1

Question 1

In a randomized clinical trial conducted to see the effect of medical therapy alone Vs medical+CABG in MI pts, 17%of patients in the medical group ( deteriorated to the point where CABG became necessary) and 91%in the M+CABG group underwent CABG at the end of follow up period. Combining both groups n analyzing outcomes for patients who actually underwent CABG Vs medical therapy introduces which bias?

Answer 1

Selection bias - known as susceptibility bias (confounding by indication)
- Initial-to-treat analysis may b used to deal with selection bias- compares the initial groups (original intention) regardless of the eventual Rx.
- the one described in the question is known as
As-treated analysis- usually performed to gauge the effectiveness of the Rx itself with less regard for potential confounders.

Question 2

What is hazard ratio? How r the values interpreted?

Answer 2

It’s the ratio of an event rate in the treatment group compared to an event rate in the nontreatment group
-<1 indicates that the treatment group had a lower event rate n
>1 indicates that the Rx group had a higher event rate

Question 3

Two studies conducted on different samples from the same population to assess OCP use as a risk factor for DVT. Study A- Relative risk of 2 n 95% CI of 1.2-2.8. Study B - RR of 2.01 and 95% CI of 0.8-3.1
R the tests statistically significant? If yes, how r the RR values interpreted?
-what can u say about the sample sizes used in the two studies?

Answer 3

Study A is statistically significant but B is not, as the CI (0.8-3.1) contains the null value 1. RR>1 and RR<1 indicate exposure is associated with increased and decreased risk of the disease respectively. Their corresponding p-values will be <0.05 for A and >0.05 for study B.
-the wider CI in study B is most likely due to smaller sample size which decreases the study’s power (its ability to detect the difference between exposed n unexposed subjects). Increased sample size➡️narrow CI➡️can help exclude the null value n increase the significance.

Question 4

What’s p-value?

Answer 4

• p-value is the probability of (observing a given or more extreme) results due to chance alone, assuming that the null hypothesis is true. P value<0.05 reflects that there is a very low probability that the result was due to chance alone

Question 5

A study is conducted to see if polyunsaturated fatty acids taken during pregnancy (22-26weeks) could reduce the risk of asthma in the offspring. 700 children born to 700 asthmatic women where followed for 3 yrs and 145 of them are found to be asthmatic. Out of these, 17.1% were born to mothers who received LcPUFA, and 24.3% placebo. Hazard ratio 0.71, 95% CI 0.53-0.95, p-value 0.01
-how many pregnant women would need to be treated to prevent a case of asthma in the offspring?

Answer 5

NNT=1/ARR
ARR=24.3-17.1=7.2%
NNT=1/7.2%=1/0.072=13.89~14

Question 6

Correlation coefficient assesses a linear relationship between 2 variables
Null value?
Direction of relationship?
Strength of relationship?

Answer 6

Null value is 0

- the closer r is to its margins (-1 or 1), the stronger the association. In general, r>0.5 or r

Question 7

Difference between two sample t test, ANOVA, chi-square

Answer 7

t test- to compare two means
ANOVA- three or more means
Chi-square- categorical data

Question 8

In a normal distribution, the values of 1,2,3 standard deviations from the mean are?

Answer 8

1SD-68%
2SD-95%
3DD-99.7 %

Question 9

In a case control study, the odds ratio can be a close approximation of relative risk based on which assumption?

Answer 9

If the disease (the outcome) is uncommon in the population - “rare disease assumption “

Question 10

If subjects of a study over or under-report exposure hx due to perceived social stigmatization eg. Responding to an interviewer Vs an anonymous mailed questionnaire
What’s the bias?

Answer 10

Reporting bias

Question 11

Significant loss of study participants may cause which type of bias?

Answer 11

Attrition bias, a kind of selection bias.

Question 12

If subjects of a study change their behavior because they r aware that they r being observed, what type of bias will take place?

Answer 12

Hawthorne effect

Question 13

In a randomized clinical trial, randomization is said to be successful when—?

Answer 13

When a similarity of baseline cxc (eg. Age, sex,…) of the patients in the Rx n placebo groups is seen

Question 14

A study design utilizing different interventions each intervention with two or more variables that are studied independently
-(eg, randomization of study subjects in to 3 different intervention groups-metoprolol, ramipril, amlodipine) to see which one of these drugs causes hyperkalemia with the greatest frequency. In addition, the drugs also have two different BP end points rather than one common endpoint to see what effect they also have on BP- the study is best classified as?

Answer 14

Factorial design

Question 15

Attributable risk represents?

It’s calculated as?

Answer 15

ARP (attributable risk percent)
Represents the excess risk in a population that can be explained by exposure to a particular risk factor
ARP= (RR-1)/RR= 1-1/RR

Question 16

Absolute risk reduction (ARR) and number needed to treat formula

Answer 16

ARR= event rate in the Rx group-event rate in the control group.
NNT = 1/ARR

Question 17

On the receiver operating characteristic curve (ORC) a shift of the curve 1) upwards 2) to the right , for a given cutoff point indicate ?

Answer 17

Upward- increased sensitivity ( true positive rate on the y axis)
To the right- decreased specificity (x-axis represented by false positive rate 1-specificity)

Question 18

Is accepting a flash drive as a gift from a pharmaceutical company ethical for a physician?

Answer 18

No! - a flash drive, although not of high monetary value, doesn’t directly benefit patient care.
- accepting nonmonetary gifts from interested parties only if the gifts directly benefit patient care and are of small monetary value.