Chapter 22

Question 1

Define
Sensitivity
Specificity
Which is affected by prevalence?

Answer 1

Sensitivity = TP/(TP+FN) proportion w/ dz that test positive

Specificity = TN/(TN+FP) proportion w/o dz that test negative

not affected by prevalence

Question 2

define positive and negative predictive values

Answer 2

PPV = TP/(TP+FP)
NPV = TN/(TN+FN)

Question 3

define prevalence

Answer 3

(TP+FN) / (TP+TN+FP+FN)

those w/ disease/those exposed at certain time (your test pop)

Question 4

which values are affected by prevalence?

choose from specificity, sensitivity, PPV, NPV

Answer 4

PPV, NPV

PPV of a test declines w/ prevalence

Question 5

describe the likelihood ratio of a positive test result

Answer 5

= sensitivity/(1-specificity) OR
= true positive rate/false positive rate
represents ratio of the prob that dz is present to the prob that its absent

Question 6

screening tests are more likely to detect patients w more rapidly/slowly progressing disease?

Answer 6

slowly

is this right?

Question 7

explain how the lead time achieved though screening sometimes leads only to perceived increase in survival

Answer 7

a tumor may be dxd earlier dt screening test. but if it is still incurable, then survival seems longer but in fact it was just dxd earlier which is why ppl seem to live longer.

Question 8

a rough guideline is that a minumum of __ outcomes must have occured for each px factor assessed

Answer 8

10

Question 9

describe the parts of the TNM system for staging cancer

Answer 9

T = extent of primary tumor, size, invasion
N = presence/absence of regional ln involvement
M = presence or absence of distant mets

Question 10

clinical trials can be divided into explanatory and pragmatic trials. describe what these terms mean. state where phase I, II, III clinical trials would fit into those categories

Answer 10

Explanatory - evaluates the biological effect of treatment. Phase I and II, not used to formulate tx policy just guide further research. Idealized conditions, tumor response is important NOT survival.

Pragmatic - evaluates the practical effects of treatment, Phase III, used to determine future tx policy, real life conditions, for target pop, see how it effects outcome such as survival

Question 11

briefly describe phases I-III in clinical trials

Answer 11

I - to determine toxicity info and develop tolerable dosing schedule

II - to screen treatment for anti-tumor effects to see what is better to evaluate

III - to determine usefulness of treatments in p management

Question 12

tumor response rate is a good measure of response for traditional chemo cytotoxic agents. what would be a more appropriate endpoint for molecular targeted therapies

Answer 12

time to progression (they are cytostatic)

Question 13

define stage migration

Answer 13

occurs when patients are assigned to a different clinical stage bc of differences in precision of staging rather than differences in true extent of disease

Question 14

the null hypothesis states that the observed difference is dt ___ while the alternate hypothesis states that the observed difference is dt ____

Answer 14

chance

treatment applied in study

Question 15

describe the processes of stratification and blocking in randomization

Answer 15

Stratification - refers to grouping of patients w/ similar px characteristics

Blocking - ensures allocation is balanced for every few patients w/in each stratum

Question 16

Name at least 4 things that influence the # of subjects that are required for a clinical trial evaluating survival

Answer 16

1 - minimum difference in survival rates that is considered significant/important. Smaller differences are harder to find

2 - # of deaths expected. if more die, smaller #s are needed bc they are not censored and longer follow up is not needed

3 - willingness to accept a false + Type I error (the less willing, the more cases you need). Standard p value is 0.05

4 - willingness to accept a false negative Type II error. Again, the more accurate you want to be the more cases you need. Standard Beta is 0.1-0.2. Power of 80-90%.

Question 17

Define power.

Answer 17

Power = 1-Beta.

The ability of a trial to detect a difference when a difference exists.

Question 18

What is meta analysis?

Answer 18

Method by which data from individual randomized clinical trials that assess similar tx are combined to give an overall estimate of treatment effect