Introduction to Biostatistics Flashcards
What are 2 key attributes of data measurements?
- magnitude
- consistency of scale (mmHg, cm, lbs)
- rational/absolute zero. (but does not have an effect of what test to use for analyzing data)
What are the three categories, that test data can be organized into based on ‘yes’/’no’ of 2 important key attributes?
- nominal
- ordinal
- interval/ratio
What components would direct the use of a nominal categorization?
- no magnitude, therefore no scale consistency, and no rational zero.
- labeled variables with no quantitative quality.
- can arrange heights into specific category of being taller or shorter than 50in.
What is meant if a null hypothesis possesses superiority?
- the collected data suggests the testing object will work better than its comparative.
What is meant if a product is deemed equivalent?
-it does not show proof of working better than the comparator, but works just as good as.
What does noninferiority mean?
-it is the new up and coming way to describe products that can be categorized as being either equivalent to or better than the comparator. Data not available to indicate superiority or equivalency, but it is not worse than the comparator.
What assumptions must be made for interval tests?
assume:
- normal distribution
- equal variances which can be tested with the Levene’s Test
- data is randomly derived and independent
If the interval data is not able to be normally distributed what two way allow for the analysis of the data?
- use of non-parametric test
- use a z score or a log test in order to transform that data to become normally-distributed
Type I error?
the ability to find the difference if one truly exists
Type II error?
claims no difference is present when in reality a difference is truly present
What is Power?
- (1 minus Beta)
-the ability of the study test to detect a true difference when one truly exists.
Explains the level of accuracy of accepting or rejecting the Null Hypothesis
How can you minimize the difference between groups that are deemed significant?
-increase the number of study subjects. Thereby reducing the difference between groups
If the p-value is lower than the pre selected value what course of action is soon to follow after?
- there is a statistical difference somewhere
- another test is performed to evaluate where the difference is
- the lower (closer to 0) the less chance of a type I error
What are confidence intervals?
-pre calculated values that display where the true relationship exists in a high and low range of data.
What are the confidence intervals based on?
- variation in sample
- sample size
What can you conclude of a ratio (ORR,RR,HR) that crosses the value 1.0?
- nothing significant can be concluded as this suggests that x amount benefit and y amount are hindered.
What can you conclude for comparison data that is not a ratio, that crosses over the value 0.0?
- nothing significant can be concluded as this suggests that x amount benefit and y amount are hindered.
What question must we always ask when reviewing the findings of a study?
- does the statistical data represent significance in the clinical setting?
What does a correlation test represent?
-provides a quantitative measure of strength and direction between variables betweem -1.0 to 1.0
What does a partial correlation test able to control?
- controls for confounding
The contingency coefficient is used when?
-nominal correlation test
Spearman correlation is used for ________
-ordinal correlation test
Pearson correlation test is used for ________
- interval correlation test
- no linear correlation if the p>0.05, but a non-linear correlation can exist.
The tip and hint for a survival test is what word?
time
How are survival tests best represented?
- Kaplan Meier graph which compares the proportion of event occurrences between groups
The log-rank test is used for ______
a nominal survival test
The Ordinal survival test is what?
cox-proportional hazards test
What test is used for interval survival test?
Kaplan- Meier test
Name a nominal regression test.
- Logistic regression
- tipped off by “prediction”
Name a ordinal regression test
- multinomial logistic regression
- tipped off with the word “prediction”
Name an interval regression test
- Linear regression
- tipped off with the word “prediction”
A nominal data set has 2 groups of independent data, what test is used?
Pearson’s chi-squared test
- ≥2 groups when cell count is less than 5 use Fisher’s Exact
Nominal data with ≥3 groups of independent data use what test?
Chi-squared test of independence
- cell count has to be >5
- if the p-value
Bonferroni Test
-subsequent test for nominal data of ≥3 independent groups, to determine specifically what group is different
Nominal data that is paired in 2 groups?
-McNemar test
Nominal data with ≥3 paired groups??
- Cochran test
- -use the bonferroni test afterwards to establish exactly where the difference lies.
What are the keywords for a paired or related data?
- pre vs post
- before vs after
- baseline vs end
Ordinal data with 2 independent groups
-Mann-Whitney Test
Ordinal data with ≥3 independent groups
- Kruskal-Wallis Test
- -compares the median values between groups
- -post-hoc analysis required with >3 significant comparisons
Ordinal Data with 2 paired groups
Wilcoxon Signed Rank Test
Ordinal data ≥3 paired groups
Friedman Test
- -compares the median values between groups
- -post-hoc analysis required with >3 significant comparisons
- -non-parametric interval test
What are the post-hoc analysis tests with >3 significant group comparisons?
- Student Newman Keul Test
- Dunnett Test
- Dunn Test
The Student-Newman-Keul Test
- compares all pair that could be possible A-B, A-C, B-C
- requires equal group sizes
Dunnett Test
- all groups are compared to a single control
- equal size groups
Dunn Test
Compares all pairs possible when the groups are NOT of equal size
Interval data with 2 independent groups??
- student T test, specific for 2 groups
- -able to use ANOVA, MANOVA as well
Interval data ≥3 independent groups
- Analysis of Variance (ANOVA)
- Multiple analysis of variance (MANOVA)
- > 3 significant group comparisons reauire post-hoc test
Interval data with ≥3 groups of independent data with Confounders
- ANCOVA analysis of Co-variance
- MANCOVA
- -controls for confounding effects
Interval data with 2 paired groups?
-paired T test: compares mean between related groups
Interval data with ≥3 paired groups??
- Repeat measures of ANOVA
- Repeat Measures of MANOVA
Interval data with ≥3 paired groups with confounders??
- Repeated measures of ANOVA
- Repeated measures of MANOVA, if there are more than 2 dependent variables
Interval data initial analysis has >3 significant comparison groups. What type of post-hoc test can be performed to isolate the cause?
- Turkey test: more conservative than the Stu.N.K.
- Scheffe test: less affected by violations
- Bonferroni Test
The Kappa statistic is what?
-grading scale of agreement between evaluators
How is a +1 Kappa score interpreted?
- observers were able to classify people the exact same way
How is a 0 Kappa score interpreted?
-no relationship other than that which could be due to chance
How is a -1 Kappa score interpreted?
- Classification is completely opposite of the initial thought process
What 4 questions guarantee the right test is selected?
1 what type of data is being evaluated
2 what type of comparison is desired
3 how many groups are being compared
4 is the data related or independent
What is the purpose of the National Clinical Trials Numbers that is assigned once the research protocol has been submitted to the clinicaltrials, before the start of the trial?
- used to reduce publication bias
- The site contains up to date information on locating interventional studies in progress or beginning
What are three things most clinicians want to know how to do with respect to studies?
- determine differences of study design and what design is most appropriate for hypothesis of question
- evaluate how the results are skewed by study design
- know the strengths and weaknesses of each design
What are two important concepts that allow a reader to accurately assess a study?
-transparency with methodology and with findings
Where can a person go to gain access to different checklists and other tools that can help in evaluating studies and study designs?
equator network
An interventional study uses what checklist, predominately?
- Consort for randomized clinical trials
- -Consolidated Standards of Reporting Trials
When would extension documents be used, in evaluating a CONSORT checklist?
- these are extensions to deal with:
- -non inferiority
- -equivalence trials
- -cluster trials
A pragmatic trial will use what type of extension document of the CONSORT checklist?
-this extension helps describe a randomized controlled trial with the sole purpose of informing about a clinical practice
What is PRISMA, and what does it stand for?
- systematic review of multiple randomized trial
- Preferred Reporting Items for Systematic Reviews and Meta-Analyses
When is STROBE used and what does it stand for?
- used for observational studies (cohort, cross-sectional, case-control)
- STrengthening the Reporting of OBservational Studies in Epidemiology
What is STROBE-ME?
This is an extension document specific for STROBE studies dealing with molecular epidemiological studies
What is STREGA?
Also another extension of STOBE, that evualates the genetic association studies
What is TREND used for and what does it stand for?
- reports evaluations with non-randomized designs of behavioral and public health interventions
- Transparent Reporting of Evaluations with Non-Randomized Designs
REMARK is another type of study test that is used for what, and stands for what?
- used for tumor marker prognostic studies
- REporting recommendations for Tumor Marker prognostic Studies
GRIPS?
-genetic risk prediction studies
Diagnostic studies utilize what test?
- STARD
- standards for reporting of diagnostic accuracy studies
QUADAS-2
- systematic review of multiple diagnostic studies
- quality assessment of studies of diagnostic accuracy in systematic reviews, 2nd edition