27 Oct For a case-control study of agricultural pesticide exposure in relation to arthritis, eligible arthritis cases were identified from rheumatology practices and eligible controls were identified from primary care practices as any patients without arthritis

For a case-control study of agricultural pesticide exposure in relation to arthritis, eligible arthritis cases were identified from rheumatology practices and eligible controls were identified from primary care practices as any patients without arthritis. The rheumatology practices were concentrated in urban areas of the study region and were unlikely to serve agricultural worker patients, whereas the primary care practices were evenly distributed in both urban and rural parts of the study region and served agricultural workers as well as non-agricultural workers.  

 

Based only on the information given in this scenario, the situation may result in: 

 

Differential selection of exposed cases, resulting in bias away from the null

 

Differential selection of exposed controls, resulting in bias toward the null

 

Differential selection of unexposed cases, resulting in bias toward the null

 

Differential selection of unexposed controls, resulting in bias away from the null

 

 

 

Question 2

 

In a cross-sectional study of agricultural pesticide exposure in relation to arthritis, rheumatologists examined study participants to classify them as to whether they had arthritis (the outcome: yes/no). The rheumatologists were more likely to determine that the participant had arthritis when they saw from the study file that the participant worked in agriculture.

 

Based only on the information given in this scenario, the situation may result in: 

 

Selection bias

 

Non-differential misclassification of the outcome

 

Differential misclassification of the outcome

 

Recall bias

 

Question 3

The lecture used an illustration like this to visualize a problem of selection bias.

ELIGIBLE POPULATION Outcome Yes NO Exposed A B Not exposed C D SAMPLE POPULATION Outcome

Yes NO Exposed a b No exposed C d

Compared to the odds ratio from the reference population (all eligibles), an odds ratio based on the sample population would be:

 

Biased toward the null

 

Biased away from the null

 

Weaker

 

The same (unbiased)

 

Question 4

 

You conduct a case-control study exploring the effect of consuming leafy greens on colon cancer incidence. Cases are selected from a clinic that specializes in cancer treatment. Controls are selected from the gastroenterology department of a local hospital. Colon cancer status is confirmed by colonoscopy and biopsy. Leafy greens consumption is determined by a nutritional survey that asks participants about their diet over the past year.

 

What can you say about information bias in this study?

 

Survey methods for determining diet are imprecise, and the resulting misclassification always causes a biased risk estimate.

 

Overall, this study has no information bias because any error in colon cancer detection occurred equally among all screened cases and controls.

 

Information bias is likely in this study, as patients with colon cancer may be more likely to recall their diet than the control group.

 

Question 5

 

You conduct a case-control study exploring the effect of consuming leafy greens on colon cancer incidence. Cases are selected from a clinic that specializes in cancer treatment. Controls are selected from the gastroenterology department of a local hospital. Colon cancer status is confirmed by colonoscopy and biopsy. Leafy greens consumption is determined by a nutritional survey that asks participants about their diet over the past year.

 

To determine the impact of measurement error, you take a random sample of 10% of cases and controls and draw blood from these patients. You measure the levels of carotenoids in their blood as a ‘gold standard’, and compare the blood carotenoid levels to self-reported leafy greens consumption. This comparison yields the following:

 

 

 

 

 

 

 

 

 

Carotenoid Test

 

 

 

 

 

 

 

High

 

Low

 

Total

 

 

 

High

 

7

 

0

 

7

 

Self-Report

 

Low

 

3

 

40

 

43

 

 

 

Total

 

10

 

40

 

50

 

What is the SPECIFICITY of the self-report? Report this as a percent (without the percent sign), rounded to the nearest 10ths place (e.g. 92.4 or 41.0)

 

Question 6

 

You conduct a case-control study exploring the effect of consuming leafy greens on colon cancer incidence. Cases are selected from a clinic that specializes in cancer treatment. Controls are selected from the gastroenterology department of a local hospital. Colon cancer status is confirmed by colonoscopy and biopsy. Leafy greens consumption is determined by a nutritional survey that asks participants about their diet over the past year.

 

The collected study data are presented in the 2 x 2 table, below:

 

 

 

Cases

 

Controls

 

High Leafy Greens Consumption

 

98

 

140

 

Low Leafy Greens Consumption

 

402

 

360

 

Total

 

500

 

500

 

After correcting for measurement error of leafy greens consumption, you get the following 2×2 table:

 

 

 

Cases

 

Controls

 

High Leafy Greens Consumption

 

98

 

200

 

Low Leafy Greens Consumption

 

402

 

300

 

Total

 

500

 

500

 

 

 

Without correction, the measurement error that occurred in the collected study data would result in an odds ratio that is:

 

Equal to 1.0

 

Unbiased

 

Biased away from the null

 

Biased toward the null

 

Question 7

 

As a rule, increasing sample size results in a lower chance of introducing selection bias.

 

True

 

False

 

Question 8

 

When collecting data on past exposures via questionnaire, a tendency for cancer patients to remember exposure with greater accuracy than healthy controls is an example of: 

 

Random error

 

Selection bias

 

Information bias

 

Low generalizability

 

Question 9

 

In the study cited below, the authors investigated various lifestyle and demographic factors in association with influenza and pneumococcal vaccination in a prospective cohort study. 

 

Fix J, Donneyong MM, Rapp SR, Sattari M, Snively BM, Wactawski-Wende J, Gower EW. Predictors of Influenza and Pneumococcal Vaccination Among Participants in the Women’s Health Initiative. Public Health Rep. 2022 Mar 18:333549221081817. doi: 10.1177/00333549221081817. Epub ahead of print. PMID: 35301881.

 

In the discussion section, the authors write, “Although to our knowledge this research represents the largest long-standing, prospective cohort study to examine vaccination among women aged ≥65 years, our study had some limitations. …  Reliance on self-reported vaccination status may present an additional limitation, because the accuracy of this information may be associated with race and ethnicity, education, and household income.”

 

What is the concern the authors are referring to, here?

 

Non-differential misclassification of the outcome

 

Differential misclassification of the outcome

 

Non-differential misclassification of the exposure

Differential misclassification of the exposure

Question 10

In the study cited below, the authors investigated various lifestyle and demographic factors in association with influenza and pneumococcal vaccination in a prospective cohort study. 

Fix J, Donneyong MM, Rapp SR, Sattari M, Snively BM, Wactawski-Wende J, Gower EW. Predictors of Influenza and Pneumococcal Vaccination Among Participants in the Women’s Health Initiative. Public Health Rep. 2022 Mar 18:333549221081817. doi: 10.1177/00333549221081817. Epub ahead of print. PMID: 35301881.

 

In the discussion section, the authors write, “Although to our knowledge this research represents the largest long-standing, prospective cohort study to examine vaccination among women aged ≥65 years, our study had some limitations. …  our study population was mostly non-Hispanic White, educated, and living in urban settings, which may limit the generalizability of our overall findings.”

What is the limitation the authors are referring to, here?

The results of the study lack external validity because the study was a randomized controlled trial.

An issue of selection bias in participant recruitment, in which non-Hispanic White participants were preferentially enrolled.

 

The concern that associations found in this study may be different in a population with a greater mix of different races, ethnicities, education levels, and urban/rural residence.

The results of this study lack internal validity due to measurement error of demographic factors like race, Hispanic ethnicity, and education.