Confounding

Example: Study reports modest alcohol consumption lowers risk of heart disease. Alcohol increases HDL, which decrease the risk of heart disease. Is HDL a confounder? How does alcohol exert its effect on heart disease?

(3) Intermediate step in causal pathway: examine biological pathway by which the exposure is thought to impact the disease Moderate alc consumption reduces risk of CVD by increasing the HDL HDL is an intermediate step (not a confounder)

Magnitude of Confounding: formula

(Crude RR - Adjusted RR)/Adjusted RR 10-20% significant

Identifying potential confounders in study design phase

-Literature review -To ascertain all known risk factors for the disease

Method to Control Confounding

-Prevention strategies -Data analysis strategies

Of the categories of systematic error: selection bias, information bias and confounding bias, ...........

.....only confounding may be controlled for in the data analysis!!!

The ideal: When you design a study, attempt to have exposed and unexposed groups comparable in every way - except the exposure under study....

...with confounding, a risk factor apart from the exposure under study is distributed differently between the exposed and unexposed groups

Criteria of Confounders: must satisfy all 4 criteria

1. Be a risk factor or marker for the disease 2. Be associated with the exposure 3. Not be an intermediate step in the causal path between exposure and disease (then it wouldn't be independent). 4. Must be present to a greater or lesser extent in the exposed (or treatment group) versus the comparison (control or placebo) group.

Confounding: Steps 1 & 2

1. Confounder and Disease: Is there an independent association between age (possible confounder) and dementia (outcome)? 2. Confounder and Exposure: Is there an independent association between: age (confounder) & diabetes (exposure)?

Analysis Strategies to Control Confounding: Stratification Example: Case control study of DDE (metabolite by-product of pesticide, DDT) exposure and breast cancer (hypothetical) Crude OR (without age stratification): 1.9

1. Each stratum (< 50 years and 50 years & older) restricts analysis to a narrow range of confounder, in this case, age 2. When stratified, no association between breast cancer and DDE among women in either stratum: < 50 and women 50 years and older. OR = 1.0 for each age category The appreciable difference between the crude odds ration (O.R=1.9) and stratum specific odds ratios (O.R=1.0) indicates that...confounding by age is present

Method to Control Confounding: Prevention strategies: Matching: Example Cohort study on exercise and risk of colon cancer

1. Lit review: potential confounders: age, gender, obesity 2. Exposed subject (exerciser) enrolls, 55 yo male, normal BMI 3. Unexposed subject (non exerciser), 55 yo, male, normal BMI

Analysis Strategies to Control Confounding: 3 types

1. Standardization 2. Stratification 3. Multivariate techniques

Ecological studies: issues

1. ecological fallacy 2. very susceptible to confounding

Example: Magnitude of confounding

= Crude RR - Adjusted RR Adjusted RR = (3.5) - (2.0) (2.0) = 75% Large amount of confounding

Confounding variable

A confounding variable is independently associated with both the risk factor (exposure) and the disease (outcome). Because of the dual association, the confounding variable creates a false association between the risk factor and disease or can exaggerate or hide a true association.

Confounding: basic definition

A mixing of effects between the exposure, the outcome and a third extraneous variable known as a confounder.

Analysis Strategies to Control Confounding: Stratification: Advantages and Disadvantages

Advantages 1. Performing analyses within strata is a direct and logical strategy. 2. The computational procedure is straightforward. 3. Allows epidemiologists to view the raw data Disadvantages 1. Small numbers of observations in some strata. (example - next slide) 2. Difficulty in interpretation when several confounding factors must be evaluated. 3. Cannot control for many variables simultaneously

Analysis Strategies to Control Confounding: Multivariate techniques: Advantages and Disadvantages

Advantages: Allow for simultaneous control of several exposure variables in a single analysis. Disadvantages: Potential for misuse.

Method to Control Confounding: Prevention strategies: Randomization: Advantages and Disadvantages

Advantages: -Convenient, inexpensive; permits straightforward data analysis. -Controls for known and unknown confounders Disadvantages: -Need large sample sizes. -Can only be used in experimental designs (not in observational studies)

Method to Control Confounding: Prevention strategies: Restriction: Advantages and Disadvantages

Advantages: -Effective, simple, convenient, inexpensive -Provides complete control of known confounder Disadvantages -Unlike randomization, cannot control for unknown confounders. -Difficult to recruit enough subjects. -Limits generalizability of study findings -You can only apply your data to the study population (white males)

Method to Control Confounding: Prevention strategies: Matching: Advantages

Advantages: -Fewer subjects are required than in unmatched studies of the same hypothesis. -Useful in small case control studies -When confounder is a complex nominal variable - occupation, neighborhood (associated with complex web of environmental variables)

Example: Risk of dementia among adults with diabetes: Confounding factors

Age: Subjects with diabetes were, on average, older than those without diabetes When confounding for age was controlled: RR=2.0 (Initial result RR = 3.5 was exaggerated by confounding by age)

Assessing Confounding: "appreciable" difference

Arbitrary. Commonly > 10% - 20% difference is considered appreciable

Method to Control Confounding: Prevention strategies: Randomization

Attempts to ensure equal distributions of the confounding variable in each exposure category (i.e. in each arm of an experimental design)

Steps in Assessing Confounding (3)

Confounding is a quantitative issue (Skipped) Is confounding present? What is the magnitude ? What is the direction?

Residual Confounding

Confounding that remains even after many confounding variables have been controlled 1. Confounder for which data was not collected. 2. Differences in risk within a category of the confounder. (e.g. too broad an age group)

Variables (SKIPPED)

Dependent Variable: an outcome (disease) we seek to explain or account for by the influence of the independent variable(s) Independent Variable: a risk factor or exposure

Methods to Control for Confounding in Design and Analysis

Design Stage: Randomization (less likely that everyone in group will have confounding factor?), Restriction, Matching Analysis Stage: Standardization, Stratified analysis, Multivariate analysis

Method to Control Confounding: Prevention strategies: Disadvantages

Disadvantages: -Costly because extensive searching and recordkeeping are required to find matches. -When one matches subjects on a potential confounder that particular exposure variable can no longer be evaluated with respect to its contribution to risk. (Not possible to study the relationship between the matching factor and outcome in case control) -Overmatch: when match on several factors e.g. age, race, gender, income, neighborhood, may make study groups that they are overmatched. The exposure becomes so similar for the two groups that association cannot be identified

Of all studies, ??????? are most susceptible to confounding

Ecological studies because it's more difficult to control for confounders at the aggregate level of data E.g. fat consumption and breast cancer (age, body weight, hormone therapy)

Decision Tree to determine if a variable is a confounder

Evaluate association between confounder and disease -> 1. association is present: Confounding POSSIBLE -> 2. Evaluate association between confounder and exposure -> Association is present: Confounding is PRESENT; OR Association Absent: NO Confounding. Evaluate association between confounder and disease -> 1. Association is absent: NO CONFOUNDING

Assessing Confounding: Direction of confounding

Exaggerate true association = POSITIVE Hide a true association = NEGATIVE

Analysis Strategies to Control Confounding: Standardization (SKIPPED)

Example - death rates in Florida and Alaska Crude mortality rates for: Florida: 1060.94/100,00 Alaska: 417.91/100,000 Excess crude mortality: 643.03/100,000 Table 3-7

Example: Potential confounders in study of chemical contamination of drinking water and risk of breast cancer. What would you like to determine from the literature review?

Known risk factors for breast cancer: family history of breast cancer, race, religion, age at first delivery, hx of radiation rx...

Method to Control Confounding: Prevention strategies: Matching

Matches subjects in the study groups according to the value of the suspected or known confounding variable to ensure equal distributions. Example - by 5 or 10 year age strata

Random Error (Types)

Poor Precision. Sampling Error Variability in Measurement

Method to Control Confounding: Prevention strategies Design Stage

Prevention strategies: attempt to control confounding through the study design itself Three types of prevention strategies: Randomization, Restriction, Matching

Method to Control Confounding: Prevention strategies: Restriction

Study admission criteria are limited. Entrance into the study is confined to individuals who fall within a specified category of the confounder e.g. age, gender, race For example, restricting participants to a narrow age category can eliminate age as a confounder.

Confounding: epi definition

The distortion of an association between an exposure and an outcome because of the influence of a third variable that was not considered in the study design nor initial analysis.

Analysis Strategies to Control Confounding: Stratification

The result of separating a sample into several subsamples according to specified criteria such as age groups, SES ... analyses performed within each stratum.

Analysis Strategies to Control Confounding: Multivariate techniques: Definitition

To control for many confounding variables simultaneously. To construct mathematical models that describe simultaneously the influence of exposure and other factors that may be confounding the effect.

Example: Study in LA reports air pollution is associated with bronchitis Possible confounders ?

Urban crowding, high population density, smoking?

Confounding: impact

When uncontrolled the effects of a confounding variable cannot be distinguished from the study exposure