14. Stats Flashcards
What type of data is categorical?
Qualitative
What type of data is numerical?
Quantitative
What are the 2 types of categorical data?
- Nominal
2. Ordinal
Data that can be in categories, but have no particular order or magnitude differences?
Nominal
Data that can be allocated to an ordered set of categories?
Ordinal
Discrete data that can only be certain whole numbers and continuous data that can be any numerical value?
Numerical
What type of data is blood groups?
Nominal
What type of data is AHA class?
Ordinal
What type of data is # of surgical procedures?
Discrete
What type of data is cardiac index?
Continuous
Case-control advantages?
- Can study rare disease
- Can study disease with long latency between exposure/manifestation
- Can be launched/conducted over short time periods
- Inexpensive (compared to cohort)
- Can study multiple causes of disease
Case-control disadvantages?
- Recall bias (information on exposure/past history based on interview)
- Validation of info on exposure is difficult
- Concerned with one disease only
- Can’t provide information on incidence rates of disease
- Incomplete control of extraneous variables
- Choice of appropriate control group can be challenging
- Methodology can be hard to comprehend for non-epidemiologist
- Correct interpretation of results can be hard
Cohort advantages?
- Complete information on subjects exposure (quality control of data)
- Clear temporal sequence of exposure/disease
- Study multiple outcomes related to a specific exposure
- Calculation of incidence rates (absolute risk and relative risk)
- Methodology/results easily understood by non-epidemiologists
- Study relatively rare exposures
Cohort disadvantages?
- Not suited for rare disease (need large # subjects)
- Not suited when time between exposure and disease manifestation is very long (can be overcome in historical cohort studies)
- Exposure patterns may change during course of study and make results irrelevant
- Maintaining high rates of follow-up can be difficult
- Expensive to carry out (need large # subjects)
- Baseline data sparse… large # of subjects doesn’t allow for long interviews
Research involving administration of a test regimen to humans to evaluate both efficacy and safety
Clinical trial
Phases of a clinical trial?
1- Safety and pharmacologic profiles
2- Pilot efficacy studies
3- Extensive clinical trial
4- Studies after FDA approval for distribution
Administration of a single subtherapeutic dose of the drug to a small group (0-15) to gather preliminary data on pharmacokinetics and pharmacodynamics
Phase 0
A small group (20-80) of volunteers to assess the safety and pharmacokinetic profile of medication
Phase 1
A large group (20-300) to assess safety in a larger group of patients as well as effectiveness of the drug
Phase 2
Randomized controlled multicenter trial on a relatively large group (300-3000+) depending on the medical condition and is to assess the effectiveness of the drug in comparison with an accepted therapy
Phase 3
Safety surveillance and ongoing technical support of a drug after permission for it to be distributed
Phase 4
What type of test is used when numbers in contingency table of categorical variables are relatively small?
Fisher exact
What test is used for two groups with paired data?
McNemar
What test is used to measure the difference between actual/expected frequencies of categorical variables?
Chi2
What test is an extension of the Chi2 test used when comparing several 2-way tables (meta-analysis)
Mantel-Haenszel
What tests are used to compare samples of normally distributed data?
Parametric
What type of tests are used when data are not normally distributed?
Non-parametric
What types of tests are student T-test, ANOVA, ANCOVA, Kolmogorov-Smirnov
Parametric
What types of tests are Wilcoxon signed-rank, Mann-Whitney U-test, Wilcoxon rank sun, Kruskal-Wallis?
Non-parametric
What test is used to compare 2 samples to test probability that samples come from population with same mean value?
Student T-test
What test is used to compare the means of 2+ samples to see whether they are derived from the same population
ANOVA
What test is used to compare the means of 2+ samples to see whether they are derived from the same population and accommodates continuous variables?
ANCOVA
What test is used to test hypothesis that the collected data are from a normal distribution so that the parametric stats can be used?
Kolmogorov-Smirnov
What test compares the difference between paired groups?
Wilcoxon signed rank
What non-parametric test is like the t-test for parametric data?
Wilcoxon signed-rank
What tests compare 2 sets of data that are derived from 2 different sets of subjects?
Mann-Whitney U-test or Wilcoxon rank sum
What test compares 2+ independent groups (like the ANOVA for parametric)?
Kruskal-Wallis
What describes the frequency of occurrence of new cases during a time period?
Incidence
What measure is useful to explore causal theories or evaluate effects of preventive measures?
Incidence
What is the equation for incidence?
new cases in a population in a period of time/Sum for each individual in population of length of time at risk for getting disease
What is the equation for cumulative incidence?
of individuals who get disease in certain period/Number of individuals in population at beginning of period
What describes what proportion of the population has a disease at a specific point in time?
Prevalence
What does prevalence depend on?
Incidence and duration
P = I * D
What measure is relevant to planning of health services or assessing need for medical care in a population
Prevalence
What is the equation for prevalence?
Existing # of individuals having disease at a specific time/Number of individuals in the population at that point in time
Does chronic disease have lower prevalence or incidence?
Lower incidence
Do acute illnesses have lower prevalence or incidence?
Lower prevalence
What is used to delineate how one set of data relates to another though a best fit line?
Regression analysis
*Regression coefficient is the slope of a line
Name examples of different types of regression analysis?
- Simple linear
- Logistic
- Poisson
- Cox proportional hazards
What is the most common survival curve method?
Kaplan-Meier curve
What does a Kaplan-Meier curve do?
Displays survival of a cohort with calculation of survival estimates upon each death or event
What is a nonparametric test to compare the survival between 2 potential Kaplan-Meier curves?
Log rank test
What is a well-recognized curve that reflects a continuous probability distribution that is bell-shaped (unimodal) and symmetrical about the mean with 2 parameters, mean and variance?
Gaussian distribution
Which continuous probability distribution most closely resembles the normal of Gaussian distribution?
T-distribution
What is the measure of dispersion or variability in a sample?
Standard deviation
What % of cases fall within 1 SD in normal distribution?
68.2%
What % of cases fall within 2 SD in normal distribution?
95.4%
What % of cases fall within 3 SD in normal distribution?
99.7%
True or False: Mean and median of a normal distribution are equal
True
What is the difference between T-distribution and normal distribution?
T-distribution is more spread out with longer tails
What distribution is right skewed and characterized by degrees of freedom?
Chi2
What distribution is right skewed used for comparing 2 variances?
F distribution
What distribution is highly skewed to the right (it is the probability distribution of a random variable whose log follows the normal distribution)?
Log normal distribution
Name 2 discrete probability distributions
- Binomial
2. Poisson
What is a confidence interval?
Range that is likely to contain the true population mean valve
What does a 95% confidence interval mean?
There is a 95% chance that the population value lies within stated limits
What indicates variability in a sample?
Standard deviation
In a normal distribution, 95% of the distribution of the sample means is within what SD of the population mean?
1.96
The size of the CI is related to what?
Sample size of study
*Larger the population, narrower the CI
How is the 95% confidence interval for the mean calculated?
Sample mean – 1.96 x SEM to sample mean + 1.96 x SEM
*SD is the SEM
Name the 2 types of applied statistics
- Descriptive
2. Inferential
What do descriptive statistics do?
Describe data in a sample
What do inferential statistics do?
Estimate whether results suggest a real difference between populations
Examples of descriptive statistics?
- Mean
- Median
- Mode
- SD
- Quartiles
- Histograms
Examples of inferential statistics?
- Student T-test
- ANOVA
- Chi2
What is a type I or alpha error?
When null hypothesis that is correct is rejected (stating a difference when there isn’t one)
What is the chance of making a type I error?
P-value
What is a type II or beta error?
When null hypothesis that is incorrect is accepted (stating no difference when there is one)
What is a type III error?
Study design that produces the right answer to the wrong question
What is the p-value?
Probability that defines how likely it is that a hypothesis is true (usually null hypothesis- no difference between groups)
What is the probability of an observed difference occurring solely by chance?
P-value
What is the usual p-value level of significance?
0.01 to 0.05
What is the method used to adjust P-value for multiple testing?
Bonferranoi adjustment
What is the power of a study?
Probability that it would detect a statistically significant difference
What is B in statistics?
Probability of accepting a hypothesis that is false
What is the equation for the power of a study?
1-B
*Probability of rejecting the null hypothesis when it is false
What is the minimum Power a study should have?
80%
What things can increase the power of a study?
- Larger significance level
- Larger effects
- Decreased variability of the observations
- Larger sample size
What is this assessment tool…Economic assessment method utilized in which costs and consequences of alternative cardiac interventions are expressed in costs per unit of health outcome. This is applicable to health programs as well as health services to determine preferred action that requires the least cost to produce a given level of effectiveness.
CEA: Cost effective analysis
What is CUA?
- Uses quality-of life measurements expressed as utilities (QALY) in the value equation.
- Disability-adjusted life year (DALY) is also a measure but is for the overall “burden of disease”
- Quantifies the impact of premature death (like QALY), but also disability on a population by combining them into a single, comparable metric
What is CBA?
Seeks to translate all relevant healthcare considerations into monetary terms by analyzing economic and social costs of medical care and benefits of reduced loss of net earnings due to preventing premature death or disability
What is a technique where results from a number of studies that are similar in nature are gathered to give one overall estimate of the effect?
Meta-analysis
List the formal steps for a meta-analysis
- Decide on effect of interest
- Check for statistical homogeneity
- Estimate average effect of interest with Cis
- Interpret the results and present the findings (forest plot)
List advantages of meta-analysis
- Refinement and reduction
- Efficiency
- Generalizability and consistency
- Reliability
- Power/precision
List disadvantages of meta-analysis
- Publication bias
- Clinical heterogeneity
- Quality differences
- Lack of independence of study subjects
What is a systematic review?
- Uses meta-analysis to render well-informed clinical decisions… essential part of evidence based medicine
- Major disease categories often have a sufficient number of randomized clinical trials for the at minimum a meta-analysis to determine the value of such an intervention
When is the risk ratio or relative risk used?
Prospective cohort studies
How is RR calculated?
Divide risk in treated/exposed group by risk in control/unexposed group
How is RR reported?
Given with a 95% CI
- Can be <1, 1 or >1
- If the CI includes 1, not statistically significant
What is RR similar to?
Odds ratio
What is the relative risk reduction?
Proportion by which the intervention reduces the event rate
*Control group risk-Intervention group risk/Control group risk
What is the absolute risk reduction?
Difference between the event rates in the intervention versus control groups
*Control group risk-Intervention group risk
What is the number needed to treat?
Number of patients who need to be treated for one to get benefit
Relationship of NNT and ARR?
NNT is reciprocal of ARR
ARR = 100/NNT
When is odds ratio used?
Retrospective case-control studies
How is the odds ratio calculated?
By comparing odds of the exposed versus control groups
*Calculated by dividing the event occurrence by the number of times that the event doesn’t happen
How is the odds ratio reported?
Given with a 95% CI
- Odds ratio can be <1, 2 >1
- If it includes 1, it isn’t statically significant
In a typical receiver operating characteristic (ROC) curve, what is the significance of the upper left corner or coordinate (0,1)?
100% sensitivity and specificity
- Percent classification
- No false negatives and no false positives
What is a ROC curve
A 2-way plot of the sensitivity (true +) against 1 minus the specificity (false + rate) for different cutoff valves for a continuous variable in a diagnostic test
What shape do you want an ROC curve to have?
Sharp upslope then taper off (versus just a straight diagonal line)
What is the measure of precision of the sample mean or how close the sample mean is likely to be to the population mean?
Standard error of the mean (SEM)
What is variance?
Square of the standard deviation
What is the coefficient of variation?
Ratio of the SD to the mean
What is a measure of spread away from the mean?
Standard deviation
What is the square root of variance?
SD
What is a measure of precision of the sample mean or how close the sample mean is likely to be to the population mean?
SEM
What is the degree of closeness of measurements to quantity’s true value?
Accuracy
What is the reproducibility of a study result with the study to be repeated under the same circumstances?
Precision
How is precision measured?
Standard error of measurement
What is a Chi squared test?
Measure of the difference actual and expected frequencies with categorical variables
What needs to be set up to calculate a Chi2 value?
Contingency Table
If there is no difference between the actual and expected values, what is the Chi2 value?
0
*Larger the difference, bigger the X2 value (and p-value accompanies X2 value)
What is the number of independent comparisons that can be made between members of a sample and is used with X2 to calculate the p-value?
Degree of freedom
In the example of some kids with SVT being treated with digoxin v. propranolol, what degrees of freedom is needed to calculate a p-value
1
*Number of independent comparisons that can be made between members of sample
What is sometime used with a Chi2 test to improve the accuracy of the p-value?
Yates continuity correction
When is a Fisher exact test used?
When numbers in a contingency table of categorical variables are small
When is a McNemar test used?
For 2 groups with paired data
What is an extension of the Chi2 test that is used when comparing several 2-way tables (like a meta-analysis)?
Mantel Haenszel test
What is a correlation coefficient?
The strength of the linear relationship between 2 variables
What is the range of a correlation coefficient?
Denoted by r and ranges from -1 to +1
What is sometimes used with correlation coefficient to correct for negatively corrected relationships?
R2
When can a correlation coefficient not be calculated?
Non-linear relationship
Outliers
What is a multiple correlation coefficient?
When degree of linear relationship is extended to several variables
When is Pearson correlation coefficient used instead of Spearman correlation coefficient?
- Pearson (“r”): Values are sampled from normally distributed populations
- Spearman (“rs”): Values are sampled from non-normally distributed populations
Equation for PPV?
True positives/Total positives (A/A+B)
Probability that a diseased individual is correctly classified as sick?
Sensitivity
Equation for sensitivity?
#Sick classified as sick/Total # sick A/(A+C)
What does sensitivity assess?
How often the test is + in patients who have the disease
What is the probability that healthy individuals are classified as healthy?
Specificity
Equation for specificity?
# Healthy classified as healthy/Total # healthy D/(D+B)
What does specificity assess?
How often a patient tests negative if they are healthy
True or false: There is interdependence between sensitivity and specficity?
True
What is the likelihood that a patient has the disease if they test positive?
PPV
Equation for PPV?
A/(A+B)
What is the likelihood that a patient is healthy if they test negative?
NPV
Equation for NPV?
D/(D+C)
True or False: The higher the sensitivity/specificity/PPV/NPV, the more valuable the test
True
*Perfect test would be calculated at a 1
What is a likelihood ratio?
Likelihood that a test result would be expected in a patient with the condition compared to the likelihood that the same result would be in a patient without the condition
Equation for likelihood ratio?
Sensitivity/(1-Specificity)
A/A+C)/(1-(D/D+B)
What does a likelihood ratio imply?
That if the test is + in a patient, the patient is many more times likely to have the disease than not
What is a randomized controlled trial?
Patients randomized to receive either new or control treamtent
What is involved in ideal randomization for a RCT?
- Equal group sizes
- Low selection bias
- Low probability of confounding (accidental bias)
What are refinements of simple randomization (as related to RCT)?
- Stratified randomization
- Blocked randomization
- Cluster randomization
What does stratified randomization control for?
Effects of factors
What does blocked randomization assure?
Treatment groups to be equal sized
What does cluster randomization allocate for?
Groups of patients
What is response-adaptive randomization (or outcome-adaptive randomization)?
Probability of being assigned to a group increases if responses of prior patients is deemed favorable
What is a placebo-controlled study?
Involves a control group that doesn’t receive the treatment
What occurs when there is a systematic difference between the results of a study and the true result?
Bias
What is bias that occurs when a spurious association is noted due to a failure to adjust fully for factors leading to an erroneous conclusion?
Confounding bias
What is observer bias?
Observer inaccurately assess variable
What is confounding bias?
Spurious association
What is selection bias?
Selected study subjects are not representative
What is information bias?
Measurements are incorrectly recorded
What is publication bias?
Only positive results are published
List types of bias
Observer, confounding, selection, information, publication, recall, assessment, allocation
What is an association?
Any relationship between 2 measured quantities that relates them to be statistically dependent
What defines a linear relationship between 2 quantities?
Correlation
What factors allow you to include causation in addition to association?
Temporality Strength of causality Dose-response Repetition in a different population Consistency with other studies Biologic plausibility
What does the Belmont report include?
3 principles of research ethics
What are the 3 principles of research ethics included in the Belmont report?
- Respect for persons
- Beneficence
- Justice
What is described by protecting the autonomy of all people and treating them with courtesy and respect and allowing for informed consent?
Respect for person
A researcher being truthful about the possibility for negative side effects associated with a study drug is abiding by what?
Respect for persons
What is described by “do no harm” while maximizing benefits for research project and minimizing risks for research subjects?
Beneficence
What is described by ensuring reasonable, non-exploitative and well-considered procedures are administered fairly and equally?
Justice
Who is the independent group of experts that continuously monitor data from various aspects of a clinical trial to ensure patient safety as well as validity and scientific merit?
Data Safety Monitoring Board
What is a committee designed to approve and review research involving human subjects to protect the rights and welfare of human research subjects?
IRB (ethical review board)
What is the difference between the IRB and DSMB?
- IRB primarily responsible for review of clinical protocols and related documents
- DSMB main responsibility is to review the trial safety and efficacy data
What is the consistency of a set of measurements or a measurement tool or its repeatability and reproducibility?
Reliability
What is the extent to which a study measures what it is intended to measure?
Validity
Validity is a measure of what?
Systematic error or bias (confounding and selection bias)
What is inversely related to random error?
Reliability
What is the degree of closeness of measurements to the quantity’s true value?
Accuracy
What is the reproducibility of a study result with the study to be repeated under the same circumstances?
Precision
What is precision measured by?
Standard error of measurement
Sample size calculation involves what?
- Power (0.8)
- Significance level (0.01 or 0.05)
- Variability of the observations (SD)
- Smallest effect of interest (the standardized difference)
List advantages of a case-control study
- Permit study of rare diseases
- Permit study of diseases with long latency between exposure and manifestation
- Can be launched and conducted over a relatively short time period
- Relatively inexpensive (compared to cohort)
- Can study multiple potential causes of disease
List disadvantages of a case-control study
- Information on exposure/past history primarily based on interview and may be subject to recall bias
- Validation of information on exposure is difficult, incomplete or impossible
- Concerned with one disease only
- Cannot usually provide information on incidence rates of disease
- Generally incomplete control of extraneous variables
- Choice of appropriate control group may be difficulty
- Methodology may be hard to comprehend for non-epidemiologists
- Correct interpretation of results may be difficult
List advantages of a cohort study
- Allow complete information on subject’s exposure, including quality control of data and experience thereafter
- Provides a clear temporal sequence of exposure and disease
- Gives an opportunity to study multiple outcomes related to a specific exposure
- Permits calculation of incidence rates (absolute risk) as well as relative risk
- Methodology and results are easily understood by non-epidemiologists
- Enable the study of relatively rare exposures
List disadvantages of a cohort study
- Not suited for the study of rare diseases (because large number of subjects are required)
- Not suited when the time between exposure and disease manifestation is very long (can be overcome in historical cohort studies)
- Exposure patterns may change during course of study and make results irrelevant
- Maintaining high rates of follow-up can be difficult
- Expensive to carry out (because a large number of subjects are usually required)
- Baseline data may be sparse (because large number of subjects doesn’t allow for long interviews)
What statistical method allows for paired comparisons of two non-normal patient populations?
Wilcoxon signed-rank test
What types of tests are used when data isn’t normally distributed?
Non-parametric
What test is used for nonparametric data when comparing difference between paired groups?
Wilcoxon signed-rank test
What test is used for parametric data when comparing difference between paired groups?
T-test
What 2 tests compare 2 set of data that are derived from 2 different sets of subjects?
Mann-Whitney U-test
Wilcoxon ranks sum test
What test compares 2+ independent groups with nonparametic data?
Kruskal-Wallis
What test compares 2+ independent groups with parametric data?
ANOVA
Numerical data, single group tests?
1-sample T-test
Sign test
Numerical data, 2 paired groups test?
Paired t-test
Wilcoxon signed-rank test
Numerical data, 2 unpaired groups test?
Unpaired t-test
Wilcoxon rank sum test (Mann-Whitney U-test)
Numerical data, multiple (>2) groups test?
ANOVA (1-way)
Kruskal-Wallis test
Categorical data, single group test?
Test of single proportion
Sign test
Categorical data, 2 paired groups test?
McNemar test
Categorical data, 2 unpaired groups test?
Chi2 test
Fisher exact test (<5)
Categorical data, multiple (>2 groups) test?
Chi2 test
What are 2 types of categorical data?
Nominal
Ordinal
What is nominal data?
Data that describe data that can be in categories, but have no order or magnitude different
Ex. SV or BV surgical strategies or antiarrhythmic agent for SVT
*Type of categorical data
What is ordinal data?
Data that can be allocated to an ordered set of categories
Ex. Severity of AVVR from mild to severe
*Type of categorical data
What are the 2 types of numerical data?
Discrete
Continuous
What is discrete data?
Can only be certain whole numbers
Ex. Number of reinterventions after Norwood
*Type of numerical data
What is continuous data?
Can be any numerical value
Ex. BP before and after ACEi
*Type of numerical data
What is a retrospective study that studies the relationship between risk factor and outcome and uses relevant exposure or condition information from a sample of individuals with the disease or condition (cases) rather than examining the entire population?
Case-control
What is a qualitative study of a single patient or small group of patients with a similar disease?
Case-series
What is a prospective observational study with study subjects (cohort) assigned to an exposure or condition category and then all followed for a defined observation period to see whether they develop disease?
Cohort
*Follow-up, longitudinal, prospective, historical
What is a historical cohort study?
A cohort study using a group of patients from the past- wouldn’t involve active enrollment of new study subjects
A Chi2 test is most closely related to what statistical test?
Fisher exact
When is a Fisher’s exact test used?
Used when numbers in the contingency table to categorical variables are relatively small
*Chi2 used with large (>5) populations
What is a chi2 test a measure of?
The difference between actual and expected frequencies with categorical variables with large (>5) populations
What type of data is a chi2 and Fisher exact test used for?
Categorical
ANOVA, student t-test, Kolmogorov-Smirnov and Wilcoxon signed-rank tests are used for what type of data?
Numerical
Parametric tests are used to compare what type of data?
Normally (Gaussian) distributed data
Student T-test, ANOVA and Kolmogorov-Smirnov tests are what type of tests?
Parametric
What test is used to compare 2 samples to test the probability that the samples come from a population with the same mean value?
Student t-test
What test is used to compare the means of 2+ samples to see whether they are derived from the same population?
ANOVA
What test is used to test the hypothesis that the collected data are from a normal distribution, so that the parametric statistics can be used?
Kolmogorov-Smirnov
What type of test is used when data isn’t normally distributed?
Non-parametric
What type of test is a Wilcoxon signed rank?
Non-parametric
What test is used for comparing the difference between paired groups (similar to a T-test for paired data) in non-parametric data?
Wilcoxon signed rank
What is the definition of prevalence?
Existing # of individuals having the disease at a specific time/number of individuals in the population at that point in time
What describes frequency of occurrence of new cases during a time period?
Incidence
What is the proportion of the population that has the disease at a specific point in time?
Prevalence
What 2 things does prevalence depend on?
Incidence and duration of the disease
P = I * D
What is an economic assessment methodology that seeks to translate all relevant healthcare considerations into monetary terms by analyzing economic and social costs of medical care and benefits of reduced loss of net earnings due to preventing premature death or disability?
CBA
What is an economic assessment method in which the costs and consequences of alternative interventions are expressed in costs per unit of health outcome?
CEA
What uses quality of life measurements expressed as utilities (QALY) in the value equation?
CUA
What are the designation levels described by the US Preventive Services Task Force used in a review article for medical therapies?
A: Good scientific evidence, benefits substantially outweigh risk
B: Fair scientific evidence, benefits outweigh risk
C: At least fair scientific evidence, benefits and risk too close
D: At least fair scientific evidence that risks outweigh the benefit
I: Scientific evidence is lacking, poor quality or conflicting
Equation for sensitivity?
A/A+C
Equation for specificity?
D/B+D
Equation for PPV?
A/A+B
Equation for NPV?
D/C+D
What is a likelihood ratio?
Likelihood that a test result would be expected in a patient with the condition compared to the likelihood that the same result would be in a patient without the condition
Equation for likelihood ratio?
Sensitivity/(1-Specificity)
How often the test is positive if the patient has the disease?
Sensitivity
How often the test is negative if the patient is healthy?
Specificity
Likelihood that the patient has the disease if the test is positive
PPV
Likelihood that the patient is healthy if the test is negative?
NPV
What is a perfect likelihood ratio?
1
*Higher the calculated value, the more valuable the test
When is the risk ratio (or relative risk) used?
Prospective cohort studies
How is risk ratio or relative risk calculated?
Dividing the risk in the treated or exposure group by the risk in the control or unexposed group
How is risk ratio reported?
<1, 1, or >1
Given with a 95% CI… if the CI includes 1, it isn’t statistically significant
What is the proportion by which the intervention reduces the event rate
Relative risk reduction (RRR)
What is the difference between the event rates in the intervention v. control groups?
Absolute risk reduction (ARR)
What is the number of patients who need to be treated for one to get benefit?
Number needed to treat (NNT)
What is the equation for NNT?
100/ARR
*Reciprocal of ARR
The most reliable results occur with that type of study?
Double-blind, placebo-controlled trial
What is the degree a study produces consistent results?
Reliability
True or False: You improve reliability to minimize bias in a study?
True
What is the practice of randomly assigning enrolled patients in one of treatment or control groups?
Randomization
What is study design in a way that providers administering intervention, measuring outcomes and patients receiving therapy are unaware who is in what group
Double Blinding
What kind of bias can randomization and double blinding minimize?
Susceptibility
What kind of bias occurs when differences in subjects at baseline between the compared groups cause differences in outcomes beyond what difference in interventions would otherwise cause?
Susceptibility
What is important in study design to assure changes wouldn’t be seen in study groups regardless of intervention?
Placebo control group
What type of study involves reviewing RFs for patients who have the disease of interest and comparable control patients who don’t?
Case-control studies
What are case control studies used for?
Determine likelihood that various RFs are more or less associated with the cases versus controls
What type of study entails prospectively following patients with a given exposure and those without?
Cohort
What are 2 advantages of case-control studies?
Study multiple RFs and rare conditions
What study design involves collection and analysis of data collected from a population at one specific point in time?
Cross-sectional
A representative cohort of families surveyed to help determine prevalence of chest pain in a pediatric population is an example of what?
Cross-sectional study
What study design is a review of records from a cohort of patients?
Retrospective cohort
What study design follows a group of patients forward in time to determine which develop disease?
Prospective cohort
What study design is useful to determine the prevalence of a disease?
Cross-sectional
The more participants in a study, the higher the what?
Sensitivity for detecting adverse events
What is a type II error?
Inability to reject null hypothesis when a difference between study groups truly exist
What represents a false-negative finding?
Type II error
How can you decrease the risk of a Type II error?
Increase the sample size studied to increase power and ability to find a difference if one truly exists
What refers to rejection of the null hypothesis when a true difference doesn’t exist?
Type I error
What represents a false-positive finding?
Type I error
How can you impact a Type I error in a study?
By adjusting the significance rate
What is a significance rate typically set at in a study?
- 05
* Statistical analysis of results must show <5% chance that results are related to chance versus true difference to be termed “significant”
What is a type of study design where patients with exposure to intervention of interest and those without are followed forward in time for development of measured outcome?
Prospective cohort
*Not retrospective because patients followed forward in time rather than record review
What is healthy entrant effect?
Lower morbidity/mortality in patients entering a study than general population due to study design
What is the ability of a test to correctly identify those without disease?
Specificity
What is the ability of a test to predict patients without disease?
NPV
What is a tests ability to demonstrate accurate value?
Validity
What is a tests ability to get consistent results?
Reliability
What does a 95% CI represent?
Range of values that are 95% certain to contain the true mean for the population based on data from respective cohort
- DOESN’T represent values between which 95% of the sample or population values fall
- Calculated around the mean value for each group
What is recall bias?
Inaccurate recollection of events by study participants
*Systematic error
What is lead-time bias?
Disease recognized earlier and survival hasn’t really changed versus test improving survival time
What is selection bias?
Nonrandom collection of participants
What is referral bias?
Only subset of population included
Ex: Study at tertiary care center with only sickest patients and not representative of population of interest
What statistically describes an association between an exposure and risk of outcome of interest?
Odds ratio
What does a negative odds ratio indicate?
Decreased risk
What does a positive odds ratio indicate?
Increased risk
If the 95% CI for an odds ratio includes 1, what does that mean?
Likely not statistically significant
What type of statistical analysis is most appropriate for assessment of a continuous variable both pre/post intervention in the same patient?
Paired student t-test
- Continuous variable with N>25 = Parametric test/Student t-test
- Paired because same patient studied before/after intervention, so 2 matched cohorts
What type of test is used to compare cohort means in samples that aren’t normally distributed or have a low number of participants?
Wilcoxon signed-rank
*Nonparametric
What type of test is used to compare categorical outcomes versus continuous variables?
Chi2
What demonstrates the odds of developing a given outcome in patients with a particular exposure and those without?
Odds ratio
What is a method to display survival results grafically?
Kaplan-Meier curve
What describes the odds of the outcome of interest in those patients with the exposure?
Odds ratio
What is the equation for an odds ratio?
(AD)/(BC)
True or False: Can calculate an odds ratio from multiple different study designs including case-control studies
True
True or False: Chi2 analysis is used to statistically evaluate continuous data?
False- Evaluates nominal data, not appropriate for continuous
What assumptions are needed to use a Chi2 analysis?
- Random/completely independent study groups
- All cells of table must have expected value >5
- Data must be arranged in table form (nominal)
What does a statistically significant P-value of <0.05 represent?
<5% chance that data distribution in the study could have occurred by random chance
What is the most appropriate test for comparison of survival between 2 non-normally distributed groups, each with 10 participants?
Fischer exact test
What does normal distribution of data in a study refer to?
Random data demonstrates a bell curve when graphed
*Data may be skewed or non-normally distributed especially if a small # of participants
What is a non-parametric test used to statistically analyze association between 2 groups which aren’t normally distributed?
Fischer exact test
What test must have normally distributed data with at least 5 participants in each cell of the table?
Chi2
*May provide falsely low p-value if used incorrectly
A Wilcoxon signed-rank test and t-test are used for what type of data analysis?
Continuous
What are censored patients in a survival curve (Kaplan-Meier)?
Drop out of the study for reasons other than event of interest (death)
What are non-censored patients in a survival curve (Kaplan-Meier)?
Had event of interest (death)
NNT cannot be calculated with that type of data?
Continuous
*Data has to be categorical and binary to calculate
Equation for NNT?
1/ARR
What type of test demonstrates agreement between 2 groups when there isn’t a gold standard?
Kappa statistic
What are the values of a Kappa statistic?
-1 (negative association)
to
+1 (positive association)
0 demonstrates no association
When can sensitivity calculations and ROC be used to evaluate a test?
When there is a clear gold standard
What does a regression analysis demonstrate?
Relationship between a dependent variable and 1+ independent variables
What is a nonparamateric test used to analyze categorical data with sample sizes too small to allow for use of Chi2?
Fischer exact test
A Chi2 test requires how many patients in each cell?
> 5
A student T-test uses what type of data?
Continuous
What is used to compare mean values from 3+ groups?
ANOVA
What test is the non-parametric equivalent of an ANOVA?
Kruskall-Wallis
What is a parametric test used to compare group means with 3+ independent groups?
ANOVA
What does it mean if the p-value for an ANOVA is <0.05?
There is a difference in the group means among multiple study groups, but not specifically where difference lies or how great a difference there is
-Need additional analysis comparing each group to other to find where exact difference is
What test is used to compare mean values among multiple groups when ANOVA isn’t appropriate (data aren’t normally distributed)?
Kruskal-Wallis
What type of data does Chi2 evaluate?
Categorical
What does it mean is a P-value is <0.05?
Null hypothesis disproven and there is a difference between study groups
What is the null hypothesis when comparing multiple independent variables to one dependent outcome?
None of the independent variables is associated with the outcome of interest
What is a type I error?
Null hypothesis rejected, but really true (False+)
What is a type II error?
Null hypothesis isn’t rejection, but really is difference (False-)