These estimates include the influence of other extraneous variables, such as confounders. Confounding is often considered a type of bias, but it is a real relationship that requires an adjustment in the study design or analysis. Understanding how to identify confounding is important as most associations have multiple causal factors. Recognizing if a study adjusted for the appropriate confounding variables is important to determine the validity of the association. To assist your proficiency with the concept of confounding, and how it ultimately affects public health, this practice assignment has been provided.

Complete Problems 1 to 4 from the “Multicausality: Confounding – Assignment” by Schoenbach, located in your Topic Materials. Check your answers against the solutions presented in the “Multicausality: Confounding – Assignment Solutions” Topic Material.

While APA style is not required for the body of this assignment, solid academic writing is expected, and documentation of sources should be presented using APA formatting guidelines, which can be found in the APA Style Guide, located in the Student Success Center.

You are not required to submit this assignment to LopesWrite.

  • © Victor J. Schoenbach Multicausality: Confounding – Assignment – 373 rev. 6/9/96, 9/6/1999, 3/30/2001

    Multicausality: Confounding – Assignment

    1. Some years ago several studies were published showing an association between reserpine (a drug used to lower blood pressure) and breast cancer in women. Since obesity is associated both with breast cancer and with hypertension (elevated blood pressure), the suspicion arose that the association between reserpine and breast cancer could be secondary to the effect of obesity. Assume that a cohort study had been conducted to address this question and produced the following data:

    Annual age-adjusted incidence of breast cancer per 100,000 women by body weight and reserpine status

    Reserpine use Yes No Total

    Obese 12.50 8.30 8.72 Not Obese 6.40 4.10 4.22 Total 10.47 6.14

    Answer the following questions on the basis of the above data (ignore considerations of statistical significance and precision). For each answer cite the most relevant figures from the table, allowing for the possibility that one factor affects the observed relation between the other factor and breast cancer risk.

    a. Is reserpine a risk factor for breast cancer?

    b. Is obesity a risk factor for breast cancer?

    c. Is reserpine use associated with obesity?

    d. Is the association between reserpine and breast cancer attributable to obesity?



    _____________________________________________________________________________________________ © Victor J. Schoenbach Multicausality: Confounding – Assignment – 374 rev. 6/9/96, 9/6/1999, 3/30/2001

    2. A 20-year retrospective cohort study of the incidence of chronic obstructive pulmonary disease (COPD) was performed in two occupational cohorts with different levels of S02, copper smelters (high SO2) and truck maintenance workers (low SO2). In 1961, when the cohort was defined, 55% of the smelter workers and 55% of the truck shop workers were smokers. The relative risk for COPD due to smoking was 10.5 among the smelters and 3.0 among the truck shop workers. Pulmonary function data taken in 1980 showed that 75% of the smelter workers had low FEV1 values (<90% predicted) and 33% of the truck shop workers had low FEV1 values. COPD and low FEV1 were strongly associated in each cohort. [FEV1 is forced expiratory volume in one second.]

    a. In the above study, is smoking a likely confounder of the association between COPD and SO2 exposure (i.e., in smelters vs. truck shop workers)? Briefly discuss (1-3 sentences).

    b. The best reason for not controlling for low FEV1 as a potential confounder is:

    A. Low FEV1 is not associated with SO2 exposure according to the data.

    B. Low FEV1 is not associated with COPD according to the data.

    C. Low FEV1 is not an independent risk factor for COPD.

    D. Low FEV1 is not associated with smoking according to the data.

    3. Diagrammed below are two possible causal models involving oral contraceptive use (OC), plasma homocysteine level (HCS) and myocardial infarction (MI). Briefly discuss the implications of the two models with respect to whether HCS would need to be considered as a potential confounder of the relationship between OC and MI.

    OC OC


    Other factors

    Other factors

    [arrows show hypothesized causal pathways]



    _____________________________________________________________________________________________ © Victor J. Schoenbach Multicausality: Confounding – Assignment – 375 rev. 6/9/96, 9/6/1999, 3/30/2001

    4. The following table, published in the Oxford Family Planning Association Contraceptive Study (Vessey et al.), shows characteristics of individuals at the time of recruitment in to the study. Based on the data presented in the table, discuss three potential sources of bias apparent from the characteristics of the three contraceptive groups. How would these factors be expected to influence the appearance of a causal association between oral contraceptive use and circulatory deaths if no adjustment for the factors were carried out?

    Some characteristics of subjects in the three contraceptive groups at time of recruitment

    Method of Contraception in use on Admission

    Characteristic Oral Diaphragm IUD Percentage aged 25-29 years 56 35 35 Percentage in Social Classes I or II* 39 49 34 Percentage smoking 15 or more cig./day 17 7 12 Mean Quetelet’s Index** 2.25 2.26 2.31 Percentage*** with history of:

    Hypertension 0.91 0.67 0.50 Pre-eclamptic toxaemia 12.58 16.26 16.07 Stroke 0.03 0.04 0.30 Rheumatic fever 0.76 0.66 1.04 Rheumatic heart disease 0.09 0.26 0.32 Congenital heart disease 0.12 0.31 0.16 Venous thromboembolism 0.87 4.30 7.96

    * Registrar General’s classification [Social Class I is highest]

    ** Weight (g) / height (cm)2. *** Standardized by indirect method for age and parity. See Vessey, et al.

    5. The following questions (from the 1985 EPID 168 second midterm exam) are based on data from Kantor AF, Hartge P, Hoover RN, et al. Urinary tract infection and risk of bladder cancer. Am J Epidemiol 1984;119:510-5). In that study, 2982 newly-diagnosed bladder carcinoma patients identified through the U.S. National Cancer Institute SEER (Surveillance, Epidemiology and End Results) Program during a one-year period beginning in December 1977 were interviewed. 5782 population controls from the same geographic areas covered by SEER were selected using an age- and sex-stratified random sample of the general populations, with 2:1 frequency-matching of controls to cases. Information on physician-diagnosed urinary tract infections (UTI) more than one year before interview (and many other factors) was obtained through personal interviews using structured questionnaires in respondents’ homes. The following data are from Table 1 in Kantor, Hartge, Hoover et al.:



    _____________________________________________________________________________________________ © Victor J. Schoenbach Multicausality: Confounding – Assignment – 376 rev. 6/9/96, 9/6/1999, 3/30/2001

    Table 1 Relative risks (RR) of bladder cancer associated with

    history of urinary tract infection,* by number of infections;

    10 geographic areas of the United States, 1978

    Males Females

    No. of urinary tract infections

    Cases Controls RR 95%

    confidence interval

    Cases Controls RR 95%

    confidence interval

    0 1758 3642 1.0+ 398 979 1.0+ 1 or 2 309 423 1.5 (1.3-1.8) 176 296 1.2 (0.9-1.5)

    3+ 146 152 2.0 (1.6-2.6) 145 206 2.1 (1.6-2.7)

    * Maximum likelihood estimate of relative risk adjusted for race, age, smoking status (never smoked, ex-smoker, current smoker) from stratified analyses.

    + Reference category

    a. The sex-specific relative risks for bladder cancer (BC) shown in Table 1 are adjusted for race, age, and smoking status. Which one of the following could be the relative risk of BC risk for 3+ urinary tract infections (UTI) adjusted for race, age, smoking status, and gender? [Choose one best answer]

    A. 1.33

    B. 1.92

    C. 2.05

    D. None of the above.

    b. Using the data in Table 1, construct a labeled 2 by 2 table for estimating the crude (with respect to race, age, smoking status, and gender) relative risk for BC in men and women who have a history of 3+ UTI compared to men and women with no history of UTI. Your answer should show the correct formula and substitution.

    c. Is gender associated with history of 3+ UTI? Support your answer with the most relevant numbers from Table 1.

    d. Is gender a confounder of the association between BC risk and history of 3+ UTI? Support your answer with data from Table 1 and/or your answers to the two preceding questions._

Do you need a similar assignment done for you from scratch? We have qualified writers to help you. We assure you an A+ quality paper that is free from plagiarism. Order now for an Amazing Discount!
Use Discount Code "Newclient" for a 15% Discount!

NB: We do not resell papers. Upon ordering, we do an original paper exclusively for you.