Academic Integrity: tutoring, explanations, and feedback — we don’t complete graded work or submit on a student’s behalf.

Question 15 A group of 1676 children were followed from birth to the 1st year of

ID: 3364963 • Letter: Q

Question

Question 15

A group of 1676 children were followed from birth to the 1st year of life for the development of otitis media (OTM), a common ear condition characterized by fluid in the middle ear (either the right or the left ear) and some clinical symptoms (e.g., fever, irritability).

Below is the output after running the Cox proportional-hazards model with the OTM data, the response variable is the time to developing OTM and the risk factors are:

1. 2. 3. 4. 5. Sex (coded as 1 if female and 0 if male) At least one sibling with Hx (history) of ear disease (coded as 1 if yes and 0 if no) Some sibs, but none with a Hx of ear disease (coded as 1 if yes and 0 if no) Ever breastfed during the 1st year of life (coded as 1 if yes and 0 if no) Mother smoking (ever during the 1st year of life) (coded as 1 if yes and 0 if no) The results are as follows: Variable Sex Sibs with Hx ear disease Sibs, no Hx ear disease Ever breastfed Mother smoking Regression coefficient -0.080 0.331 0.207 0.047 0.223 Standard error 0.069 0.073 0.120 0.075 0.106 1.16 4.53 1.73 0.63 2.10 value Relative risk .25 .001 .085 .53 035 0.92 1.39 1.23 1.05 1.25

Explanation / Answer

Question 15

Solution:

There are significant effects of maternal smoking, even after controlling for the other covariates (p = 0.035) The relative risk = hazard ratio = 1.25, thus, children whose mothers smoke are 25% more likely to develop a 1st episode of OTM at any point in time 1st year of life given that they haven't developed OTM previously compared with children whose mothers do not smoke, all other risk factors being equal. In addition, there was a significant effect of having a sibiling with a Hx of ear disease varsus not having any sibs (RR = 139, p < 0.001). The other covariates were not statistically significant.
In addition, it is possible to obtain confidence limits for both the true regression coefficients and RR's . Specifically, a 95% CI for the regression coefficient for maternal smoking is
0.223 ± 1.96(0.106) = (0.015, 0.431)
The corresponding 95% CI for RR is [exp(0.015), exp(0.431) = (102, 154)