ITE CA-2 stats stuff Flashcards
nominal data
nominal data (also known as nominal scale) is a type of data that is used to label variables without providing any quantitative value. … One of the most notable features of ordinal data is that, nominal data cannot be ordered and cannot be measured
compare 2 different groups of nominal data
chi-squared or fisher’s exact test
parametric data
When we assume that the distribution of some variable (like course grades) follows a well-known distribution (like the Normal distribution), that can be boiled down to knowledge of just a couple of parameters (like mu and sigma), and then we use that assumption in the performance of some statistical test, we are said to be using a parametric test.
nonparametric data
When you can’t make such an assumption about the underlying distribution of a variable, before looking at the data, and must instead use more robust (but frequently less powerful) methods as a result, to answer the same kinds of questions, then you are using a nonparametric test.
compare 2 different parametric interval groups
un-paired t-test (2-sample)
The unpaired t-test allows for comparison of two populations with respect to a single variable with continuous data. In our example, one population is the group of patients receiving remifentanil and the other population is the group receiving sevoflurane. Our single variable is the mean arterial pressure.
compare 2 different ordinal or nonparametric interval groups
Wilcoxon-Mann-Whitney test
Wilcoxon-Mann-Whitney is a nonparametric test designed for studies for ordinal numbers (ranking: 1st, 2nd, 3rd, etc.).
ordinal data
nominal variables are used to “name,” or label a series of values. Ordinal scales provide good information about the order of choices, such as in a customer satisfaction survey. Interval scales give us the order of values + the ability to quantify the difference between each one.
compare more than 2 different nominal groups
chi-square or fisher’s exact test
Chi-square testing is for comparison of two populations with respect to a single variable with discrete (not continuous) data.
Chi-square test is used to compare categorical data and not means.
compare more than 2 parametric interval groups
one-way ANOVA
Analysis of the variance (ANOVA) is similar to a t-test except that it is designed to analyze >1 variable.
Analysis of variance (ANOVA) is a statistical test used to compare means between more than two groups or test differences in repeated measurements within the same group.
compare more than 2 different ordinal or non-parametric groups
kruskall-wallis
compare 2 paired nominal groups
McNemar
compare 2 paired parametric groups
paired t-test
compare 2 paired ordinal or non-parametric groups
Wilcoxon-signed-rank test
compare more than 2 matched nominal groups
Repeated measures logistic regression
compare more than 2 matched parametric interval groups
Repeated measures ANOVA
compare more than 2 matched ordinal or non-parametric groups
Friedman test
Case control cannot measure
In a case-control study, you cannot measure incidence, because you start with diseased people and non-diseased people, so you cannot calculate relative risk
cohort study
The cohort study design identifies a people exposed to a particular factor and a comparison group that was not exposed to that factor and measures and compares the incidence of disease in the two groups
case control
The case-control design uses a different sampling strategy in which the investigators identify a group of individuals who had developed the disease (the cases) and a comparison of individuals who did not have the disease of interest. The cases and controls are then compared with respect to the frequency of one or more past exposures. If the cases have a substantially higher odds of exposure to a particular factor compared to the control subjects, it suggests an association.
odds ratio
AD/CB
Number needed to treat
1/absolute risk reduction
relative risk
(A/[A+B])/(C/[C+D})
2 std dev contain
95%
risk
frequency at which a specific event occurs
What is AIMS
The use of computerized anesthesia information management systems (AIMS) is increasing in modern anesthesia practice. AIMS are a form of the electronic health record which allows automatic collection, storage, and presentation of patient data during the perioperative period.
Standard error of the mean
SE = SD / square root n
1 std dev contain
68%
Specificity
Specificity = TN / (TN+FP). Specificity is the percentage of time a test is truly negative. By definition, it takes into account the cases which yield a false positive (FP) test.
Sensitivity
Sensitivity = TP / (TP+FN), the chance (%) to correctly detect the disease or problem. Sensitivity “rules out” the disease
Positive predictive value
Positive predictive value = TP / (TP+FP), the chance (%) that a positive test result means that the subject actually has the disease or problem
Negative predictive value
Negative predictive value = TN / (TN+FN), the chance (%) that a negative test result means that the subject does not actually have the disease or problem
Multivariate logistic regression produces what measure for the outcome
Multivariate logistic regression produces “adjusted” odds ratio, where “adjusted” is short for “adjusted for known confounding variables”. Typically the adjusted odds ratio is lower than the unadjusted odds ratio, and some exposure factors may no longer be significantly associated with the outcome after adjusting.
difference between odds and risk
The odds of an outcome are calculated as number of people with the outcome / number of people without the outcome. For example patients with perioperative MI / patients without perioperative MI. This is not calculated as people with outcome / all people, which would be probability (i.e. risk).
unadjusted odds ratio
The “unadjusted” odds ratio (OR) is calculated as the odds of the exposed / odds of non exposed. For example, odds of perioperative MI in COPD patients / odds of perioperative MI in non-COPD patients.
Multivariate logistic regression
Multivariate logistic regression is a technique to address confounding, and produces “adjusted” odds ratios. For example, the unadjusted OR for COPD patients developing perioperative MI might be 2.4 (95% CI 1.5-3.3), and the corresponding “adjusted” OR might be 1.2 (95% CI 0.6-1.8).
power
power = 1-β
Practically speaking, the power tells us the chance that the null hypothesis (e.g. no treatment effect from a particular drug) will be rejected when an alternative hypothesis is actually true (e.g. a drug does have a treatment effect). The larger the sample population (usually denoted as “n = some number”), the greater the power.
3 std dev contain
99%
1 std dev contains
68%
Which of the following statistical tests should be used to compare the effect of sevoflurane versus the effect of remifentanil on mean arterial blood pressure in patients undergoing craniotomy?
unpaired t test
What test is used to compare multiple variables each with discrete values.
Logistics regression is used to compare multiple variables each with discrete values.
Logistics Regression
(“Chi-Squared on steroids”)
test to analyze single variable survival (events)
kaplan-meier analysis
test to analyze multivariable survival (events)
Cox Proportional Hazard Analysis
Three surgeons wish to compare their mean operative durations for identical procedures. Each has done ten of these operations and the data for each are normally distributed. Which test is MOST appropriate?
Analysis of variance (ANOVA)
compare means of two groups
Student T-test is used to compare means of two groups
main advantage of registry studies
describe rare events (but cannot calculate incidence)
what test use for comparing mean length of stay between patients having open or endovascular aortic aneurysm repair
t-test
what test to compare rates of infection between 2 groups of patients
chi square
graph to evaluate the agreement between two measurement techniques
A Bland-Altman plot is used evaluate the agreement between two measurement techniques.
best measure of central tendency for ordinal data
Ordinal data sets refer to data sets that have a natural numerical ordering, e.g. the verbal numeric pain rating scale. With ordinal data, however, the intervals between the numbers may not be the same. For example, the difference between a 5 and a 6 on the verbal numerical pain rating scale may not be the same as the difference between a 9 and a 10 on the same scale. When using an ordinal scale, the central tendency of a group of items can be described by using the group’s mode.
Factors that increase statistical power include:
Factors that increase statistical power include: increasing sample size, increasing the effect size, increasing alpha (e.g. p-value), and reducing population variability (e.g. standard deviation).