Posts: 3,675,937
Threads: 734,344
Joined:
Sep 2021
Reputation:
5
wo programs for the treatment of patients with newly detected hypercholestrolemia were tried in a community.Program A was used in one district of the community and program B was used in another.After four yrs 45% of the 2200 patients on program A and 49% of the 1900 patients on program B had been successfully treated for hypercholestrolemia.The difference between the success rates for the two programs was stastically significant pvalue is less than .01.Health officials however decided not to change to Program B in the 1st district because the magnitude of the difference was so small.Which of the following best explains their decision ?
A)They attributed the difference in success rate to chance alone
B)They distinguished between statistical significance and practical importance of the difference in success rates.
C)They felt the p value was too small to justify a decision in favor pf program B
D)They felt the samples were too small to justify a decision in favor of Program B
Posts: 3,675,937
Threads: 734,344
Joined:
Sep 2021
Reputation:
5
anyone please can explain this
i think the answer is B,by excluding the other options
Posts: 3,675,937
Threads: 734,344
Joined:
Sep 2021
Reputation:
5
well based on the question stem what can we say - wasnt due to chance based on p value
samples seem adequate-ish --- not like 100 pts or less
i used the same - exclusion to come to the asnewer - def not A
b is plausible because it is significant but mere difference of 4%
d i excluded, wasnt overtly small sample size
ultimately i chose B
the following is an explanation i found in an old post
Explanation ......Q4
-Statistical significance= Is determined by P value.....and as see the p value is less than .05....hence for sure there is a difference between the two interventions (NOT BY CHANCE ALONE).....If the P Value is greater than .05 u can SAY the difference in the two interventions is due to CHANCE ALONE .....So A is out
-Practical Importance means CLINICAL SIGNIFICANCE not Statistical....and clinical significance MUST ANSWER THE QUESTION ...how effective is the intervention or treatment in the CLINICAL SETTING....And the doctor made the CALL in this case not the researcher....And the way u measure CLINICAL SIGNIFICANCE IS by some of the following methods...EFFECT SIZE ,Number Needed to treat,and Preventive fraction.....
-Already in the question u stated the magnitude of the difference was so small....How do they know ? They calculated NNT....
as u know NNT= 1/49%-45 %=1/4%=25.....Taking Program A as a control (as u know control groups can get placebo or as in this case STANDARD CARE)....
-This means if the health official decided to change Program B as a STANDARD of CARE .....to treat HYPERCHOLSTEROLEMIA in 1 patient 25 should be INVOLVED .....
---NNT=25 is not clinical significant.....hence choice B......
about the other choices....
Choice D....once you find the difference is STATISTICALLY SIGNIFICANT ...the sample size should be assumed LARGE and Enough...Only question sample size if the STUDY is not STATISTICALLY SIGNIFICANT....i.e When Calculated P Value is greater than .05.....
Choice C....Many times the investigator will report the lowest COMPUTED P Value ....In general the smaller the p the BETTER......
And I remember a question on NBME 7 where the computed P value was greater than .05.....and the answer was ....Due to chance alone....but the scenario was almost the same as this question.....
is posted by yeabiruh
Posts: 3,675,937
Threads: 734,344
Joined:
Sep 2021
Reputation:
5
Thanks alot whatsfordinner
That was exactly what i was looking for thank u
