Download PHM2612 Assignments PHM2612 Module 13 Confounding & Effect Modification Correct Answers and more Quizzes Public Health in PDF only on Docsity!
Assignments PHM Module 13. Confounding & Effect Modification Complete this quiz on Canvas by or before the due date listed on the calendar. Please note that you will only have one attempt to complete the quiz; however, there is no time limit. Note that this is NOT the document you will submit for a grade. You may use this document to answer the questions prior to submitting your answers on Canvas. (100 points)
- Which of the following images represents confounding? (e= exposure, o= outcome, v= other variable) (2 points) C.
- Like information bias, confounding can result in inaccurate estimates of association. ( points) A. True B. False
- In the situation where the True RR = 0.70 and the confounded RR = 0.85, which of the following statements is true? (2 points) A. This is an example of positive confounding, where the confounded RR is biased towards the null B. This is an example of positive confounding, where the confounded RR is biased away from the null C. This is an example of negative confounding, where the confounded RR is biased towards the null
D. This is an example of negative confounding, where the confounded RR is biased away from the null
- If an effect modifier is present, the stratum-specific ORs will be different from each other. (2 points) A. True B. False
- Confounding can be controlled only in the analysis phase of the study. (2 points) A. True B. False
- According to the readings/lecture, what is considered to be one of the most important problems in observational epidemiologic studies? (2 points) A. Issues with external generalizability B. Confounding C. Interviewer bias D. Random error E. Effect measure modification
- Under what type of bias is confounding classified? (2 points) A. Selection Bias B. Information Bias C. Random Error D. Systematic Error
- A research study looking at the association between head injury and hearing loss among construction workers and office workers resulted in the following data. (2 points) Crude OR: 0. - OR among those who are construction workers: 0. - OR among those who are office workers: 0. This is an example of _____ A. Confounding B. Effect Modification C. Selection bias D. Unable to determine with the information provided
Young and older adults were recruited for a study aimed at evaluating the association between sun exposure and skin cancer. For questions 12 & 13, indicate which method is being used to control for confounding in the following scenarios. In each scenario, “sun exposure” is the exposure, “skin cancer” is the outcome/disease, and “age” is the confounder.
- During the analysis, the researchers limit their study to assess the association between sun exposure and skin cancer in young adults only. What method are the researchers using to control for confounding? (2 points) A. Restriction B. Stratification C. Adjustment (statistical) D. Matching E. Randomization
- The researchers then determine the relative risk separately for young and older adults and compare these to the crude relative risk. What method are the researchers using to control for confounding? (2 points) A. Restriction B. Stratification C. Adjustment (statistical) D. Matching E. Randomization Use the information below to answer the following questions. In a case-control study, the researchers studied the association between calorie restriction and type 2 diabetes. The study included 2700 cases and 3040 controls, and the following results were observed: Type 2 diabetes No diabetes Total Calorie restriction Yes 1230 2030 3260 No 1470 1010 2480 Total 2700 3040 5740
- Calculate the crude odds ratio for the association between calorie restriction and type 2 diabetes, based on the data above. Round your final answer to two decimal places. ( points) A. 0. B. 0. C. 1. D. 0.
Use the following information to answer questions 15-17. In the case-control study described above, age was investigated as a potential confounder or effect modifier and thus, the data were stratified by younger adults/older adults. Based on the stratified data in the following table, calculate the association between calorie restriction and type 2 diabetes separately for young and old adults. Younger adult Older adult Cases Controls Cases Controls Calorie restriction Yes 770 450 460 1480 No 1140 550 330 560 Total 1910 1000 790 2040
- Calculate the OR for younger adults. Round your final answer to two decimal places. ( points) A. 0. B. 0. C. 0. D. 0.
- Calculate the OR for older adults. Round your final answer to two decimal places. ( points) A. 0. B. 3. C. 0. D. 0.
- Use the data provided in the stratified tables to calculate a Mantel-Haenszel adjusted Odds Ratio. Remember to keep 5 digits for your intermediate variables. Round your final answer to two decimal places (4 points). .
- Using the table below and for each of the 4 datasets, state whether there is evidence of confounding, effect modification, or neither (choose only one). (8 points) Data Set Crude OR Stratum 1 OR Stratum 2 OR Answer: Data Set 1 1.10 0.40 2.20 Effect Measure Modifer Data Set 2 0.51 0.49 0.52 Nothing Data Set 3 2.72 1.61 1.64 Confounder Data Set 4 1.90 1.21 1.22 Confounder
- _Negative (positive/negative) confounding underestimates the strength of the relationship, biasing the estimate _Towards (away from /towards) the null while _Positive (positive/negative) confounding overestimates the strength of the relationship biasing the estimate _Away from (away from/towards) the null. ( points)
- A study was conducted to understand the relationship between contact lens use and the risk of eye ulcers. The crude relative risk was equal to 0.7 and the age-adjusted relative risk was equal to 0.45. What is the role of age in this study after adjusting for it? ( points) A. Age is the exposure under study B. Age is the outcome under study C. Age is a 3rd^ variable with no effect on the contact lens use—eye ulcer association D. Age is a positive confounder E. Age is a negative confounder
- Researchers performed a study to examine the relationship between breast cancer incidence in women and number of offsprings. Because they believed that age could be a confounder, participants were stratified into two age categories: <40 years old and 40+ years old, but the researchers suspect that residual confounding with age is still present. What is their best option for removing most or all of the residual confounding due to age? (2 points) a. Restrict their analysis to women with breast cancer who have had at least one child. b. Stratify their age data using more age categories so the age ranges of each category are smaller. c. Exclude age from their analysis.
d. Adjust their data for other demographic variables, such as race/ethnicity and socio- economic status. As you might already know, most working epidemiologists do not calculate measures of association or Mantel-Haenszel adjusted estimates by hand. Rather, statistical software like Stata, R or SAS is used to perform these calculations. Stata, in particular, was designed with epidemiology in mind and has several functions to create tables for epidemiologists, including functions that calculate the crude, stratum-specific and adjusted measures of association in a single output. Use the Stata output in the graphic below to help you understand how Stata output can be used to determine if confounding, effect modification, or neither is present. After reviewing the explanations in the figure below and fully understand how to interpret Stata output, proceed to answer the following questions.
- A case-control study was conducted to assess the association between drinking status (current drinker vs not a current drinker) (variable name: drinker) and heart disease (variable name: CHD). As part of their analysis, investigators considered the effect of several other variables including biological sex. Use the Stata output below to determine if biological sex (variable name: sex) is a confounder, effect modifier, or neither. . cc chd drinker, by(sex) Sex | Odds ratio [95% Conf. Interval] -----------------+--------------------------------------- Female | .4622825 .3276208.
In this dataset, the variable age is … (2 points) A. A Positive Confounder B. A Negative Confounder C. An Effect Measure Modifier D. Nothing but a 3rd^ variable that has no effect on the exposure--outcome association
- Provide a rationale for the answer you selected above. (3 points) Variable age is similar to crude rate. Read the assigned article by Epstein et al. (2015) and answer questions 31-37.
- What is the exposure? (2 points) A. Age B. Dietary Patterns C. Classic Hodgkin lymphoma (cHL) D. Epstein-Barr virus E. Sex F. None of the above
- What is the outcome? (2 points) A. Age B. Dietary Patterns C. Classic Hodgkin’s lymphoma (cHL) D. Epstein-Barr virus E. Sex F. None of the above
- What potential confounder/effect modifier did the authors stratify by in the primary analysis (Factor X)? (2 points) A. Age B. Dietary Patterns C. Classic Hodgkin lymphoma (cHL)
D. Epstein-Barr virus E. Sex F. None of the above
- Based on the information provided below, calculate the crude OR. Keep 5 decimal places for intermediate steps and round your final answer to two decimal places. (2 points) Cases Controls Total E+ (Quartile 3 ) 120 130 250 E- (Quartile 1) 92 157 249 Total 212 287 499 1.
- Below are tables of the study comparing Quartile 3 and Quartile 1 stratified by Factor X for those with desserts/sweets diet. Using the data provided from the paper, calculate the stratum specific ORs. Round your final answer to two decimal places. (4 points) Factor X Absent Cases Controls Total Factor X Present Cases Controls Total E+ (Quartile 3) 101 97 198 E+ (Quartile 3) 19 33 52 E- (Quartile 1) 80 135 215 E- (Quartile 1) 12 22 34 Total 181 232 413 Total 31 55 86 Factor X Absent: ______1.76___________ Factor X Present: ______1.06_______________
- Based on your calculations for questions 34 & 35 (rounded to 2 decimal places), when comparing Quartile 3 to Quartile 1, is Factor X a confounder or an effect modifier? ( points) A. Factor X is a confounder. B. Factor X is an effect modifier C. Both A & B. D. None of the above.
- Justify your answer for Q36. Check all that apply. (4 points) ☐ There is not enough information to determine what Factor X is. ☐ The stratum-specific estimates are different from each other. ☐ The stratum-specific estimates are different from the crude OR. ☐ The stratum-specific estimates are similar to each other.
