Statistics Flashcards by Luke Cielonko

True positives/Total amount with disease

Sensitivity

How well did you know this?

Not at all

Perfectly

true negative/total negative

Negative predictive value

How well did you know this?

Not at all

Perfectly

True negatives/total disease negatives for disease

Specificity

How well did you know this?

Not at all

Perfectly

True positives/Total positives

Positive predictive value

How well did you know this?

Not at all

Perfectly

1/attributable risk

Number needed to harm (nhh)

How well did you know this?

Not at all

Perfectly

A/ (A+B) - C/ (C+D)

Attributable risk

How well did you know this?

Not at all

Perfectly

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

Relative risk

How well did you know this?

Not at all

Perfectly

New cases divided by total population at risk

Incidence

How well did you know this?

Not at all

Perfectly

Total cases divided by total population

Prevelance

How well did you know this?

Not at all

Perfectly

What is the relationship between prevalence and incidence?

Prevalence = Incidence (duration)

How well did you know this?

Not at all

Perfectly

High incidence, low duration. What is prevalence?

Prevalence is similar to incidence

How well did you know this?

Not at all

Perfectly

Low incidence, high duration. What is prevalence?

Prevalence increased relative to incidence

How well did you know this?

Not at all

Perfectly

High incidence, high duration. What is prevalence?

High prevalence

How well did you know this?

Not at all

Perfectly

Directionality of a study design that begins with outcome

Backward directionality

How well did you know this?

Not at all

Perfectly

Directionality of a study design that begins with exposure

Forward directionality

How well did you know this?

Not at all

Perfectly

What are the three types of observational studies?

1) Cross-sectional study
2) Cohort study
3) Case control study

How well did you know this?

Not at all

Perfectly

What is the directionality of 1) a cross-sectional study, 2) a cohort study, and 3) a case control study

1) Non-directional
2) Forward
3) Backward

How well did you know this?

Not at all

Perfectly

Variation of disease occurrence among populations

Distribution

How well did you know this?

Not at all

Perfectly

What observational study type allows for the calculation of relative risk?

Cohort study

How well did you know this?

Not at all

Perfectly

What observational study type allows for the calculation of odds ratio?

Case control study

How well did you know this?

Not at all

Perfectly

What observational study type allows for the determination of prevalence?

Cross-sectional study

How well did you know this?

Not at all

Perfectly

Write out a 2x2 table used for the determination of RR or OR

See paper

How well did you know this?

Not at all

Perfectly

What is the calculation for relative risk in a 2x2 table?

A/(A+B) divided by C/(C+D)

How well did you know this?

Not at all

Perfectly

What does a RR of 1 mean?

There is no association between exposure and outcome

How well did you know this?

Not at all

Perfectly

What does a RR of >1 mean?

The exposure increases the risk of the outcome

What does a RR of <1 mean?

The exposure decreases the risk of the outcome

What should a confidence interval never include within its range?

1) 1 | 2) makes it statistically insignificant

Measure of how likely the observed assication would be to occur by chance alone, in absence of a true association

p-Value

Bias caused by a non-random assignment to participation in a study group or as a selective loss of subjects

Selection bias

Bias caused by a knowledge of presence of disorder that is altered by recall from subjects; common in retrospective studies

Recall bias

Bias caused by subjects not representing the general population

Sampling bias

Bias caused by information being gathered at an inappropriate time

Late-Look Bias

Bias cause by subjects in different groups being treated differently

Procedure bias

Bias caused by the mixing of effects between the exposure, the disease, and a third unknown factor

Confounding bias

Bias caused by early detection being confused with increased survival

Lead-time bias

Bias that occurs when a researcher's belief in the efficacy of a treatment changes the outcome of that treatment

Observer-expectancy effect

Bias that occurs when the group being studied changes its behavior due to knowing they are being studied

Hawthorne effect

What is the difference between an observational and an experimental study?

Experimental induces an exposure; observational just observes an exposure

Where does nicotine act?

Binds to the VTA and nucleus accumbens

Prevention type aimed at preventing disease or injury from occurring and reducing incidence of disease

Think: PDR (Prevent, Detect, Reduce) | Primary prevention

Prevention type aimed at delaying or preventing the occurrence of full-blown or symptomatic disease by early detection

Think: PDR (Prevent, Detect, Reduce) | Secondary prevention

Prevention type that is focused on limiting the extent of impairment

Think: PDR (Prevent, Detect, Reduce) | Tertiary prevention

Proportion of individual screened positive by the test who actually have the disease

Positive Predictive Value

Proportion of individual screened negative by the test who actually do not have the disease

Negative Predictive Value

Study sample with Small number of healthy volunteers

Phase I

Study sample with small number of pt. with disease of interest

Phase II

Study sample with large number of pt. randomly assigned either to the treatment under investigation or to the best available treatment

Phase III

Postmarketing surveillance trail of pt. after approval

Phase IV

Longer tail on the right; how are mean, median, and mode related

1) Positive Skew | 2) Mean > median > mode

Longer tail on the left; how are mean, median, and mode related

1) Negative skew | 2) Mode>Median>Mean

Relative Risk Reduction

1 - RR = RRR

Absolute reduction in risk associated with a treatment as compared to a control

Absolute risk reduction

1/absolute risk reduction

Number needed to treat

1/absolute risk

Number needed to harm

Error caused by a difference that did exist

Type II error (beta) | Think: Beta = blind to a difference

Error caused by a difference that did not exist

Type I error (alpha) | Think: alpha = you sAw a difference that did not exist

Checks difference between the means of 2 groups

T test | Think: T=Two

Tests checks differene between 2 or more percentages or proportions

Chi-square

Checks difference between the means of 3 or more groups

ANOVA

Determines strength of a linear correlation between two variable

Pearson's correlation

What does a phase I trial consist of? phase II? phase III? phase IV?

1) Small group of volunteers with no disease 2) Small group of pt. with disease of interest 3) Large group of pt. with disease that are divided by new treatment and best known available treatment 4) Postmarketing surveillance trial of pt. after trial approval

Power (1 - beta)

Probability of rejecting null hypotehesis

What increases Power (1-beta)

1) Increase sample size (POWER in Numbers) 2) Increase expected effect size 3) Increase precision of measurement

When an effect on the main exposure on the outcoume is modified by the presence of another variable (aka: doing a test on a drug and separating groups into smokers/nonsmokers, and then seeing the result)

Effect modification

Tendency of a study population to affect an outcome due to knowledge of being studied

Hawthorne effect

Selection bias created by selecting hospitalized patients as the control group

Berkson's bias

Researcher's beliefs in the efficacy of a drug that can potentially affect the outcome

Pygmalion effect

Likelihood ratio of 2, 5, and 10 = what % probability of disease

1) 15% 2) 30% 3) 45%