Data Collection Tools and Creation of a Usable Database

1
Data Collection Tools and Creation of a Usable
Database

• Adam Schlichting
• University of Illinois at Chicago
• Department of Emergency Medicine

Last updated 10.18.2002
2
IntroductionWhy?

• DO NOT COPY INFORMATION DIRECTLY FROM CHARTS INTO
  A COMPUTER DATABASE!
• Decrease mistakes
• Lose charts
• Concisely store all of the data for your study
• Data sheets can be checked in seconds, charts
  take minutes to hours

3
IntroductionThe Questions

• Before designing a data collection tool, you must
  have a clear answer to the following questions
• What is it you are trying to show with your
  study?
• What is the main hypothesis?
• What is the 2nd, 3rd, 4th, hypothesis?
• What data will you need to form your conclusion?
• Think about your final manuscript and work
  backwards
• Use your abstract as a template

4
IntroductionThe Balancing Act

• Collecting data is time consuming
• Collect only what you need
• Going back to collect more data later is painful
• Collect ALL the data you need for each case

5
Data Collection ToolThe Fields

• What information do you need for every case?
• Trackable case ID
• Allows you to find case again
• MR
• Study ID
• Date
• Demographic information (who you studied)
• Sex
• Age
• Race/ethnicity

6
Data Collection ToolThe Fields

• If multiple investigators/sites
• Who recorded the case (initials)?
• Which site?
• Specify formats
• Time 24 hour time or AM/PM?
• Dates July 4, 1987
• 04.07.87 (European)
• 07.04.87 (U.S.)

7
Data Collection ToolSimplify and Specify Data
Collection

• Circle or check-boxes
• Forces conformity on data recorders
• Reduces problems caused by illegibility
• Include all possible selections that may be of
  interest
• Include Other/Other Specify
• How many choices should be selected?
• All that apply?
• Most pertinent?
• Most severe?

8
Data Collection ToolExamples

• Chest Pain
• Prospective with chart review component
• Violence in EMS
• Anonymous, cross-sectional survey
• No specific identifying information

9
Data Collection ToolStoring

• Once you have begun collecting data, make copies
• Keep a reserve, off-site copy of all of your work
• Dont store it in your car

10
Creating a DatabaseWhy?

• Once you have collected your data
• Pile of paper with information
• To draw any conclusions from these data
• Analyze en-mass
• Analyze many variables
• Analyze relationships between variables

11
Creating a DatabaseGeneral Tips

• Make it simple but complete
• Incorporate all info used at the beginning
  Demographics (age, sex, race.), hypothesis
  questions, inclusion and exclusion questions,
  outcomes
• No chances to go back and recollect data once the
  study is over
• Make it yes/no as much as possible
• Easiest to enter data on a spreadsheet when 1yes
  and 2no

12
Creating a DatabaseGeneral Tips

• Data sheets should have a logical flow
• Needs to make sense to people not directly
  involved in the study
• Provide a way of accounting for data sheets (i.e.
  subjects)
• Numbering system works best
• Remember to maintain subject confidentiality
• Make data sheet simple if subject needs to fill
  it out
• Use lay terms, not medical terms

13
Creating a DatabaseGeneral Tips

• Opening paragraph at the top of each data sheet
  that provides an explanation of the project
• Use language appropriate to reader
• If appropriate, have a second sheet that includes
  reference phone numbers
• Subject may call these numbers if he or she has
  any questions
• Spell check your work!

14
Creating a DatabaseWhat To Use

• Microsoft Excel
• Easy to use
• Microsoft Access
• More difficult to setup
• Once constructed, more difficult to screw-up
• D-BASE
• Well focus on Excel

15
Creating a DatabaseMicrosoft Excel
16
Creating a DatabaseMicrosoft Excel

• Rows
• Columns

17
Creating a DatabaseMicrosoft Excel

• 1 case 1 row
• Each variable for that case goes in a separate
  column
• Label column using first row only
• Variable labels may be up to 8 characters long
  (for analysis)

18
Creating a DatabaseMicrosoft Excel Encoding
Variables

• Use a blank data collection tool as a decoder
• KEEP MULTIPLE COPIES OF THIS DECODER SHEET

19
Creating a DatabaseMicrosoft Excel Encoding
Variables

• Can enter numbers or text
• Male / Female
• M / F
• 1 / 0
• 1 / 2
• During later analysis
• Can do frequency counts on text variables
• Cant do regressions using text variables

20
Creating a DatabaseMicrosoft Excel Encoding
Variables

• For simplicity, use M / F
• Should you need to do regressions, recoding
  variables later to 1 / 0 is not difficult

21
Creating a DatabaseMicrosoft Excel Encoding
Variables

• Date / Time
• Separate columns for month, date, year
• Can merge later
• Separate columns for hours, minutes
• Can merge later

22
Creating a DatabaseMicrosoft Excel Encoding
Variables

• Multiple choice questions
• 1 choice one
• 2 choice two
• 3 choice 3
• No data
• Leave blank excluded from freq. counts
• 99 how many cases had missing data

23
Creating a DatabaseMicrosoft Excel Encoding
Variables

• Multiple choice questions with other
• Create variable for specify if other
• Input text
• If no data, leave blank

24
Creating a DatabaseMicrosoft Excel Encoding
Variables

• Free response questions
• Age _________
• Blood Pressure _________
• Divide into DBP and SBP
• If no data, leave it blank

25
Creating a DatabaseMicrosoft Excel Encoding
Variables

• Free response questions
• Age _________
• Blood Pressure _________
• Divide into DBP and SBP
• If no data, leave it blank

26
Creating a DatabaseMicrosoft Excel Save Early
and Often
27
Creating a DatabaseMicrosoft Excel Save Early
and Often
28
Creating a DatabaseMicrosoft Excel Save Early
and Often

• Save in multiple locations
• Keep a copy for yourself
• Harddrives crash
• Keep a copy on floppy disk

29
Creating a DatabaseMicrosoft Excel Example
30
Creating a DatabaseMicrosoft Excel Example
31
Questions?