P1 Spring Pp2 Exam 1

Front 1

What is the relative risk equation, given smokers and disease X?

Back 1

RR = smokers with disease X/all smokers divided by
non smokers with disease X/all nonsmokers

eg RR = 27/482 smokers divided by 77/1908 nonsmokers
= 5.6%/4.0%
= 1.4 (40% inc in risk)

RR is the risk of getting disease if exposed vs not exposed

Front 2

How do you determine ARR (absolute risk reduction)?
Give example with smokers.

Back 2

ARR = smokers with disease X/all smokers - non smokers with diseaseX/all nonsmokers

eg ARR = 5.6% - 4.0% = 1.6% (absolute risk reduction)

ARR = Event rate (of exposed) - Event rate (of unexposed)

Front 3

What is NNT (number needed to treat), and it's equation?

Back 3

The number of people that need to be treated to prevent one bad outcome.

NNT = 1/ARR (its the inverse of ARR)

eg NNT = 1/1.6% = 1/0.016 = 62.5 (people would need to quit smoking to avoid one myocardial infarction that would've occurred)

These 62.5 people are smokers and nonsmokers IBLV, not just smokers who have yet to get a MI

Front 4

What's a "construct"?

Back 4

Something that you want to measure, eg blood pressure

Contrast this to a "measure" which is the actual number representing a particular instance of the construct.

Front 5

What's an "odds ratio"?

Back 5

The ratio of: odds of exposure among diseased TO odds of exposure among nondiseased.

OR = smokers with X/nonsmokers with X divided by
smokers w/o X/nonsmokers w/o X

eg 27 / 77 divided by (482-27) / (1908-77) = 35% / 25%
odds ratio is 1.4

Front 6

What's a "confidence interval"?

Back 6

A confidence interval is a range of values that our true value will likely fall within.

eg the confidence interval of our RR value is 1.2 to 1.5 (and our RR was 1.4).

They're used in hypothesis testing, gives us confidence that our value (eg 1.4) is correct.

RR = Relative Risk

Front 7

Types of errors in hypotheses?

Back 7

Random error: accidental(screwup), independent of sample size
Systemic error: an introduced bias (not accidental)

Front 8

Whats a p-value?

Back 8

The probability that the result of our experiment could have occurred by chance alone. If p=0.05 (then there was a 5% probability that could've happened by chance. Anything less than 0.05 is "statistically significant". Note that the p-value assumes the null hypothesis was true (no association, no effect, no difference). The p-value is NOT the prob that the null hypo is true! A low p-value just implies the rarely of the result occuring by chance.

Front 9

How does p-value relate to confidence interval (CI)?

Back 9

When the p < 0.05, we have 95% CI (think of p<0.04 giving us a 96% CI)

Front 10

What is "significance level"?

Back 10

5% typically (comes from p<0.05 which is 5%)

Front 11

With most point values (like relative risk and odds ratio), you get a number above 1.0 (eg 1.4 or 1.6). What is the significance if the CI include 1.0 in its range?

Back 11

When CI includes 1.0 our result is not considered statistically significant, we have no confidence. This is because even though we got 1.4, we couldv'e gotten 1.0, which shows no association between independent and dependent variables.

So a "statistically significant" result has CI that excludes 1.0 AND p<0.05

Front 12

How does confidence interval and sample size relate?

Back 12

The larger the sample size, the more narrow our CI gets, because we're getting closer to the true value. eg sample size of 10 ppl versus 1 million.

The narrower the CI is, the better estimate we have.

Front 13

How does your CI change if you want your confidence to be at least 99% (eg p<0.01)

Back 13

You're CI range gets wider, because like a bell curve it has to encompass the outliers/extremes.

Front 14

What can you find in PDR?

Back 14

•reprinting of package inserts
•for malpractice suits, contains record of older drugs
•listing of poison control centers

Front 15

Whats in Drugs, facts and comparisons?

Back 15

•thorough monographs
•cost comparisons
•drug comparisons
•widely used

Front 16

Whats in AHFS?

Back 16

•extensive monographs on single drug ingredients
•contains off label uses of drugs
•widely used
•highly respected
•includes monthly updates

Front 17

Whats in Martindale’s?

Back 17

•Inter <MISSING>

Front 18

Whats in USP-DI Vol 1?

Back 18

•Its for health care professionals
•Contains off label uses of drugs

Front 19

Whats in USP-DI Vol 2?

Back 19

•Its for the consumer
•Simplified monographs
•Easy to read, good pictures
•No copywrite, so consumer can photocopy it

Front 20

Whats in USP-DI Vol 3?

Back 20

•Legal requirements for perscribing, dispensing, packaging, storage, labeling

Front 21

What are the pregnancy risk categories from Briggs?

Back 21

•A: Controlled studies show no risk
•B: No evidence of risk in humans
•C: Risk cannot be ruled out
•D: Positive evidence of risk
•X: Contraindicated in pregnancy

Front 22

List some major journal indexes used in primary literature

Back 22

•Ovid Medline or PubMed
•PsycInfo
•Education Abstracts
•Social Science Citation Index

Front 23

List the pharmacy related sites available to us.

Back 23

•Access pharmacy (mcgraw hill pharm books)
•StatRef (DI Vol II, AHFS drug info, Taber's med dictionary)
•Lexi-comp
•Facts & comparisons online (Wolters Klower)
•Clinical Pharmacology
•MicroMedex

Front 24

How do you judge a journal/article intellectual influence?

Back 24

Impact factor (number of times its been cited IBLV)

Front 25

Primary vs secondary vs tertiary references?

Back 25

•primary: original research, case reports, clinical trials, cohort studies, eg JAMA, NEJM, AJHP, Pharmacotherapy
•secondary: abstracts (no interpretation/judgement), eg PubMed, Micromedex, Medline(bibliographic DB)
•tertiary: textbooks, reviews, polls, comparisons

Front 26

Whats a MeSH term?

Back 26

•Medical Subject Heading
•Vocabulary used for hierarchical classifying (taxonomies, coding system)
•Articles are given MeSH terms when indexed in Medline (used as a search descriptor, eg "type 1 vs type 2 DM")

Front 27

Whats the standardized way to frame a clinical question?

Back 27

•P (population or patient)
•I (intervention, prognostic factor, exposure)
•C (comparitive intervention)
•O (outcome)

Front 28

Specificity vs sensitivity?

Back 28

When performing a search of online databases/articles, these are the metrics that compare different search engines (eg google)

sensitivity: the proportion of docs you get that are related to my subject (broad)

specificity: the proportion of docs relevant to my question (narrower)



If too few hits, inc sensitivity (widens search), be less specific

If too many hits, inc specificity (narrows focus), by using limiters