"Torture the data, and it will confess to anything.” - Ronald Coase
bruno-nascimento-PHIgYUGQPvU-unsplash.jpg

Who worries about their health?

 

Who worries about their health?

 

Heart disease is a prevalent issue in the United States. By 2010, about 600,000 died of heart disease; one in every four deaths is caused by heart disease. To make matters worse, the prevalence for heart diseases increases for those with a family history of heart disease.

Those who have a family history of heart disease are also ones who report greater levels of worry than those without a family history of heart disease. These perceptions of a greater personal risk of heart disease cause people to take protective action. Levels of worry are related to perceived risk and, in turn, is related to health behaviors.

The goal of this analysis is to examine the connection between the presence of heart disease in family history, levels of worry, and health behaviors. I aim to measure the correlation of levels of worry and aspirin use among individuals based on a family history of heart attacks., changes in aspirin use in response to levels of worry and possible differences in effects of family members of different genders

 
 

The data set

To conduct this research, I used the Midlife in the United States (MIDUS) data set; a data set that assesses behavioral and psycho-social factors on physical and mental health.

The MIDUS data set consists of a 30 minute phone interview to participants via a Random Digit Dialing process as well as self-administered questionnaires.

Below you will see the summary of responses to three key questions:

  1. How much does an individual worry about their heart health: “not at all” , “a little” , “some” and “a lot”

  2. Do you use aspirin to prevent heart disease: “Yes” or “No”

  3. Do you have a family member with a history of heart disease: “Yes”, “No”, “Male”, “Female”, “Both”

worry picture 1.png
worry picture 2.png

the analysis

worry picture 3.png

There are two main outcomes I am focusing on in this project: level of worry and preventative behaviors (aspirin use). These outcomes (dependent variables) are predicted using the prevalence of a family history of heart disease (independent variable).

To predict the two main outcomes, I used a series of binary logistic regressions:

worry picture 4.png
worry picture 5.png
  1. Predicting level of worry (ref: A Lot) by family history

  2. Predicting aspirin use by family history

  3. Predicting aspirin use by levels of worry (ref: Not at all)


The result

worry picture 6.png

From the binary logistic regressions, it is shown that respondents with a family history of heart attack have higher odds of…

  • Worrying a lot about their health

  • Taking aspirin to prevent heart disease

In turn, respondents who worry about their heart health have higher odds of taking aspirin to prevent heart disease. It is not too far fetched to say then that family history of heart attacks is an effective predictor of the likelihood an individual will take some sort of preventive behavior, from simply using aspirin to potentially changing their lifestyles completely.