BIOSTATISTICS

Question 1

specificity

Answer 1

TN / (TN + FP)

1-FP rate

Question 2

what is beta value in biostatics

Answer 2

type II error
b = %

don’t confuse with power = 1 - b

Question 3

what is 1 - beta in biostatics

Answer 3

power

100 - b = %

Question 4

chai square test assess for statistically significant ____ between two or more ____ values

Answer 4

differences
categorical

i.e. percentages, alleles
(not means) = t-test

Question 5

cross sectional study
aims to identify

Answer 5

prevelance

‘snap shot in time’ - **does not **follow patients over time

Question 6

in cross over study, subjects recieve multiple treatments in ___

Answer 6

sequence

Question 7

two or more treatments given to different patients groups
evaluates multiple independant variables simultaneously + interactions between variables

Answer 7

factorial study design

i.e. pts given vitamins: A only, C only, C+E, placebo

Question 8

2 groups: 1 disease, 1 disease free. variables then compared.
study design?

Answer 8

case control

(OR)
disease –> risk factor
(retrospective)

Question 9

2 groups: 1 risk factor, 1 risk factor free. disease then compared.
study design?

Answer 9

cohort

(RR)
risk factor —> disease
(anterograde)
* can be retrospective

Question 10

ranges of correlation coefficient range between

and its strength

Answer 10

1 to +1

close to its margins = stronger

Question 11

in correlation coefficient how are
r>0
r<0
interpreted

Answer 11

r>0 = positive = both values increase/decrease together
r<0 = negative = one value increases, other decreases

direction (positive, negative)
strength (-1 to +1) = stronger at poles

Question 12

what study design can be useful in identifying risk factors potentially contributing to a disparity in the community being observed to help generate a research hypothesis

Answer 12

i.e. cross sectional study

(identify risk factors correlated to obesity in different settings - helps generate hypothesis)

Question 13

OR of 1 means
OR >1 means

Answer 13

equally likely
exposure related to outcome

Question 14

most time efficient experimental design to run

Answer 14

case control

You would need a case (people drinking high arsenic) and control (people with no arsenic).

Question 15

intention to treat requires

Answer 15

all patients to be analysed in the group they were initally randomized

ignores deviations, withdrawals, compliance

used to preserve randomisation with attrition/crossover is introduced to study

Question 16

which point locates ‘maximum sensitivity’

Answer 16

(A) = 100% sensitivity
all people who have disease test positive

moving cut point higher for a test =
decreases sensitivity
increases specificity
(increases FN, decreases FP)

Question 17

equation for NNT

Answer 17

NTT = 1/ARR

(ARR = CER-EER)

CER = control event rate
EER = experimental control rate

don’t forget to do the 1/ at the end (not just calculate ARR)

Question 18

Assume trace elements is control group, how many nulliparous women with confirmed pregnancies would have to be treated with folic acid to prevent one congenital abnormality?

Answer 18

NNT = 1/ARR
(ARR = CER-EER)

ARR = 2.3 - 1.3 = 1%

1/0.01 = 100 women

Question 19

A study is conducted to assess body mass index (BMI) in a group of 100 patients with type 2 diabetes
mellitus. Results show a mean (± standard error of the mean) BMI of 31 kg/m 2 2
(± 4 kg/m ). The 99%
confidence intervals for this measurement are 20.7 to 41.3. The 95% confidence intervals are 23.3 to 38.8.
Based on this information, which of the following best represents the number of individuals who have a BMI
2 2
between 38.8 kg/m and 41.3 kg/m ?
A. 2
B. 4
C. 6
D. 8
E. 10

Answer 19

2

gaussian distribution question

roughly 2.5% between 95-99.7% confidence intervals
(here uses 99%). So 2% x 100 patients = 2 people

Question 20

role of IKB in nuclear factor kappa B signal transuction pathway

Answer 20

NFKB bound to inhibitor protein IKB in cytoplasm (inactive)

when ligand binds (i.e. IL-1) –> activates IKB kiase (IKK complex) –> activated IKK phosphorylates IKB –> releases NFKB –> NFKB translocates to nucleus –> upregulates transcription/translation proinflammatory genes

increased** IL-6 **synthesis induced by this pathway

IKB = inhibitor of kappa B
IKK = IKB kinase complex

Question 21

1-B (power) is the probabilty of ___ the null hypothesis when it is ___

Answer 21

rejecting
truly false

probability of finding a true relationship

= the probability of a statistical test correctly rejecting a false null hypothesis
detects a real effect when one exists

1-type II error

depends on sample size and outcomes being tested

typically set at 80% i.e. if B set at 0.2 = 1-B = 80%

Question 22

B is probablity of commiting a

Answer 22

type II error

fail to reject null hypthesis when it is truly false

will miss true relationships

Question 23

type I error

Answer 23

reject hypothesis when it is true

Question 24

type II error

Answer 24

**fail to reject **hypothesis when it is false

Question 25

a value

Answer 25

maximum probability making a type I error that a researcher will accept | a = 0.05 less than 5% chance it occured under null hypothesis ## Footnote p value = probability of observing a given result due to chance alone (assuming null is true) compared to p value

Question 26

in prospective studies, loss to follow up poses the risk of

Answer 26

atritiion bias | form of SELECTION BIAS ## Footnote doesn't ocur if random loss

Question 27

what study design shows risk factor association with disease, but does not determien causality

Answer 27

cross-sectional | 'snap shot' prevelence ## Footnote i.e. genome mutation and HTN analysed in random population over 1 week, association determined

Question 28

correlation analysis assesses the relationship between 2 _\_\_ variables

Answer 28

**qualitative** i.e. categorical *temp, glucose levels, body composition* | correlation coefficient ## Footnote **chai square - assoc between 2 categorical variables** i.e. male, female i.e. presence or absence myocardial infarction