OPT 323 Epidemiology Terms and Studies

What are the 2 assumptions of epidemiology?

diseases and risk factors DON'T occur at random

AND

disease is NOT randomly distributed throughout a population

What are the 3 main practical goals of epidemiology?

describe the burden of disease, disability, or death within a population

predict the probability that an individual within a defined population will develop a disease

compare the disease frequencies/”factors” within or between populations (or between exposed and unexposed groups)

What is epidemiology?

study of distribution and determinants of disease in a human population

What is the exposure?

entity = toxical chemical, micro-organism

OR

behaviour = where one works, socializes

OR

attribute = age, race, sex, genes

NOTE: these may be associated with an increased OR decreased occurence of disease

What is the outcome?

disease state, event, or behaviour that is under condition

What are 3 main shortcomings of epidemiology?

difficulty in measuring exposure = duration, some things are hard to measure

separating causal contributions of exposure from other exposures to remove confounds

difficulty assessing role of multiple small risks and how they interact

What is the main impractical goal of epidemiology?

does exposure A cause outcome B?

establishing causation is always our ultimate goal

What are Hill's 9 causal criteria?

temporality = exposure must be BEFORE outcome

strength of association

biological gradiant = dose-response

plausability and mechanism

constistency and replication

specificity

coherence

experimental evidence

analogy = alternate explanations

What is the ideal design for an epidemiological study?

counterfactual model = have 1 singular population and compare the measured incidence if exposed vs not

What is the real design for an epidemiological study?

have 2 populations (each as similar as possible), 1 exposed and 1 not, then measure incidence in both

What is the difference between an experimental vs observational epidemiologic study?

experimental = exposure status assigned by researcher

observational = exposure status NOT assigned by researcher

What is a randomized controlled trial experimental study?

1 group of very similar people without outcome are randomly assigned either a treatment/exposure or placebo, then outcome is compared after

What does it mean if a RCT is double or triple blinded?

double = subject and examiner do not know who belongs to what group

triple = statistician also does not know

What are some advantages of a RCT experimental study?

standard approach to test new treatments (gold standard)

prospective = we look to see how things progress temporally

exposure is assigned

What are some disadvantages of a RCT experimental study?

time-consuming

expensive

same effect in real world?

non-compliance

ethical issues (life-threatening conditions, who gets the placebo)

What is the difference between analytic vs descriptive observational studies?

analytic = WHAT, HOW, WHY = comparing groups helps establish cause or risk factors of exposure and outcomes

descriptive = WHERE, WHEN, WHO = identifies patterns in cases and populations, including rates and risk

What is incidence rate?

# of new cases that develop in a population at risk during a given person-time period at risk = expressed as a rate (comparing 2 quantities with different units)

How do we calculate incidence rate?

# new cases / person-time at risk

What is cumulative incidence?

risk of acquiring disease within a defined time period, expressed as a percent

How do we calculate cumulative incidence?

= # new cases / total population at risk x 100

NOTE: longer time period = higher CI

What is prevalence?

# of current cases that are present in a population at risk at a given time point / snapshot = expressed as a proportion (comparing 2 quantities of the same thing, parts of a whole)

How do we calculate prevalence?

= incidence rate x disease duration

What is a population study?

analyzes and predicts populations trends and challenges = often used to establish risk of an individual in a population getting a disease

What is a cohort analytic (observational) study?

observe a population of people with and without a certain exposure, and monitor them over time prospectively to see if exposed vs non-exposed develop the outcome

What are some advantages of a cohort analytic (observational) study?

prospective with a comparison group

info on multiple exposures and/or multiple outcomes

suitable/efficient for studying rare exposures

temporality = exposure before outcome

calculation of outcome incidence in exposed and unexposed subjects

allow for calculation of risk

What are some disadvantages of a cohort analytic (observational) study?

relatively expensive

time consuming = study subject drop-out, prolonged follow-up

not suitable/efficient for studying rare (or very “delayed”) outcomes

What is a true positive?

(+)exposure

(+)outcome

What is a false positive?

(+)exposure

(-)outcome

What is a true negative?

(-)exposure

(-)outcome

What is a false negative?

(-)exposure

(+)outcome

What is a case-control analytic (observational) study?

retrospective analysis of people with and without an outcome (cases and controls) to check for exposure in each group

What are some advantages of a case-control analytic (observational) study?

comparison group

relatively inexpensive

quick = already have data

very suitable/efficient for studying rare (or very “delayed”) outcomes bc easy to find cases that already exist

can study outcomes with long latency

can test multiple exposures for a single outcome

generate exposure “odds ratio”

What are some disadvantages of a case-control analytic (observational) study?

retrospective

control selection can be troublesome

exposure may be difficult to measure

subject to bias (e.g. recall)

temporality may be equivocal

cannot study multiple outcomes (diseases)

cannot directly determine incidence or risk

What is a cross-sectional desriptive (observational) study?

assess exposure and outcome simultaneously at a snapshot in time = compares subjects with exposure and subjects with exposure and outcome

What are some advantages of a cross-sectional desriptive (observational) study?

relatively inexpensive

quick (e.g. surveys)

practical for outcomes (diseases) of long duration and/or prolonged, progressive onset

practical for biological characteristics that do not change over time

establish prevalence

What are some disadvantages of a cross-sectional desriptive (observational) study?

no cause and effect

no temporality

can miss rare exposures/outcomes (diseases)

can miss outcomes (diseases) of short duration and/or having sporadic manifestations

not useful for rapidly emerging outcomes (diseases)

What is an ecological desriptive (observational) study?

analyzes groups/populations rather than individuals, trying to determine relationship between exposure and outcome

What are some advantages of an ecological desriptive (observational) study?

relatively inexpensive

quick (can sometimes use existing data)

big picture (population trends, broad social processes)

may detect previously unknown/unappreciated associations (as an exposure could be more variable across groups than within groups)

What are some disadvantages of an ecological desriptive (observational) study?

cannot establish causation

unable to control for confounding variables

ecological fallacy = trying to infer individual-level correlation/causation from group-level associations

What is a case report/case series desriptive (observational) study?

documents one or a few unusual medical occurrences

What are some advantages of a case report/case series desriptive (observational) study?

piques interest and raises awareness for a potential problem

What are some disadvantages of a case report/case series desriptive (observational) study?

anecdotal and non-scientific

no comparison group

What is a meta-analysis / systematic review?

summarizes multiple studies (even if they are conflicting) to draw a deduction or determine an overall trend

How do we perform a meta-analysis?

1. decide exclusion and inclusion criteria for studies

2. find all relevant studies

3. extract required info

4. do the meta statistics

5. interpret results

What is the hierarchy of study strength?

meta-analysis

systematic review

RCT

cohort studies

case control studies

case reports

animal research

What types of studies get a grade A for study design strength?

meta-analysis

systemiatic review

RCT

diagnostic studies that have a wide population, good reference standard, no case control studies

What types of studies get a grade B for study design strength?

RCT with weaker design

cohort studies (retrospective or prospective)

diagnostic studies that have ONE of narrow population, sample doesn't reflect population, poor reference standard, NOT blinded, may use case control studies

What types of studies get a grade C for study design strength?

case control studies (retrospective or prospective)

diagnostic studies that have TWO+ of narrow population, sample doesn't reflect population, poor reference standard, NOT blinded

strong study designs but with substantial uncertainty about conclusions or doubts about generalizations, bias, research design, sample size

nonrandomized trials

What types of studies get a grade D for study design strength?

cross sectional studies

case reports/series

reviews

position papers

expert opinion

reasoning from principle

What is absolute risk?

risk of a health event happening over a certain time period, reflecting the overall public health impact of the risk factor

How do we calculate absolute risk?

# of health events in a population / total populations at risk

What is relative risk?

risk of experiencing a health event based on group membership, reflecting how strong the association is between factor and disease

How do we calculate relative risk?

risk of population A / risk of population B

What does it mean if relative risk is 1 or >1?

RR > 1 = population A has a higher risk

RR = 1 = risk is equal between populations

How do these graphs show the difference between prevalence and relative risk?

prevalence = most diabetic ret cases are in white people

relative risk = hispanic people have a higher relative risk of developing AMD

What is morbidity?

reported cases of disease, illness, and disability = shows current health status of population

What is mortality?

# of deaths in a population

What does sporadic mean?

occasional, irregular occurence such as...

food-borne disease

Creutzfeld-Jakob disease

Acanthamoeba keratitis

What does endemic mean?

disease frequency is maintained without external inputs such as...

cataracts

Rx

common cold

What does epidemic mean?

sudden increase in cases above endemic/baseline level = typically > 1.645 standard dev above seasonal baseline is considered threshold

What are the 2 types of epidemic?

common source = not transmissible like chlorine gas leak, obesity, diabetes, myopia

propogated = transmissible between people like COVID

What does pandemic mean?

epidemic affecting a wide area, such as...

HIV

flu in 2009

COVID in 2019

Ex) the statement "14% of people develop PDR after a 20 year duration of DM" is an example of what type of public health measure?

cumulative incidence

Ex) the statement "8.6% of Americans > age 40 with DM have diabetic retinopathy" is an example of what type of public health measure?

prevalence

Ex) the statement "ambulatory eye injuries among US military personnel is 983/100,000 person-years" is an example of what type of public health measure?

incidence rate

Ex) the statement "the risk of retinal detachments in patients with lattice is < 1%" is an example of what type of public health measure?

absolute risk

Ex) the statement "10% of adults > age 60 in the US have ARMD" is an example of what type of public health measure?

prevalence

Ex) the statement "over a 15 year period, 8% of people > 75 years of age have late AMD" is an example of what type of public health measure?

cumulative incidence

Ex) the statement "2.5 million Americans suffer eye injuries annually" is an example of what type of public health measure?

incidence

Ex) the statement "AA are 3.29x more likely to be diagnosed with glaucoma than Caucasians" is an example of what type of public health measure?

relative risk

Ex) the statement "1 in 3 Americans have astigmatism" is an example of what type of public health measure?

prevalence

Ex) the statement "1.08 deaths per 100 million vehicle miles in the US in 2014" is an example of what type of public health measure?

incidence

What was the goal of the Orinda Vision Screening study (1954-1956)?

design the least expensive, least technical and most effective screening program for finding essentially all elementary-school children with vision problems

What was the study design of the Orinda Vision Screening study (1954-1956)?

hybrid study design = descriptive epidemiology + retrospective clinical trial

What were the 2 main outcomes of the Orinda Vision Screening study (1954-1956)?

modified clinical technique (MCT) = Snellen VA, retinoscopy with lens rack, CT, DO

>50% of children with previous eye care still failed the screening

What are some weaknesses of the Orinda Vision Screening study (1954-1956)?

not applicable to all populations

must be performed by clinicians, not laypeople

not the cheapest program

does not include visual skills testing

What was the goal of the Vision in Preschoolers study (1997)?

developing evidence-based guidelines for preschool vision screening tests to specificially target amblyopia, strabismus, significant Rx, unexplained reduced VA

What was the study design of the Vision in Preschoolers study (1997)?

experimental with 2 phases

What was phase I of the Vision in Preschoolers study (1997)?

eye care providers = refractive test (ret) and CT with crowded, linear Lea was best to assess strab

What was phase II of the Vision in Preschoolers study (1997)?

lay screeners and nurses = retinomax, suresight + stereo smile with single, crowded Lea was best for strab

What was the goal of the Framingham Eye Survey (1973)?

prevalence data gathered for common eye disease and disorders

What was the study design of the Framingham Eye Survey (1973)?

cross-sectional study with data taken from the original 1st generation heart study (which was a prospective cohort study)

What were the 3 main outcomes of the Framingham Eye Survey (1973)?

vision impairment is lower than initially thought

POAG prevalence is higher that initially thought

eye disease prevalence for other conditions such as CATs, DR, ARMD

What was the goal of the Baltimore Eye Survey (1985-1988)?

evaluation of glaucoma screening tests such as tonometry, VF, stereo fundus photos, PMHx, POHx

What was the study design of the Baltimore Eye Survey (1985-1988)?

prospective, cross-sectional study

What were the 2 main outcomes of the Baltimore Eye Survey (1985-1988)?

risk factors for POAG like race, IOP, HTN, perfusion pressure, FHx, ONH structure

visual impairment increases with age, black race, general health, and SES

What was the goal of the Beaver Dam Eye Study (1988-2008)?

prevalence and incidence of CATs, AMD, DR

What was the study design of the Beaver Dam Eye Study (1988-2008)?

prospective cohort study

What were the 3 main outcomes of the Beaver Dam Eye Study (1988-2008)?

cigarette smoking associated with CATs and AMD

retinal drusen and pigment associated with AMD progression

grading scale images

What was the goal of the Wisconsin Epidemiological Study of Diabetic Ret (1979)?

incidence and progression of diabetic complications

What was the study design of the Wisconsin Epidemiological Study of Diabetic Ret (1979)?

prospective cohort study

What were the 2 main outcomes of the Wisconsin Epidemiological Study of Diabetic Ret (1979)?

DR and kidney disease risk increases if HbA1c > 6.5%

higher BG, smoking, HTN associated with DM complications (retinopathy, nephropathy, neuropathy)

What was the goal of the Los Angeles Latino Eye Study (2000-2010)?

incidence and prevalence of DR, POAG, ARMD, CATs

What was the study design of the Los Angeles Latino Eye Study (2000-2010)?

prospective cohort study

What were the 2 main outcomes of the Los Angeles Latino Eye Study (2000-2010)?

60% of eye disease is undiagnosed in Latino populations

Latinos have highest risk of vision impairment, blindness, and developing DR

What was the goal of the Barbados Eye Studies (1986-2003)?

incidence, prevalence, and risk factors for POAG, CATs, DR, ARMD

What was the study design of the Barbados Eye Studies (1986-2003)?

simple random sample

What were the 2 main outcomes of the Barbados Eye Studies (1986-2003)?

highest prevalence of low vision/blindness than elsewhere in the world, largerly due to CATs and POAG

POAG prevalence differed by race, age, gender = older, black, males more affected

What was the goal of the Multi-Ethnic Pediatric Eye Disease study (early 2000s)?

determine prevalence of Rx, strabismus, amblyopia in Baltimore and LA children

What was the study design of the Multi-Ethnic Pediatric Eye Disease study (early 2000s)?

cross-sectional study