EPI 202 Lab 2 Practice Problem Solutionsa.In a study of malnutrition and diarrhea in children under age five in the Sudan, 154 episodes of diarrhea were observed in 38.51 person-years of observation among children who were 75% or less of their expected weight for age, and 287 episodes of diarrhea were observed in 88.83 person-years of observation among children who were 90% or more of their expected weight for age (El Samani et al.; Am J Epidemiol, 1988). Suppose the data from this study were broken down by sex as follows:MalesFemales<75% WFA>90% WFA<75% WFA>90% WFACases75150Cases79137PY20.548PY18.0140.83From this data, we can see that, among unexposed children (those with >90% weight forage), the IRR of diarrhea comparing females & males is given by: IRR=IR(Females|≥90% WFA)/IR(Males|≥90% WFA)=(150/48)/(137/40.83)=0.931Additionally, we can see that the OR comparing the odds of exposure between the female and male study bases is given as: OR=Odds(Exposed|Female)/Odds(Exposed|Male)=(18.01/40.83)/(20.5/48)=1.03The crude IRR = IR(<75% WFA)/IR(>90% WFA) = (154/38.51)/(287/88.83) = 1.24 a.Calculate the summary incidence rate ratio for the association between weight for age and rate of diarrhea. Interpret your answer.aibiN0iN1iTii=1 (male)751504820.569i=2 (female)7913740.8318.0159Summary Incidence Rate Ratio

¿[(75∗4869)+(79∗40.8359)][(150∗20.569)+(137∗18.0159)]=106.88486.385=1.237After adjusting for sex, the incidence rate of diarrhea is 1.24 times higher in children whoare less than 75% of their expected weight for age compared to children who are 90% ormore of their expected weight for age, assuming no residual confounding, confounding by other variables, selection bias, or information bias.b.Calculate the summary incidence rate difference for the association between weight for age and rate of diarrhea. Interpret your answer.Summary Incidence Rate DifferenceValues for a, b, and N can be found in the previous table.*Note: the weight is inverse variance of the strata (difference)Weights for males: (20.52)(482)/(75(482)+150(20.52))= 4.1056066147Weights for females: (18.012)(40.832)/(79(40.832)+137(18.012))= 3.0699740937Sum of weights: 4.1056066147+ 3.0699740937= 7.17558IRD for males: (75/20.5) - (150/48) = 0.5335IRD for females: (79/18.01) - (137/40.83) = 1.03107Numerator of summary IRD: 4.1056066147(0.5335) + 3.0699740937(1.03107)= 5.3557Denominator of summary IRD= sum of the weights= 7.175585.3557/7.17558= 0.746 cases/person-year =74.6/100 PYAfter adjusting for sex, there were 74.6 more cases of diarrhea per 100 person-years in children who are less than 75% of their expected weight for age compared to children who are 90% or more of their expected weight for age, assuming no residual confounding, confounding by other variables, selection bias, or information bias.c.Test the hypothesis that weight for age has no association with rate of diarrhea, after stratifying by sex. Interpret the numeric results.

H0: The rate of diarrhea does not differ between those children who are less than 75% oftheir expected weight for age compared to children who are 90% or more of their expected weight for age, after stratifying by sex.

#### You've reached the end of your free preview.

Want to read all 9 pages?

- Summer '14
- FrancisCook