teratogens: environmental

• infections (e.g., CMV, syphilis, rubella, HIV, EBV)

• maternal toxic exposures (e.g., substance use, vitamin A, chemicals)

• maternal conditions/environment (e.g., diabetes, obesity, fever, thyroid disorder, occupation)

congenital malformations

structural birth defects (present at birth)

developmental abnormalities

adverse environmental exposure: consequences
5 levels

1. no effect
   least severe

2. functional defect

3. IUGR (growth abnormality)

4. structural defect

5. death
   most severe