Cofounding and Effect Modification Jeopardy!

Cofounding

100

What does confounding mean?

Confounding occurs when the relationship between two variables is explained by each of their association with a third variable. (For the answer to this question and more information, see slide 2 of Confounding & Effect Modification on Blackboard)

200

What are the three criterias a variable must meet in order to be considered a confounder?

• The confounder must be associated with the exposure and must be an independent risk factor for the disease. • A confounder cannot result from the exposure or be in the causal pathway. • The confounder must be associated with the outcome (To find this answer you have to go to slides 3-4 on Confounding & Effect Modification)

300

What are three ways investigators can minimize confounding?

• Match for the potential confounding variable • Adjust for confounder during analysis • Restrict study to population group lacking confounding factor (To find this answer please see slides 15 on the Confounding & Effect Modification slides)

400

• The confounder must be associated with the exposure and must be an independent risk factor for the disease. • A confounder cannot result from the exposure or be in the causal pathway. • The confounder must be associated with the outcome

What are the three criterias a variable must meet in order to be considered a confounder? (To find this answer you have to go to slides 3-4 on Confounding & Effect Modification)

500

The first step to assess confounding is that it must be associated with the exposure and must be a(n) ___________________ for the disease:

Independent risk factor (For the answer to this question and more information, see slide 3 in notes)

Effect Modification

100

When is effect modification present?

When the exposure-disease relationship varies across levels of another factor. (For more information about this answer please refer to slide 16)

200

Think of an example of effect modification in relation to a smoker, the likelihood of them getting cancer, and asbestos.

If a person is a smoker, they are likely to get cancer. If the person is a smoker and is exposed to asbestos they they are more lily to get cancer. The third factor and effective modified is the asbestos. (For more information please refer to slide 16-17)

300

What are the ways of assessing effective modification?

• We can use stratified analysis • test for effect modification prior to testing for confounding • If the variable is concluded to be effect modifier, then the question of confounding by variable is unsettled. (For the answer to this question and more information, see slide 16 in notes)

400

True or false? If a variable is identified as an effect modifier, we do not worry about cofounding and directly present the stratum specific results. We don't have to worry about finding the adjusted measure of association.

True (For more information please refer to the brochure and slide 16)

500

Reye's syndrome is a severe condition characterized by the sudden development of brain damage and liver dysfunction after a viral illness. The syndrome is most commonly seen in children between the ages of 4-14 who have been treated with aspirin while recovering from a viral illness, most commonly chickenpox or influenza. Fortunately, Reye's syndrome has become very uncommon since aspirin is no longer recommended for routine use in children. While Reye's syndrome can occur in adults, it is distinctly more common in children. Thus, the effect of aspirin treatment for a viral illness is very clearly modified by age. In this situation if you were to calculate effect modification using an overall estimate of the data would the result be applicable?

No, in this situation, computing an overall estimate of association is misleading.One common way of dealing with effect modification is examine the association separately for each level of the third variable. For example, if one were to calculate the odds ratio for the association between aspirin treatment during a viral infection and development of Reye's syndrome, the odds ratio would be substantially greater in children than in adults. As another example, suppose a clinical trial is conducted and the drug is shown to result in a statistically significant reduction in total cholesterol. (For more information, please refer to slides 16-18 and the following link: http://sphweb.bumc.bu.edu/otlt/MPH-Modules/BS/BS704-EP713_Confounding-EM/BS704-EP713_Confounding-EM_print.html)

Math Problems

100

If your teacher proclaims that teenagers that eat daily meals with their family are 20 times more likely to graduate college than those who do not, what is the relative risk?

-20 (To find other examples in that has to deal with relative risk please see slides 10-11)

200

Lung cancer Drinking status Yes No Total Heavy Drinker 35 1500 1535 Non- Drinker 23 2100 2123 Are those with lung cancer more likely than non smokers to drink alcoholic beverages? What is the relative risk?

• RR=incident rate in exposed/ incident rate in unexposed • 35/1535=.0228 for exposed • 23/2123=.0108 for unexposed • .0228/.0108=2.11 (For examples and more information on this question, see slides 5, 10, and 11)

300

If statistics show that teenagers who grow up in neighborhoods with high crime rates are 52 times more likely to end up becoming criminals compared to teenagers who grow up in safe neighborhoods, what is the relative risk?

-52 (To find other examples in that has to deal with relative risk please see slides 10-11)

400

Diabetes Sugar Levels Yes No Total High Blood Sugar 120 1532 1652 Normal Blood Sugar 143 2432 2575 Are those with high blood sugar more than those with normal blood sugars to prone to get diabetes? What is the relative risk?

• RR=incident rate in exposed/ incident rate in unexposed • 120/1532=.0783 for exposed • 143/2432=.0588 for unexposed • .0783/.0588=1.33 (For examples and more information on this question, see slides 5, 10, and 11)

500

Hospitalized Not hospitalized Total Male 1330 7018 8348 Female 798 6400 7198 Calculate the Risk Ratio using the following data.

Risk ratio= 1.44 (For examples and more information on this question, see slides 5, 10, and 11)

Formulas/Rates

100

What is the formula for adjusted relative risk?

Incidence rate in exposed group/Incidence rate in unexposed group (For more information please refer to slide 14)

200

What is the formula for relative risk?

Incidence rate in exposed/Incidence rate in unexposed (For more information please refer to slide 11)

300

What is crude rate? How is it expressed?

Rate that exclusively accounts for two factors for a population. It is expressed as a ration. (For more information please refer to brochure)

400

What is adjusted measurment?

Includes factors such as age, gender, or any other confounding factor that should be taken into account in order to get a more specific ratio as the result (For more information, please refer to brochure)

500

When is the adjusted odds ratio used?

This ratio is used when there is no significant difference between the two groups is distinctly identified. Preforming an adjusted odds ratio would cause you to lose the meaning of what it is that is making the two distinct groups so different. (For more information please refer to slide 18)

Random

100

Blake is working on a project to measure the rate of obesity in the entire population of St. Louis, Missouri. This is an example of _______

Crude Rate (For the answer to this question and more information please see the link:http://www.healthknowledge.org.uk/public-health-textbook/research-methods/1a-epidemiology/confounding-interactions-methods Search “crude” in document for location of information.)

200

If Blake then takes into account the age of the population, and creates subgroups, the new rate would be an example of _____

Adjusted Rate (For examples of this question and more information, see slide 18 of the notes)

300

People who drink more coffee also tend to smoke more and people who smoke are at an increased risk for pancreatic cancer. This is an example of confounding or effect modification?

Confounding (In looking for the difference between effect modification and confounding please see slides 2 & 16)

400

A confounder cannot result from the __________ nor be in the causal pathway of the disease:

Exposure (For answer to this question and more information, see slides 2-4)

500

What is the difference between effect modification and confounding?

• Confounding is more of the distortions of the true exposure/disease relationship by another factor. Bias is formed that must be controlled throughout your analysis • Effect modification desires observation, you have to identify subgroups that will result in a higher or lower effect. • Confounding is a distortion of the true exposure/disease relationship by another factor. It is a form of bias that must be controlled for in your analysis. • Effect modification is a desired observation, it means that you have identified subgroups for which there is a higher or lower effect (For difference between the two, see slides 2-10 on confounding and 16-18 on effect modification.)