Epi test 3

Necessary Cause

required for the disease to occur. More applicable to infectious disease and is caused by a single pathogen.

Sufficient Casual set

set of factors working together, requires joint action of component causes, each component plays a necessary role. More applicable to non-communicable diseases with multiple causes.

Ronald Ross

A British physician known for his work in identifying the malaria parasite's transmission through mosquitoes, significantly advancing the understanding of infectious diseases.

Tuberculosis

Top infectious killer in the world

1.5 million died in 2018

HIV and TB high correlation

Criteria for causation and which are different from Bradford Hill

Temporal relation- Does the cause precede the effect

Plausibility is the association consistent with other knowledge

Consistency- Have similar results to other studies

Strength- Is there a strong association between the cause and effect?

Dose- response relationship- Is increased exposure possible cause leas to the reduction of disease risk

Different

Reversibility- Does the removal of a possible cause lead to reduction of disease risk

Study design- Is the evidence based on a strong study design

Judging the Evidence- How many lines of evidence lead to the conclusion

Ability to prove causation

RCT

Cohort studies

case-control

cross section

Ecological

Bias?

Evidence quality?

strong

moderate

moderate

weak

weak

Risk of bias lowest in RCT and increases down the list highest in Cross sectional (ecological not on pyramid)

Evidence quality lowest in RCT and increases down the list highest in Cross sectional (ecological not on pyramid)

Systematic review

uses standardized methods to select and review studies on a particular topic

Meta-Analysis

Pools and re-analyzes data from other studies often used in tandem with systematic review

Upstream

root causes

ex. access to healthy foods, access to quality health care, reliable transportation, stable housing, and economic stability

Downstream

Effects

Ex. chronic disease, infection, mortality

Distal causes

factors that are further removed from the immediate health outcomes, often encompassing broader social and environmental determinants.

proximal cause

factors that are closely related to health outcomes, such as individual behaviors or specific medical conditions.

causal interference

The process of determining whether an observed association is likely to be causal.

Hierarchy of causes

some causal factors lead to exposure that are the direct cause of the disease

calculate the attributable fractions of smoking and asbestos exposure for lung cancer

AF of smoking among asbestos workers: (602-58)-602=0.904

AF of smoking among non-abestos workers: (123-11)/123=0.910

AF of eliminating asbestos exposure among smokers:(602-123)/602=.0796

AF of eliminating asbestos exposure among non-smokers:(58-11)/56=0.810

Why can attributable fractions add up to more than 100%

The fractions can add up to more than 100% because they are each dependent on the other.

Which criteria of causation would you want to establish first

Temporal relationship

Why is time-series analysis of short term association considered an acceptable method to assess causality

a short time frame reduces chance of other confounding variables

What is meta-analysis and which conditions have to be met

Meta- analysis combines data from more than one study in order to achieve more stable and precise conclusions concerning causal assocations. To use this method each study needs to have used the same exposure and health outcomes varaibles and the basic population characteriscs (age, sex, etc. )

Sample

subset of the population

We dont really care about the sample but about describing the whole population.

Biostats gives us tools to determine how well the sample describes the whole population

Binary Variable

Two categories eg yes or no

Nominal variables

Multiple categories which are not numbered or ordered eg hair color

Ordinal variables

multiple categories which are ordered e.g. level of education

Continuous variables

Numeric values with theoretically infinite numbers of value e.g. height and weight

Discrete

Numeric values with finite numbers e.g. number of siblings

categorical variables

Binary, nominal, ordinal

Quantitative variables

continuous and discrete

Mean

average value

Standard deviation

What percent falls in one and two SD from the mean

How is SD related to variance

numerical value which tells us how spread out the normal distribution is from the sample population

One SD is 68%

Two SD is 95%

SD= Square root of the variance

Variance

is the distance between points and the mean

95% CI

What would happen to the CI if the sample size is reduced

Typically includes the mean

95% of the true population mean is in the range x-y

CI would be larger

Standard Error

how spread out the sample means would be, if we were drawing multiple random samples from the same population and measuring the distribution of those sample means

H_o

Null hypothesis there is no difference between two things

H₁

Alternative statement that there is a difference

When do we reject the H_o

observed outcome is rare or unlikely p<0.05

Null Hypothesis rejected when it is false

correct

true positive

probability =1-B

AKA Power

Null hypothesis is rejected when it is true

Type 1 error

false positive

Probability = a

Null hypothesis is not rejected when it is true

correct decision

true negative

probability 1-a

Null hypothesis not rejected when it is false

Type II error

False negative

Probability =B

Binary outcomes

Two possible results

chi-squared or logistic regression

for all statical methods p-value or CI can be used to accept or reject null hypothesis

Continuous outcomes

wide range of numerical data like temperature

t-test or linear regression

for all statical methods p-value or CI can be used to accept or reject null hypothesis

Time to event

Time elapsed before an event occurs

survival analysis (Kaplan-Meier Curves)

for all statical methods p-value or CI can be used to accept or reject null hypothesis

P value= 0.59 do we accept or reject H_o

Accept null hypothesis Treatment did not have significant effect

P value = 0.01 do we accept or reject H_o

Reject Null hypothesis treatment did have a significant effect.

If CI overlap do we accept or reject H_o

Accept the null hypothesis no significant effect

Causation in chronic disease

Chronic disease is increasing in high, middle and low income countries

major cause of morbidity in most of the world

No single cause

4 levels of prevention

Primordial, primary, secondary, and tertiary

Primordial prevention

addressing social determinants of health

Targeted towards the entire population

ex. income, access to care

Primary prevention

Addressing specific causal factors, education, laws, vaccines, sanitation

Suseptible population aims to stop the disease form ever occuring

Secondary prevention

Treating disease in its early stages to shorten duration

Healthy appearing individuals

early detection and treatment

screening to find early stages of disease

Tertiary prevention

Late stages to reduce complications

Targets symptomatic patients

screening in public health

process of using test on a large scale to identify the presence of disease in apparently healthy people

Sometimes establishes diagnosis, more often identifies risk factor like high BP

Disorder must be well defined, known, long period between first signs and development, test simple safe and cheap, treatment should be accessible and affordable.

How does sensitivity and specificity relate to these variables

Sensitivity= probability of a positive test in people with the diseae a/a+c

Specificity= probability of a negative test in people without the disease d/(d+b)

Sensitivity

Proportion of people with disease who have a positive test result

high sensitivity reduces false negatives

Specificity

proportion of people without disease who have a negative test result

Positive predictive value

probability of the person having the disease when the test is a positive a/a+b

Negative Predictive value

Probability of the person not having the disease when the test is negative d/d+c

Odds ratio

a statistical measure that compares the likelihood of an event occurring in a group exposed to something versus a group that was not exposed.

(a/b)/(c/d)

population had ___ times the odds of having outcome if exposure than compared to group without exposure

Risk ratio

the ratio of the probability of an outcome in an exposed group to the probability of an outcome in an unexposed group.

RR=[a/(a+b)]/[c/(c+d)]

population had ___ times the risk of having outcome if exposure than compared to group without exposure

Epidemic

refers to an increase, often sudden, in the number of cases of a disease above what is normally expected in that population in that area

Pandemic

refers to an epidemic that has spread over several countries or continents, usually affecting a larger number of people

Endemic

refers to a relatively stable pattern of occurrence in a given geographical area or population group at relatively high prevalence and incidence.

Communicable disease

caused by transmission of specific pathogenic agent to a susceptible host, either directly from infected humans or animals or indirectly through vectors, airborne particles or vehicles.

vector

living organism carrying infectious agent on/in its body

vehicle

non-living substance e.g. soil, water, air carries infectious agent to a new host

Infectious disease

same as communicable disease

contagious disease

spread directly between humans without intervening vector or vehicle

R-naught

Number of people that one sick person will infect on average

What three things affect the transmission process

Environment, host, and agent

Linear regression

dependent variable is continuous

logistic regression

dependent variable is binary

survival analysis

dependent variable is a time interval until a prespecified event