IIHE Education, Epi Info
IIHE  |  Education  |  Class Curriculum |  Epi Info

During this session, you will learn how to use Microsoft Access to explore data culled from the Henry Ford Corporate Data Store:

Objectives of session:

1. Residents will learn how to use EpiInfo to construct a database.
2. Residents will how to enter data into EpiInfo databases.
3. Residents will learn how to use EpiInfo to conduct epidemiological analyses.
4. Residents will explore risk behaviors in their practice populations.

Outline

1. Overview of EpiInfo
   • contrast of functionality of EpiInfo vs. ACCESS
   • uses of EpiInfo
2. Navigating the main menu of EpiInfo
   • editor
   • entry of data
   • analysis
   • stat calculator
   • importing and exporting data
3. Building a questionnaire
   • field formats
   • saving your file
4. Entering data
5. Conducting analyses

• Setting up an Epi Info database

• Use EPED text editor
• Set up a field for each variable in your study
• Variable types
  • Text field <A > <A>
  • Logical field <Y>
  • Date field <MM/DD/YY>
  • Numeric field ## ##.#
• Save file as a .QES file
• Enter data in a .REC file using ENTER

• Epi Info Analysis

• To read a data file
  • READ (file name)
  • EXAMPLE: READ hradnw97.rec
  • READ a:study.rec
• To calculate the mean of a variable
  • MEANS (numeric variable)
  • EXAMPLE: MEANS age
• To calculate the means of a variable of two or more groups
  • MEANS (numeric variable) (categorical variable)
  • EXAMPLE: MEANS age sex
• To get a frequency distribution of a variable
  • FREQ (variable) or TABLES (variable)
  • EXAMPLE: FREQ race TABLES race
• To get a cross tabulation of two or more variables
  • TABLES (variable) (variable)
  • EXAMPLE: TABLES sex race
• To select a subset fitting criteria
  • SELECT (criteria clause)
  • EXAMPLE: SELECT age < 18
  • EXAMPLE: SELECT race = "B"
  • EXAMPLE: SELECT (race = "B") and (age > 17)
  • EXAMPLE: SELECT smoker = "Y"
  • EXAMPLE: SELECT finc <> "1" (NOT EQUAL TO)
• To remove select criteria
  • SELECT
• To recode a variable
  • DEFINE (new variable) (type of variable)
  • RECODE (variable) TO (new variable) BY 10
  • EXAMPLE: DEFINE AGEGROUPS STRING RECODE AGE TO AGEGROUPS BY 10
  • OR
  • DEFINE RACEGROUP STRING
  • RECODE RACE TO RACEGROUP "C" = CAUCASIAN "B" = BLACK ELSE = OTHER
• To create a bar graph
  • BAR (variable)
  • EXAMPLE: BAR sex
• To create a pie chart
  • PIE variable
  • EXAMPLE: PIE race

Computer lab exercise - using Epi Info to analyze patient health risk appraisal data

You have access to data containing four different health risk appraisal surveys administered to populations at the Family Practice sites. A data dictionary that contains the variable names and coding schemes has also been provided. The content and method of administration for each of the datasets is as follows:

• HRADNW97.REC - 71 records from a convenience sample of patients at the Detroit Northwest Center in 1997. Most of these were collected in the waiting room.
• HRA_EJ97.REC - 346 records from a mailed survey. This was sent to a random sample of HAP adult patients assigned to East Jefferson. The survey was mailed in Dec. 96 with returns in Dec. and Jan. 97.
• HRA_M96.rec - 89 records from a waiting room survey conducted on a convenience sample of patients at the Mercy Family Practice Center in 1996.
• HRA_EJ96.rec - 89 records from a waiting room survey conducted on a convenience sample of patients at the East Jefferson Family Practice Center in 1996.

Use the data in one or more of the datasets to describe risk patterns. You may want to use more than one dataset to make comparisons between different populations.

1. Assess patient risk from smoking. Identify overall smoking patterns. Characterize current smokers. As a clinician, what sort of intervention would you plan to deal with current smokers?
2. Examine dietary habits of your surveyed group. Describe these habits in terms of healthy behaviors. Are any improvements needed? What would you recommend?
3. Examine alcohol use. Is there anyone at special risk? Analyze and describe the "CAGE" factors of drinkers.
4. Describe the prevalence of the health problems listed in the survey. Do a "crosstab" of personal history of HTN and family history of HTN. Do the same for diabetes. What are your impressions of this?
5. What proportion of the patients in your survey sample know their cholesterol level? For these patients, describe the range and mean value. Are there any patients with really alarming cholesterol levels? What would you do about this?
6. Analyze some of the more social factors associated with health (like feeling safe, exposed to violence, etc.). Profile your patients with respect to some of these variables.

 

Return to Curriculum Page