Read the following article:
Gilliam WS et al. COVID-19 transmission in US child care programs. Pediatrics 2021;147(1): e2020031971.
Part A: Critical appraisal of a new epidemiological study (Gilliam 2021) (i.e., for questions 1-9, please relate your answers to the particular methods used in the Gilliam et al study).
| Q1. (/ 2 points) What research question (RQ) were Gilliam et al trying to answer? Be sure your RQ includes an exposure, outcome, and population. | 1 sentence or bullet |
| Q2. (/ 4 points) Do you think that selective survival (or survival bias) is a concern in this study? If yes, why? If no, why not? | max. 2 sentences or bullets |
| Q3a. (/ 1 point) How was the outcome measured, meaning what was asked and how was the information collected? | max 2 sentences or bullets |
| Q3b. (/ 1 point) What is one limitation of this method of measurement (for the outcome)? | 1 sentence or bullet |
| Q3c. (/ 1 point) How may this method of measurement (for the outcome) lead to information bias? | 1 sentence or bullet |
| Q3d. (/ 1 point) Is the outcome measure a risk or a rate? | 1 word |
| Q3e. (/ 1 point) How was the exposure measured, meaning what was asked and how was the information collected? | max 2 sentences or bullets |
| Q3f. (/ 1 point) What is one limitation of this method of measurement (for the exposure)? | 1 sentence or bullet |
| Q3g. (/ 1 point) How may this method of measurement (for the exposure) lead to information bias? | 1 sentence or bullet |
| Q4. (/ 2 points) Using the odds ratio (OR) and confidence interval (CI) presented in Table 6, what conclusion would you make about the association between exposure to child care and COVID-19? Is this association statistically significant? Explain how you know. | max. 2 sentences or bullets |
For question 5, below, apply each of the 3 conceptual criteria to evaluate whether race/ethnicity should be considered a potential confounder for this RQ. For each part of the question (5a, 5b, 5c) write one sentence/bullet explaining how the criterion is met or not. The following two pieces of background information may be helpful in formulating your answers.
According to the US Bureau of Labor Statistics, Black (or African American) and Hispanic (or Latinx) people are more likely than white people to work in child care. While Black and Hispanic people make up 30% of the general population, they make up 38% of child care workers.
Black and Hispanic people in the US have also been found to “experience disproportionately higher rates of SARS-CoV-2 infection,” as indicated in a recent systematic review by Mackey et al.
| Q5a. (/ 2 points) Confounder must be an independent cause/risk factor of the outcome. | 1 sentence or bullet |
| Q5b. (/2 points) Confounder must be associated with the exposure. | 1 sentence or bullet |
| Q5c. (/2 points) Confounder must not be an intermediate step in the causal pathway. | 1 sentence or bullet |
| Q6a. ( /2 points) A letter to the editor was published in response to the Gilliam et al study. The author of this letter writes the following: “From numerous occupational studies, we know that many daycare workers take up additional, part-time work to supplement the low wages they receive as daycare workers. This means that their free time may be limited, calling to question the kinds of workers who are able to answer a survey like the one conducted by Gilliam et al.” What kind of bias is the author of this letter describing? | 1 sentence or bullet |
| Q6b. (/2 points) How could Gilliam et al have avoided this bias? | 1 sentence or bullet |
| Q7a. (/ 2 points) Interpret (i.e., explain the meaning in words) the odds ratio (OR) for child care type, presented in Table 6. | 1 sentence or bullet |
| Q7b. (/ 2 points) Is this association statistically significant? Explain how you know. | max. 2 sentences or bullets |
| Q8a. (/ 2 points) To address some of the biases in the cross-sectional design of the Gillam et al study, you have been given permission to collect data from laboratory confirmed cases of COVID-19 identified by the BC Centre for Disease Control Public Health Laboratory. You decide to do a case-control study, where the cases are lab-confirmed cases of COVID-19 who reside in British Columbia. Describe one method for sampling controls and be sure to explain how the controls are defined, and how you will find the controls and invite them to participate in the study. | max. 2 sentences or bullets |
| Q8b. (/ 2 points) Comment on the degree to which selection bias remains a concern with your sampling approach for controls, specifically commenting on whether the distribution (%) of the exposure in the controls may differ from the distribution (%) of the exposure in the source population that gave rise to the cases, and if so, how. | max. 2 sentences or bullets |
| Q9. (/ 3 points) Think of a confounding variable that was not measured in the Gilliam et al study but which you would want to measure in the case-control study. Describe this variable (including how you will measure it), and confirm that it satisfies the 3 conceptual criteria for confounding; i.e., explain how the criteria are met, using either your stated assumptions or information you can find about the relationship between the potential confounder and the exposure and outcome. | max 4 sentences or bullets |
Part B: Write your own epidemiological research question
| Q10a. (/ 2 points) Write an epidemiological research question that you would like to know the answer to. Use your imagination. You can choose any RQ, including something that came up during HSCI330, something you read in the news, or something that you are curious about. Be sure your RQ includes an exposure, outcome, and population. | 1 sentence or bullet |
| Q10b. (/ 2 points) Choose a “third variable” that might be a potential confounder of the exposure/outcome association identified in part a. Confirm that it satisfies the 3 conceptual criteria for confounding; i.e., explain how the criteria are met, using either your stated assumptions or information you can find about the relationship between the potential confounder and the exposure and outcome. | max 4 sentences or bullets |
| Q10c. (/ 2 points) Choose a “third variable” that might be a potential effect modifier of the exposure/outcome association identified in part a. This could be the same variable as used in part b, or a different variable. Explain how you would expect this third variable to modify the exposure/outcome effect (i.e., how might you predict that the effect will vary across levels of the third variable). | 1 sentence or bullet |