Biostatistics I Flashcards
check on learning
All people in a defined geographical area (e.g., Montgomery county, VA) or with certain defined characteristics (e.g., age>65 years & having a thyroid nodule).
population
Why to define a population?
We know the group of individuals about whom a statement is made.
We could get some clues on how to get a sample.
How to determine a population?
from research question
available resources
We know the group of individuals about whom a statement is made.
We could get some clues on how to get a sample.
Sample
why use a sample?
less costly
more quick and timely
reliable
what is the most important criterion for a sample?
representative of the population
a numerical summary of a population?
parameter
a numerical summary of sample
statistic
Example 1: the average age of all adults living in Blacksburg (μ)
Example 2: the proportion of adults with Type II diabetes living in Blacksburg (p)
parameter or statistic
parameter
Example 1: the average age of the 30 samples (𝑥̅)
Example 2: the proportion of Type II diabetes patients in the sample (𝑝̂)
parameter or statistic
statistic
Deals with summarizing data from a sample via graphical or numerical approaches
descriptive statistics
Refers to making decisions and/or predictions about a population (parameter) based on a sample (statistic)
inferential statistics
Examples:
Histogram
Scatterplot
Mean/Median/Mode
descriptive statistics
Examples:
Confidence Interval
Hypothesis Testing
inferential statistics
inferential statistics are related to inferences about?
population parameters
a sample in which every unit in the population has a fixed and known or equal chance of being selected.
what does this allow?
random sample
allows powerful inferences about the population without having to observe the entire population
what are the types of sampling strategies?
what are the types of these?
probability sampling
non probability sampling
simple random, systemic, stratified, cluster (involves random selection and gold standard for generalizability)
*probability sampling
convenience sampling and quota sampling
*non-probability sampling
Every member of the population has an equal chance of being selected
Probability Sampling _ Simple Random Sampling
It is a sample that is selected by a predetermined periodic process
Probability Sampling _ Systematic Sampling
When known categories (strata) exist in a population, and a simple random sample can be taken from each of these “strata.”
Probability Sampling _ Stratified Sampling
It is a random sample of natural groupings (clusters) of individuals in the population.
Probability Sampling _ Cluster Sampling
Get a list of all the people in Blacksburg, put their names or ID in a drawer, and randomly pick a number of people (nowadays people use random number generator).
Probability Sampling _ Simple Random Sampling
patient satisfaction with service at a clinic
Probability Sampling _ Systematic Sampling
Given the previous knowledge, the effectiveness of flu vaccine differs in terms of age. Divide the whole population into different age groups (e.g. 0-17, 18-60, 60+), and take simple random samples from each group.
Probability Sampling _ Stratified Sampling
phone interview of lung cluster patients
Probability Sampling _ Cluster Sampling
within stratum
between strata
to increase precision and representation
Probability Sampling _ Stratified Sampling
between clusters
within cluster
to reduce cost and improve efficiency
Probability Sampling _ Cluster Sampling
A sample that is selected in a convenient way that introduces a high risk of biasedness.
Non-Probability Sampling _ Convenience Sampling
Collect info from all VCOM GCP students.
Non-Probability Sampling _ Convenience Sampling
what type of sampling is the following?
Steps:
Break the whole population into strata based on certain characteristics: e.g. age, gender.
Determine the quota for each stratum.
Sampling from each stratum until the quota is reached.
Notes:
Number of subjects (quota) for each stratum is fixed.
Subjects from different strata are as heterogeneous as possible.
It differs from stratified sampling in the way samples are drawn within each stratum.
It is quicker and easier to carry out.
Non-Probability Sampling _ Quota Sampling
what is the purpose of research questions?
What questions will the study address?
what is the purpose of background and significance?
Why are these questions important?
what is the purpose of design, like:
time frame
research design
How is the study structured?
what is the purpose of subjects, like:
selection criteria
sampling design
Who are the subjects and how will they be selected?
what is the purpose of variables:
Predictor variables
Outcome variables
Confounding variables
What measurements will be made?
what is the purpose of Statistical Issues, like:
Hypotheses
Sample size
Analytic approach
How large is the study and how will it be analyzed?
what is the overall research goal and general question?
concrete and researchable questions
Example:
General concern: Should people eat more fish?
Specific questions:
How often do Americans eat fish?
Does eating more fish lower the risk of cardiovascular disease?
Is there a risk of mercury toxicity from increasing fish intake in older adults?
Do fish oil supplements have the same effects on cardiovascular disease as dietary fish?
Which fish oil supplements don’t make you breath smell like fish?
a good research question should pass the what test?
so what test, getting the answer should contribute to our state of knowledge
what are the five essential characteristics of good research?
Feasible Interesting Novel Ethical Relevant
Describes some interesting or intriguing observations that occurred for a small number of patients
Case series study
A group examined at one point in time (e.g., survey)
Cross-sectional study
Two groups selected based on presence or absence of an outcome
Case-control study
Groups of subjects identified at the beginning and followed over time (retrospective cohort study)
Cohort study (Retrospective and Prospective)
Two groups (control and experimental) created by a random assignment
Randomized controlled trial (RCT)
explore the lay of the land and lead to generation of research hypotheses.
Descriptive observational studies
test one or more predetermined hypotheses about association between exposure and outcome variables.
Analytical observational studies
test one or more predetermined hypotheses about association between exposure and outcome variables.
Experimental studies
The investigator goes through the medical records of several patients who has been taking fish oil supplements for the past year, and notices their current coronary calcium scores are lower than those from one year ago.
Case Series Study
The researcher interviews a group of subjects about current and past history of fish intake and correlates results with history of coronary heart disease (CHD) and current coronary calcium score.
Cross-Sectional Study
The researcher examines a group of patients with coronary heart disease (CHD) (the “cases”) and compares them with a group who do not have CHD (the “controls”), asking about past fish intake.
Case-Control Study
The investigator measures fish intake in a group of subjects at baseline and periodically examines them at follow-up visits to see if those who eat more fish have fewer coronary heart disease (CHD) events.
Cohort Study
The investigator randomly assigns subjects to receive a fish oil supplement or a placebo that is identical in appearance, then follows both groups for several years to observe the incidence of CHD.
Randomized Controlled Trial (RCT)