BIAS

ATTRIBUTABLE RISK

(a / (a+b) ) - (c / (c+d) ) (same as RR, but subtract instead of divide) The amount of proportion of disease incidence that can be attributed to a certain exposure and can be calculated by exposed groups or the total population.

Which of the following is the detmiernant of pop health in the category of living + working conditions a. health care services B. individual beh C. Social, family, and community ntwk D. Election Cycle E. The length of sumemr

A

Selection Bias

A form of bias that can result when the selection of subjects into a study or their likelihood of being retained in the study leads to a result that is systematically different from what you would have gotten if you had enrolled the entire target population (i.e. the relationship between exposure and outcome) Also seen as a distortion that results by which subjects are selected in to the study population

Differential Misclassification Bias

A form of misclassification bias where the rate of misclassification differs in DIFFERENT STUDY GROUPS that can lead to either an APPARENT ASSOC WHEN THERE IS NONE or or APPARENT LACK OF ASSOC WHEN ONE IS THERE *you can also say that this type of bias, biases the measure of association either towards or away from the null

CHANCE

A random error appearing to cause an association between an exposure and outcome

Information Bias

A systematic error in measuring outcome or exposure that results in differential accuracy of information between the comparison groups. Also seen as the distortion that results from inaccuracies in the measurement and incorrect classification of subject characteristics

WISH BIAS

A type of bias given to individuals who have developed a disease and are trying to answer the question "Why me?" This causes them to (often unintentionally) deny certain exposures related to lifestyle. Ultimately this bias is a result of an error in the design or conduct of a study. This is also a subcategory of reporting bias.

Bias From Surrogate Interviews

A type of bias occurring when family members are approached on behalf of a recently deceased participant (of a study) that provides information about that participant. The information provided includes: case's employment history, diet, and other exposures and characteristics. The family member generally is a significant other or child.

REPORTING BIAS

A type of bias when multiple different subjects are reluctant to report an exposure they are aware of because of attitudes, beliefs, and perceptions.

HEALTHY WORKER EFFECT

A type of selection bias where employed people have a more favorable morbidity and mortality experience than the general population. It tends to dilute the results

Withdrawal / Loss To Follow-Up

After randomization (if applicable), this occurs when participants do not respond after a period of time in the study. They stop participating (with or without notice to the researchers) and can create bias within the data collected.

bias

Any systematic error in the design, conduct or analysis of a study that results in a mistaken estimate of an exposure's effect on the risk of disease

Pick achievement by CDC in 20th ce Control of infectious dx Legalization of marijuana elimination of opioid crisis HIV in US Affordable Health Insurance

Control of Infectious Dx

There is adjustment for interaction T/F

False - once interaction is detected, stratified analysis must be used to obtain stratum-specific results

T/F indirect steps between exp + dx (in causal ptwy) are confounding variable

False they are not

Nondifferential Misclassification Bias

Form of misclassification bias that results from degree of inaccuracy that characterizes how information is obtained from any study group - cases + controls or exposed + nonexposed persons. -Not related to exp status or dx status - INHERENT PROBLEM IN DATA COLLECTION METHODS that results in RR or OR to be SHIFTED TOWARDS 1.0 (diluted) (less likely to detect an association) *you can also say that this type of bias, biases the measure of association towards the null

ADDITIVE MODEL

In interaction, this is the model used when the risk factors being examined are independent of one another. Thus, the effect of one exposure is added to the effect of the second exposure.

MULTIPLICATIVE MODEL

In interaction, this is the model used when the risk factors being examined are not independent of one another but there effect builds on the other. Thus, the effect of the second exposure multiplies the effect of the first exposure.

HOW TO ADDRESS CONFOUDNING IN DESIGN

Individual Matching Group Matching *Matching generally occurs to remove a suspected possible confounder

NON_RESPONSE BIAS

Individuals who do not respond to requests to be studied generally have different baseline characteristics than responders

VOLUNTEER BIAS

Individuals who volunteer for a study may possess different characterizes than the average individual in the source population

Is it interaction or confounding: Crude RR/OR Diff from Stratified RR/OR AND Stratified RR/OR are different

Interaction

What supersedes confoudning?

Interaction Effect Modification

IS CONFOUNDING PRESENT? Crude RR /OR are same as Stratified RR/OR

Nope

Is confounding an all-or-none phenomenon?

Nope

METHODS OF BIAS CONTROL

Randomization Treatment Assignment Ratio (allocation Ratio) Stratification Blocking Masking (Blinding) Standardization

HOW TO ADDRESS CONFOUNDING IN ANALYSIS

Stratification Adjustment

Describe Synergism + Antagonism w/ Additive Interaction

Synergism: Add Risk factors together to see an effect greater than the sum of independent effects Antagonism: add risk factors together to see a deficit c/c to the sum of the independent effects

Describe Synergism + Antagonism w/ Multiplicative Interaction

Synergism: Multiply Risk factors together to see an effect greater than the sum of independent effects Multiply: add risk factors together to see a deficit c/c to the sum of the independent effects

Bias in Abstracting Records

The bias that occurs when conducting a study where the researcher has to obtain different forms of information (ex. medical records, local public health records, etc.) that may be insufficient, located in multiple different areas, or any other reason that would make pulling these records produce incomplete data

CONFOUNDING

The distortion of association between an exposure and disease outcome by a third variable that may be present in any study design because it is not a result of the study design. This third variable is a risk factor that is associated with both the exposure and outcome and is not in the causal pathway of the association.

SURVEILLANCE BIAS

This bias arises when patients in one exposure group have a higher probability of having the study outcome detected, due to increased surveillance, screening, or testing of the outcome itself, or an associated symptom.

Prevalence - Incidence Bias (Neyman's Bias)

This is a selection bias where the very sick or very well are erroneously excluded from a study. Excluding patients who have died makes the results seem less severe while excluding patients who have recovered makes the results seem more severe.

NONCOMPLIANCE

This occurs after randomization where participants in a study are not completing the actions or following instructions that are asked of them. Participants may or may not tell researchers if they are doing what is being asked of them.

RECALL BIAS

This type of bias is a systematic error that occurs when participants have differences in accuracy or completeness when remembering past events or experiences.

Interviewer Bias / Bias in Interviewing

This type of bias occurs because of the intentional or unrecognized behavior of an interviewer (only) that lets their own beliefs influence how they record data, whom they will interview, and/or what and how questions would be asked.

Outcome Assessment Bias (Diagnostic Suspicion Bias)

This type of bias occurs from interviewers and/or data analysts where the interviewer/analyst is aware of the exposure and/or outcomes and lets this knowledge influence their diagnosis/determination of what the outcome is.

Exclusion Bias

This type of bias occurs when investigators apply different eligibility criteria to the cases and to the controls regarding which clinical conditions in the past would permit eligibility in the study + which would serve as the basis for exclusion

Berkson's Bias

This type of selection bias can arise when the sample is taken not from the general population but from a subpopulation. An example is sampling cases and controls from a hospital instead of from a community.

Systematic / Differential Error

Use of an invalid measure that differentially misclassifies study groups; can create or hide an association Can also be defined as any error different than random error

Nondifferential Error

Use of invalid measure that equally misclassifies study groups; dilutes an association (this means that we as experimenters are less likely to detect an association even if one does exist)

SYNERGISM POSITIVE IXN

When an interaction produces an effect greater than what we would expect

ANTAGONISM NEGATIVE IXN

When an interaction produces an effect less than what we would expect

OVERALL IXN

When the incidence rate of disease in the presence of two or more risk factors differs from incidence rate expected to result from their individual effects

IS CONFOUNDING PRESENT? Crude RR/OR are different from RR/OR

YES ONLY IF the stratified RR/OR are equivalent

What do you do if there is confdouning?

adjust

Mediator

associated w/exp, associated w/ outocme, in causal ptwy of association exp --> _______ --> outcome

Pick the correct option regarding evidence based PH framework A. 3 types of data collection : primary, secondary, tertiary B. Quantifying the issue after community assessment facilitates the dvp't of a summary of the issue c. social ntwk searching is an optional step in the framework; helps us understand w/the evidence tells us D.It is critical to dv'p + prioritize programs + policy options that can reach goals _ objectives E. 1 of the common approaches for a plan of action is phase in total -> revision implementation

b

In the study of a dx in which all cases that were develoepd were ascertained, if the RR for the assoc btw a factor + dx is 1.0+ then a. there is no assoc btw factor + dx b. the factor protects against development of dx c. either matching or randomziation has been unsuccessful d. there is positive association btw factor + dx e. there is either no association or a negative assoc btw factor + dx

d

Pick the correct description of epidemiology A. there are a total of 4 different levels of dx in epidemiology b. epidemic is the usual dx prevalence with a given geographic area c. epidemic is the usualy prevalence of dx with a given pop group d. CFR represents the deaths from a certain dx in the gen pop during a specific time pd e. surveillance data are often numerator data

e c is missing the in excess

What do you do if there is interaction?

report separate measures for levels of covariate

Interaction effects differs in

size/direction among strata