261947
Description
Flashcards by Michael Riben, updated more than 1 year ago
|
|
Created by Michael Riben
over 10 years ago
|
|
| Question | Answer |
| how do you calculate the expected value? | [Prob*value of win] + [prob*value of lose] |
| What is expected Utility | function of value and also risk aversion, personal pref |
| What are sources of Decision Bias? | diagnostic inference, blois funnel, making decision based on opinion is subject to predictable patterns of bias |
| What is availability Bias? | overestimation of the probability of unusual events because of recent or memorable instances |
| What is Representative Bias ? | overestimating rare disease by matching patients to typical picture of that disease |
| What is anchoring Bias? | Failure to adjust probability of a disease or outocme based on new information akin to premature closure |
| What is a value induced bias | Overestimate of the probability of an outcome based on value associated with outcome i..e not wanting to miss brain tumor |
| What do sum of all probabilities for a situation equal to? | 1 |
| What are conditional probabilities | Probability of X given y p(X|Y) |
| What are conditional probabilities | Probability of X given y p(X|Y) |
| what is a Cost effective analysis | Value of outcome nodes becomes units of cost instead of the values used in our example (life = 1, death = 0) |
| What is time trade off? | choose between X time in state of illness vs. Y time in state of perfect health |
| What is the standard gamble? | choose between X time in state of illness vs. therapy with known risk of cure or death |
| What is the incremental cost/effectiveness Ratio? | ICER = (C1-C2)/(E1-E2) measure of change in cost with change in unit of effectiveness (usually the QALY) |
| What is the incremental cost/effectiveness Ratio? | ICER = (C1-C2)/(E1-E2) measure of change in cost with change in unit of effectiveness (usually the QALY) |
| What is sensitivity? | Equal to True Positive Rate = TP/(TP+FN) |
| What is Specificity ? | Equal to True Negative Rate - TN/(FP+TN) |
| What is sensitivity out or snOUT? | sensitivity test allows you to rule a condition out . you want to minimize false negative so you can trust a negative test |
| What is spIN | Specifity tests allow you to rule condition in, minimize false positives so you can trust a positive test |
| What is the Positive Predictive value? | Characteristic of test in context of condition and dependent on prevalence of a condition in population PPV = TP/(TP+FP) |
| What is negative predictive Value? | Characteristic of test in context of condition and dependent on prevalence of a condition in population NPV = TN/(TN+FN) |
| What is negative predictive Value? | Characteristic of test in context of condition and dependent on prevalence of a condition in population NPV = TN/(TN+FN) |
| What do Sensitivity and Specificity of test assume? | Assume you know the disease state and it is used to determine if test performs as tested |
| What does PPV and NPV assume? | Assumes you know the test results and your trying to determine the disease state |
| What is a useful way to remember S, Sp, PPV, and NPV? | s= given a known disease,what is probability of a positive test Sp= Given the patient doesn't have disease, what is the probability of negative test? PPV- Given a positive test, what is probability of patient having disease NPV= given a negative test, what is probability patient doesn't have disease |
| What is prevalence of disease? | TP+FN/Total # |
| What is prevalence of disease? | TP+FN/Total # |
| What is the Receiver Operating Characteristics? | Graph of sensitivity vs. 1-specificity (TPR vs FPR) |
| What is the area under curve in a ROC curve --> | Allows to compare different tests to determine which is better at distinguishing disease from normal ( a higher under care is betteR) |
| What is the relative risk? | RR = prob (Disease|Exposure)/ P(disease|no Exposure) |
| Bayes theorem | Probability of disease given a pos test = prob of positive test given disease (sensitivity) multiplied by Prob of disease (prior prob) devided by prob of positive test ( |
| What are Odds? | Odds X = P(X)/1 - P(X) |
| what does Post test odds =? | Pre-test odds X Likelihood Ratio |
| What is the Positive Likelihood Ratio? | LR+= sensitivity/1-Specificity = TPR/FPR |
| What does negative Likelihood ratio ? | LR- = 1-sensitivity/Specificity = FNR/TNR |
| What is the Fagan Nomogram | Allows you to determine the post-test probabiltiy given the pre-test probability (prevalence) and the positive Likelihood ratio |
Want to create your own Flashcards for free with GoConqr? Learn more.