Biostatistics Flashcards
Cross-sectional study
assess frequency of disease (and related risk factors) at a PARTICULAR POINT IN TIME. Measures disease PREVALENCE. Cannot establish causality
Case-control study (retrospective)
“known disease” compares a group of people with disease to a group without disease. Looks for prior exposure or risk factor Measures OR
Cohort Study (prospective or retrospective)
“known exposure” compares a group with a given exposure or risk factor to a group without such experience. Looks to see if exposure increases the likelihood of disease. Measures RR
Sensitivity
TP/(TP+FN) Highly SeNsitive when Negative rules disease OUT “SNNOUT” Screening test
Specificity
TN/(TN+FP) Highly SPecific when Positive test rules disease IN “SPPIN” Confirmatory testing
Positive Predictive Value (PPV)
proportion of positive test results that are true positive TP/(TP+FP) Varies directly with prevalence/pre-test probability
Negative Predictive Value (NPV)
proportion of negative test results that are true negative TN/(TN+FN) Varies inversely with prevalence/pre-test probability
Incidence
looks at new cases incidence rate = # of new cases / # of people at risk
Prevalence
looks at all current cases = # of existing case / # of people at risk ~pretest probability
Odds Ratio
used for case control studies

Relative Risk
used in cohort studies

Attributable Risk

Relative Risk Reduction (RRR)
RRR = 1 - RR
proportion of risk reduction attributable to intervention
e.g. if 2% of patients who receive a flu shot develop the flu, while 8% of unvaccinated patient develop the flu, then:
RR = 2/8 = 0.25 and RRR = 0.75
Absolute Relative Risk (ARR)
Difference in Risk attributable to the intervention as compared to a control
e.g. if 8% of people who receive a placebo vaccine develop the flu vs. 2% of people who receive a flu vaccine, then:
ARR = 8% - 2% = 6% = 0.06
Number Needed to Treat (NNT)
NNT = 1 / ARR
(for the benefit of one patient)
Number Needed to Harm (NNH)
NNH = 1 / AR
who need to be exposed for a risk factor for 1 patient to be harmed
Selection Bias
Error in assigning subjects to a study group reulting in an unrepresentative sample.
Reduce bias: Randomization, correct comparison group
Recall Bias
Awareness of disorder alters recall by subjects; common in retrospective studies
Reduce Bias: decrease time from exposure to follow-up
Measurement Bias
Information is gathered in a way that distorts it
e.g. miscalibrated scale consistently overstates weight of subjects
reduce bias: use standardized method of data collection
Procedure Bias
Subjects in different groups are not treated the same
reduce bias: blinding and use of placebo
Observer-expectancy bias
Researcher’s belief in the efficacy of a treatment changes the outcome of that treatment (‘self-fulfilling prophecy’)
reduce bias: blinding and use of placebo
Confounding bias
When a factor is related to both the exposure and outcome, but not on the causal pathway –> factor distorts of confuses effect of exposure on outcome
e.g. pulmonary disease more common in coal workers than general population; people who work in coal mines also smoke more frequently.
reduce bias: multiple/repeated studies; crossover studies (subjects act as own controls); matching (patients with similar characteristics in both treatment and control goup)
Lead-time bias
Early detection is confused with increased survival
reduce bias: measure “back-end” survival (adjust survival according to the severity of disease at time of diagnosis)
Mean
average
most affected by ouliers
Median
Middle value
Mode
Most common value
least affected by outliers
Normal Distribution

Type I Error (alpha)
Stating there is an effect or difference when non exists (null hypothesis incorrectly rejected in favor of alternative hypothesis)
alpha = probability of making a type I error.
p = judge against a preset alpha (usually <0.05)
“false positive error”
Type II Error (beta)
Stating there is not an effect or difference when one exists (null hypothesis is not rejected when it is in fact false)
beta = probably of making a type II error
statistical power = 1 - beta
Increase power by: increasing sample size, expected effect size and precision of measurement
t-test
“tea for 2”
Check differences between means of 2 groups
ANOVA
Checks differences between means of 3 or more groups
Chi-Square
“chi-tegorical”
Checks differences between 2 or more percentages (%) or proportions of categorical outcomes
(not mean, like t-test)
Medicare
(medicarE for Elderly; medicaiD for Destitute)
Medicare:
Part A: hospitAl
Part B: Basic medical Bills (fees, dx testing)
Part C: (A+B) delivered by private Companies
Part D: prescription Drugs
APGAR
Appearance
Pulse
Grimace
Activity
Respiration
at 1 and 5 minutes
>6 good, 4-6 assist and stimulate, <4 resuscitate