A study was performed relating baldness pattern to MI (heart...

A study was performed relating baldness pattern to MI (heart attack) among men in the Atherosclerosis Risk in Communities (ARIC) study [49]. Baldness pattern and prevalent MI were determined at the same examination during the period $1996-1998$. Baldness was categorized into 5 categories (none/frontal/mild vertex/moderate vertex/severe vertex). For this example, we focus on the comparison of severe vertex baldness to no baldness. The data in Table 13.57 were reported by age group. What is the estimated OR for Ml comparing men with severe vertex baldness vs. no baldness after controlling for age?

A study was performed relating baldness pattern to MI (heart attack) among men in the Atherosclerosis Risk in Communities (ARIC) study [49]. Baldness pattern and prevalent MI were determined at the same examination during the period $1996-1998$. Baldness was categorized into 5 categories (none/frontal/mild vertex/moderate vertex/severe vertex). For this example, we focus on the comparison of severe vertex baldness to no baldness. The data in Table 13.57 were reported by age group.
What is the estimated OR for Ml comparing men with severe vertex baldness vs. no baldness after controlling for age?

Key Concepts

Stratification

Stratification is an analytical method used to control confounding by dividing the data into homogeneous subgroups based on a variable, such as age. By calculating the odds ratio within these age strata and then combining the results, researchers can better isolate the effect of the primary exposure (severe vertex baldness) on the outcome (MI) from the influence of age.

Cross-sectional Study Design

A cross-sectional study design involves collecting data on both exposure and outcome at a single point in time. This design is useful for assessing prevalence and associations between variables, such as examining the relationship between baldness and MI. However, because exposure and outcome are measured simultaneously, causal relationships cannot be definitively established.

Odds Ratio (OR)

The odds ratio is a measure of association used in epidemiological studies to compare the odds of an outcome occurring in one group with the odds of it occurring in another group. In this context, it compares the likelihood of myocardial infarction (MI) between men with severe vertex baldness and those with no baldness. An OR greater than 1 suggests a higher odds of MI among the severe vertex baldness group compared to the reference group.

Control for Confounding

Confounding occurs when an outside variable influences both the exposure and the outcome, potentially distorting the true association. In this study, age is a confounding variable that could affect both baldness patterns and the risk of MI. Controlling for age, either through stratification or multivariable methods, helps ensure that the estimated odds ratio more accurately reflects the relationship between baldness pattern and MI, independent of age.

Transcript

00:01 I'm going to begin by writing down everything that we are told in this problem.

00:05 So all of the numbers that we are given, i'm going to write down.

00:09 So were first told there is 1,435.

00:15 Um men in this study um that's being followed by researchers were also told that this study is and research is happening over a period of five years.

00:29 And we know that we are looking at boldness and we are looking at heart disease.

00:42 And when you take it out of this study and you ask, are these two things are correlated? we already know the answer is known.

00:50 But we now need to prove why why is it not correlated? so we're ranking these the baldness on a scale from 1 to 5 and heart disease were saying yes or no.

01:04 That's how that's being raped.

01:05 So so what we are going to start with is um defining a few things.

01:14 So first off we have a bunch of data here, right? we're not data necessarily big numbers.

01:21 So we're gonna need to define something called quantitative quantitative data, quantitative data.

01:31 You've probably heard it before.

01:32 It's really any kind of data that has to do with numbers that have meaning.

01:43 So the numbers tell us something about the study such as the 1,435 people.

01:50 That's aquantive number.

01:52 It's a number informing us.

01:54 Um and it's measuring.

01:56 It's basically a way of measuring, let's put measure the opposite of quantitative data is something called qualitative data.

02:10 And qualitative data is not numbers.

02:13 Usually it's words.

02:16 And that's because it's usually used in descriptive terms and it's something that's observed rather than measured.

02:28 So anything you get with qualitative data you can't measure it, only observe it, it's describing something.

02:35 Um and there's no numbers usually.

02:37 So when you look at this data of baldness 1- five and heart disease yes or no.

02:44 Trying to determine if it's quantity for qualitative.

02:47 You see numbers right? there's a scale of 1-5 but you also see words.

02:52 So it's kind of tricky to determine which one it is.

02:55 Well that's where i'm going to introduce this term which is called categorical variables.

03:08 Categorical variables is exactly what it sounds like.

03:12 It's putting things into categories...

Please give Ace some feedback