need help with a question - (Jan/25/2014 )

Am truly stuck on this question. Can someone help me know the answer to this and how you reached your answer? Thanks in advance. 

In a study designed to examine the effect of salt intake (high vs low) on blood pressure, if the group with the usual (higher) salt intake is being more intensively screened for hypertension than those with low salt intake, and the hypothesis being tested is that lowering salt intake lowers risk of hypertension, then will the observed measure of association between reducing salt intake and lowering risk of hypertension be an underestimate or overestimate of the true benefit of salt reduction on hypertension?

A. Over estimate

B. Under estimate 

So the fault with this study is that the group with high salt diets is more intensively screened for high blood pressure than the group with low salt diets.  What will probably happen is that more cases of high blood pressure will be found in the high salt group than the low salt group.  Hence you will have a study the finds more high blood pressure with a high salt diet than a low salt diet.

So my guess is that you would have an over estimate.  (I will give you 75% odds I am correct--either way you have a 50/50 chance of getting this answer correct!)

The description of the study is a little ambiguous.  Are the participants going to be put on a low salt diet and then re-evaluated for blood pressure?  Or is this just a correlational study?  Either way the participants in either group need to be screen to the same level of stringency to have an unbiased result.

Good luck.

Thanks. Overestimate makes sense. Because the high salt group was being screened more, then the results would show that low salt would be better (overestimate) because the high salt group was overscreened. 

I need help with another question. I am stuck between answers of chance or confounding. I don't think generalizability is the answer. 

A mother brings her 10-year old child to a physician, to ask if her child should use a new drug for asthma that is being widely advertised in the media. The physician reviews the literature, and finds that the claims in the advertisements are based on the results of a single trial, in which 40 consecutive adults in the asthma clinic of the investigator’s hospital were randomly allocated to the new drug versus the current standard of care. Neither the patients nor the investigators could be blinded, as the new drug had a characteristic aftertaste that could not be masked. The trial compared the number of visits to the emergency room over the next six months, and found a 50% reduction in emergency room visits with the new drug, that was not statistically significant at the 0.05 level (P=0.27). The physician is concerned because since there are so few patients in the study, it may give a misleading impression of the effectiveness of the new drug. What is the underlying basis for this concern?

A. The role of chance

B. The role of confounding

C. Generalizability

DeeAnnVisk on Sat Jan 25 17:20:56 2014 said:

So the fault with this study is that the group with high salt diets is more intensively screened for high blood pressure than the group with low salt diets.  What will probably happen is that more cases of high blood pressure will be found in the high salt group than the low salt group.  Hence you will have a study the finds more high blood pressure with a high salt diet than a low salt diet.

 

So my guess is that you would have an over estimate.  (I will give you 75% odds I am correct--either way you have a 50/50 chance of getting this answer correct!)

 

The description of the study is a little ambiguous.  Are the participants going to be put on a low salt diet and then re-evaluated for blood pressure?  Or is this just a correlational study?  Either way the participants in either group need to be screen to the same level of stringency to have an unbiased result.

 

Good luck.

Huh, I guess its comparing means issue. Here high salt mean has higher reliability than low salt diet mean. will depend on response variability in samples of low salt group.don't think its 75% odds if r not aware of previous salt - hypertension relation. 

On the 2nd problem: I would exclude confounding because, although the study was less-than-optimal, the text does not .suggest that the effect may be due to another parameter somehow related to drug intake.

Of course generalizability is a problem here because the tested population were only adults and the patient in question is a child. Viewed this way, generalizability might be an answer, but I'd go for something else.

The text states: "The physician is concerned because since there are so few patients in the study, it may give a misleading impression of the effectiveness of the new drug. What is the underlying basis for this concern?"

Since the question explicitly asks for the rationale to the physician's concern (and not to what else may be worried about, like generalizabitlity, which is not mentioned by the physician), I'd go for the role of chance because with so few people and a even then a nonsignificant result, the probability is high that it is only a chance finding and not a real effect.

However, I'd like to add that this is just my interpretation of the question and I may be wrong, of course.

I agree with the above post.  The clue is in the question "The physician is concerned because since there are so few patients in the study".  So I am pretty sure the answer would be chance.

The study could easily be just showing a placebo effect, because the taste of the drug could not be masked.  Hence the participants KNOW whether or not they are getting the drug.  I would also be concerned as a pediatrician that the child might react to the drug differently, but 10 year old lungs are probably much closer to an adult lung than to say an infant lung. This would be an issue of generalization.

Very unlikely to be a confounding variable...