Stats Flashcards
Which type of test has few false negatives, or negative test results are reliable?
Sensitive
Which type of test could have lots of false positives, making positive results not as reliable?
Sensitive
What type of test is a good screening test?
Sensitive
*Negative rules out disease, but positive doesn’t necessarily rule in disease
What test has reliable negative and no false positives?
Sensitive
What type of test has positive results that are very reliable, but false negatives can occur?
Specific
What type of test can be used to determine treatment since positive results are reliable?
Specific
What is a common example of a specific test used in peds?
Rapid strep- Positive you can trust and treat, negative you send for culture
What type of test has reliable positives with rare false positives?
Specific
What is the sensitivity of a test?
The probability that the test will produce a true positive result when used on the population with the disease
What is the equation for sensitivity?
TP/(TP + FN)
True or False: In a test with high sensitivity, positive results may not be reliable
True
Negative results in what type of test can rule our disease?
Sensitive (very few false negatives)
What type of test is a good screening test?
Sensitive (if negative, probably true negative and no further workup needed)
Negative results are reliable in what type of test?
Sensitive
What is the equation for specificity?
TN/(TN + FP)
What is the specificity of a test?
Probability that is will produce a true negative result when used on a population without disease
What type of test has very few false positives?
Specific
In what type of test are positive results reliable?
Specific
What type of test will have very few false negatives?
Specific
A specific test is good for what?
To confirm the presence of a disease
In assessment for stroke, what type of test is a head CT?
Specific: If positive, helps to diagnose acute stroke, if negative still need further workup like MRI, ect. (may be false negative)
What test is good for confirming positive disease?
High specificity
A test with good positive and negative predictive values is what (in terms of sensitivity and specificity)?
Sensitive enough to minimize false negatives
Specific enough to minimize false positives
What is the positive predictive value of a test?
Probability that someone who tests positive actually has the disease and doesn’t represent a false positive.
What is the equation for positive predictive value?
TP/(TP + FP)
When is positive predictive value most useful?
When prevalence is high
What happens to positive predictive value as prevalence increases?
Positive predictive value increases
What happens to positive predictive value as prevalence falls?
Positive predictive value falls
True or False: The higher the positive predictive value, the higher the reliability of a positive result
True
What is negative predictive value?
The probability that someone who tests negative is actually disease free and doesn’t represent a false negative
When is negative predictive value most useful?
When prevalence if low
What happens to negative predictive value as prevalence increases?
Negative predictive value falls
What happens to negative predictive value as prevalence falls?
Negative predictive value increases
True or False: A high negative predictive value increases the reliability of negative result
True
If you are testing a group where a specific disease is prevalent, will the negative predictive value for that disease be lower or higher?
Lower
What is the equation for negative predictive value?
TN/(TN + FN)
What is the null hypothesis?
That results obtained were due to chance factors and not the variables being studied
What is generally a synonym to “rejecting the null hypothesis”?
“Statistically significant”
What is generally the goal of research in terms of a null hypothesis?
To reject the null hypothesis
What is the P value?
The chance that the null hypothesis was rejected in error
What represents the odds that the null hypothesis was correct and that the results were due to chance and not attributed to the studied variables?
P
What does a p value <0.01 mean?
Less than 1% chance that the null hypothesis was rejected in error and results are due to chance alone
What P value is significant enough to reject the null hypothesis?
P < 0.05
True or False: The lower the P value the more significant the study?
True
What is a type 1 error?
Rejecting the null hypothesis in error- thinking something was significant/different when it really wasn’t
What is the probability of a type 1 error?
P-value
What is it called if you labeled an insignificant study as significant?
Type 1 error (study was “overestimated”)
What is a type 2 error?
Where you rejected the null hypothesis in error
What is it called if you determined that the findings could occur by chance and accepted the null hypothesis when in reality they should have been attributed to the study variables?
Type 2 error
What is it called if you labeled a significant study as insignificant?
Type 2 error (study was “underestimated”).
In most cases of medicine, which is worse, a Type 1 or Type 2 error?
Type 1
In research, what addresses whether an instrument or test actually measured what it intended to measure?
Validity
What in research refers to the consistency or repeatability of scores>
Raliability
What does an intentention to treat analysis insist on?
That all patients remain in the original groups to which they were initially randomly assigned in the study
How does an intention to treat analysis help maintain power in a study?
If patient can be moved to another group as study progresses, benefits of original randomization process would be lost and bias would creep in
What is number needed to treat?
The number of patients that you would need to treat to prevent one additional adverse effect
True or False, depending on your study group characteristics, the NNT changes?
True
What is the equation for NNT?
NNT = 1/ARR
ARR = Rate in untreated group - Rate in treated group
What is the number of newly diagnosed cases of a disease in a period of time?
Incidence
What is the total number of cases of disease existing in a population?
Prevalence
What is a measure of the spread of individual values around the mean or average value?
Standard deviation
What does standard error describe?
How accurate the sample mean value used in the analysis is compared to the “true” population mean value
True or False: The close the sample mean is to the “true” mean, the more accurate the study results are
True
What is the confidence interval a measure of?
The reliability of your result
What does confidence interval consist of?
Range of values in which you are confident that the true population result will be found
In applied practice, what confidence level are confidence intervals typically stated at?
95%
What does a confidence interval stated at the 95% confidence level mean?
The level at which you are 95% confident that your result lies within the true result of the general population
True or False: Greater levels of variance within a population yield larger confidence intervals and therefore less precise estimates of the measure
True
What does statistically significant mean?
The results have a small chance of having occurred by chance alone
True or False: Statistically significant implies clinical importance
False
What is pre-test probability?
Best estimate of the probability that a condition is present before you start diagnostic testing
What helps to establish post-test probability of a suspected condition?
Diagnostic test result
What is a likelihood ratio of a test?
The likelihood that a person who has the condition will have a positive test result
What indicates by how much the result of the test will raise or lower your pre-test probability and get you closer to the post-test probability?
Likelihood ratio
What is the difference between positive predictive value and pre-test probability?
PPV applies to the test and Pre-test Probability applies to the condition
What does a likelihood ratio greater than 1 do?
Increases the probability that the target condition is present
What does a likelihood ratio under 1 do?
Makes the target condition less likely
How is the absolute risk of developing a condition calculated?
Divide the number of patients who develop disease by the total patients exposed
What does relative risk compare?
The probability of an outcome in the exposed group to the probability of the outcome in an unexposed group
What does an odds ratio represent?
The odds of an outcome in one group compared to the odds of that outcome in another group
Best practice in medical care is derived from what?
Evidence based care
Rank the 5 types of research studies from most valid to least valid
- Randomized, controlled trials
- Cohort studies
- Case-control studies
- Cross-sectional studies
- Case studies
What is considered the gold standard of research designs?
Randomized controlled trials
*Study participants are assigned randomly to one treatment or another
What are 2 major strengths of a randomized controlled trial?
- The minimialization of confounding variables by making the baseline groups equivalent
- Usually blind to participants and those administering the study
True or False: Randomized controlled trials reduce the expectation of bias of the patient and reduce bias in interpreting the results
True
What are limitations of randomized controlled trials?
Cost and time
What type of study follows a group forward or traces then backwards to investigate associated risks for a certain outcome?
Cohort
What are limitations of cohort studies?
Require large sample size, possible confounding variables
When are case-control studies useful?
When you only have a small sample size or a rare outcome
How does a case-control study work?
You start with patients with the disease and then patients without the disease and search for risk factors
What timing component is important for a case-control study?
The risk factor has to come in time before the outcome
What are limitations of case-control studies
Small sample size, risk of confounding variables
What type of study is observational and great at looking at associations between two measured factors at one point in time
Cross-sectional studies
True or False: Cross-sectional studies look at measures over time
False
What are limitations of a cross-sectional study?
Risk of confounding variables and results cannot be used to infer causality
What type of study involves studying the same group of individuals over an extended period of time?
Longitudinal
What are 2 types of studies based on individual outcomes
- Case reports
- Anecdotal reports
*Limited because they have many confounding variables
What uses a well-defined approach to locate and summarize articles related to a clinical question?
Systematic review
What 4 things do systematic reviews require?
- Specific clinical question
- Thorough search for studies
- Clear explanation for which studies were considered and which were not
- Descriptive results section summarizing the findings
What does a meta-analysis do?
Pools all the statistical results collected from a group of articles
Name 7 things a meta-analysis should do
- Specific clinical question
- Thorough search for studies
- Clear explanation for which studies were considered and which were not
- Descriptive results section summarizing the findings
- Present a summary statistic using data from all included studies
- Report whether combining statistics was possible
- Assess for publication bias
What is publication bias?
When some studies weren’t published, but would have affected the data
Which is better, systematic review or meta-analysis?
Meta-analysis
What is a limiting factor for both meta-analysis and systematic review?
Conclusions in both are based on data collected outside of your control
