Homework Assignment #5: chance, bias, and confounding in an experimental study

Homework Assignment #5: chance, bias, and confounding in an experimental study

Part 1: Match the following methods for minimizing chance, bias, and confounding in an experimental study. (1 point each)

  1. Blind
  2. Randomization
  3. Increase sample size
  1. Increase sample size Chance
  2. Blinding                         Bias
  3. Randomization             Confounding

 

 

Part 2: For each of the following statements, indicated whether the statistical association is likely a result of chance, bias or confounding and provide a brief explanation. (2 points each)

  • A case-control study found a positive association between self-reported chest radiographs during pregnancy and breast cancer.

Likely Cause: Confounding

Brief Explanation:various studiesconducted have not found any positive association between diagnostic chest x-rays or radiographs and breast cancer risk before age 50 years. Therefore the positive association obtained between self-reported chest radiographs during pregnancy and breast cancer can only be attributable to other factors and not self-reported chest radiographs.

 

 

  • A case-control study showed a strong association exists between birth order and Down syndrome.

Likely Cause:  Confounding

Brief Explanation: the strong association does not mean that we can conclude that birth order causes Down’s. The relationship between birth order and Down’s is confounded by the mother’s age.  Older women are more likely to have children with Down’s.  Older women are also likely to be having children who are late in the birth order (it’s a lot more common for a 35-year-old woman than for a 21-year-old woman to have 6 children).  Mother’s age confounds the association between birth order and Down’s syndrome: it looks like there is an association when there is not.

  • A randomized clinical study found that participation in a health education program to reduce smoking initiation versus not participation in the study did not significantly reduce smoking cessation. The risk ratio was <1 but the confidence interval was very wide and included 1. (RR= 0.35; 95% confidence interval 0.14-55.01).

Likely Cause:  Chance

Brief Explanation: Pvalues and confidential intervals are often misused to dichotomize finding into “associations” and “no associations”, but this is an overly simplistic way of describing a biologic mechanism such as smoking cessation. Common sense and clinical experience are necessary to separate a meaningful, although not statistically significant association from a statistically significant association that has no meaning. Any association will eventually have a statistically significant point estimate if the investigator keeps adding to the sample size

 

  • A hospital-based case-control study identified a strong association between oral contraceptives and blood clots. Many doctors suspected the association and hospitalized some women who used oral contraceptives for evaluation.

Likely Cause: Bias

Brief Explanation:This is because doctors  hospitalized only those women who they felt to be significant for the evaluation rather than all the adverse outcomes they had identified. That is, the doctors did not include a sample of women who were not on contraceptives

 

Part 3: Identify the following studies as either case-control, retrospective cohort, or prospective cohort (1 point each).

  1. Vietnamese Experience Study: Subjects were several thousand soldiers stationed in Vietnam from 1969-1971 and several thousand soldiers stationed in Europe form 1969-1971. In 2000, investigators initiated a study to determine and compare the death rates and prevalence of illness in both groups.

A Cohort Study: This is because subjects were enrolled on the basis of their exposure (Vietnam or Europe)

 

  1. Subjects were persons with laboratory confirmed Campylobacter jejuni infection and one healthy friend of each. All subjects were asked about their consumption of meat, poultry and milk products.

Case control study because subjects were enrolled on the basis of whether they had Campylobacter jejuni infection(cases) or not (controls)

 

  1. In 1948, researchers recruited 5,209 men and women between the ages of 30 and 62 from the town of Framingham, Massachusetts, and began the first round of extensive physical examinations and lifestyle interviews that they would later analyze for common patterns related to CVD development. Since 1948, the subjects have continued to return to the study every two years for a detailed medical history, physical examination, and laboratory tests.

Experimental study because the investigators rather than the subjects themselves controlled the exposure

 

 

Part 4: For each of the following list whether the risk ratio will be equal to one, greater than one, or less than one (1 point each).

  1. Exposure is associated with decreased risk of disease: Less Than One
  2. Exposure is associated with increased risk of disease: Greater Than One
  3. There is no association between exposure and disease: Equal To One

 

 

Part 5: Read each of the following studies, construct the appropriate 2×2 table, calculate the measure of association, state the study design, and state the interpretation of the results.(4 Points for each scenario.)

  • An investigator wanted to determine whether playing video games constantly (chronic videogamer: CVG) was associated with carpal tunnel syndrome (CTS). By interviewing males ages 15-30 in a Midwestern city, the researchers located 550 males with CTS. A total of 600 additional males without CTS were selected at random and interviewed. Of those with CST, 84 reported not being a CVG. A total of 365 control males reported being a CVG.

(1) Identify what type of study this is:           Case Control Study

(2) Construct the appropriate 2×2 table regarding chronic videogaming and carpal tunnel syndrome

 

Males were CVG Were not CVG Total Incidence of CVG
With CTS 466 84 550 466/550
Without CTS 365 235 600 365/600

 

 

(3) Calculate the odds ratio between chronic videogaming and carpal tunnel syndrome

 

Relative risk   = (466/550)/ (365/600)

                        =0.8472/0.6083

                        =1.3927

 

(4) Interpret the results.

The results denote a computation of a relative risk of 1.3927. this RR is greater than 1 which means that the chances of a person suffering from CTS being CVG is higher than than the likelihood of an individual not suffering from CTS being CVG.

 

 

 

  • A group of 1,000 female college seniors from a large Midwestern university were followed for 20 years to investigate whether the use of oral contraceptive increases women’s risk of developing blood clots. Of these females, 550 were oral contraceptive users and at the end of 20 years, 48 had developed blood clots. Among those not using oral contraceptives 17 developed blood clots. None of the females had clots when the study began.

(1) Identify what type of study this is:           Cohort study

(2) Construct the appropriate 2×2 table regarding the use of oral contraceptives and the development of blood clots

 

 

Number of females who: Developed clots Did not develop clots total Incidence of disease
OC Users 48 502 550 48/550
Non OC users 17 1433 1450 17/1450

 

 

(3) Calculate the risk ratio between the use of oral contraceptives and the development of blood clots

 

Relative risk   = (48/550)/ (17/1450)

                        =0.0873/0.011724

                        =7.4462

(4) Interpret the results.

Having obtained a relative risk of 7.4462 which is more than 1 denotes that the risk in users of oral contraceptives is less than the risk in non users of oral contraceptives.

 

 

 

Part 7: Identify the following features with appropriate study design:

    1. Case-control study
    2. Cohort study

 

  1. Cohort            Categorize subjects on the basis of exposure and then follow up to see if they develop the health condition we are studying.

 

  1. Case-control study                Generally used to explore rare diseases.

 

 

  1. Cohort  study             Useful for exploring several risk factors for a given outcome.

 

  1. Case control study                 After some time compare the disease rate for the exposed

with that of the unexposed.

 

  1. Cohort study                          Generally used when the exposure is rare.

 

 

  1. CohortStudy              Useful when there are several outcomes related to a given exposure.

 

  1. Case-control study    Enrolls a group with disease and an appropriate group without disease and compares their patterns of previous exposures.

 

 

Part 8: Match the sources of bias with the type of study design and the best way to minimize the bias.

 

Type of study design:                                               

  1. Case-Control
  2. Cohort

 

Way to minimize bias

  1. Obtain controls from the same work environment as cases
  2. Obtain controls from the general population.
  3. Incentives
  4. Take steps to assure that mild, clinically resolved or fatal cases are represented in the case group
  5. Refer to medical records containing exposure information if they exist

 

Study Design                          Way to minimize bias

 

33–34    Loss to follow-up                    Cohort Study                                   C

 

34–35    Recall, interviewer                  Case Control                                    E

 

36–37    Healthy worker effect             Cohort Study                                   A

 

38–39    Berkson’s bias                         Case Control                                    B

 

40–41    Neyman’s bias                         Case Control                                    D

Do you need an Original High Quality Academic Custom Essay?