Legionnaires' disease outbreak investigation toolbox

Download Page

Study questionnaire analysis template

The following template can be used in conjunction with the hypothesis-testing questionnaire (section 9). The following template is designed for use in Epidata, but should provide indicative steer for analysis with other software - i.e. EpiInfo or R. An electronic EpiInfo version of the questionnaire is provided.

Analysis template

• Close any open data file

• Close any open log file: logclose

• Clear the screen: cls

• Change working directory for example: cd "C:\Documents and Settings\user.name\Desktop"

• Open log file: logopen outbreaklog

• Read in data file as entered into Epidata Entry: read outbreak

• List the variable names: var

• Visually check data for accuracy and missing data: browse

• Tell Epidata Analysis which code (here 999) used to denote a missing variable: missingvalue age /999

• Look for data entry errors by summarising and plotting each quantifiable variable

• Summarise each quantitative variable: describe ___

• Plot each quantitative variable: histogram age

• Correct outlier after checking paper form: select age > 100 list if id=x then age=y

• Look for data entry errors by tabulating each categorical variable: tables * /f, tables sex /m

• Correct data entry error after checking form if sex="ff" then sex="f", tables sex

• Create any new variables required and add labels define agegp #, agegp=999

missingvalue age /999

if age < 40 then agegp=0

if age >=40 then agegp=1

tables agegp

label agegp "Age group"

labelvalue agegp /0="Under 40"

labelvalue agegp /1="40 and over"

• Describe outbreak in terms of time with an epidemic curve: epicurve status dateonset /edit

• Describe outbreak in terms of place: place

• Describe outbreak in terms of person characteristics (for Case Control Studies, often stratified by case/control status)

means age /by=status

tables status sex /c

• Univariate analysis: tables status exposure 1 exposure 2 exposure 3....../ct /obs

• Further analysis to look at issues of confounding or interaction

• Epidata Analysis to ask user where to save data: savedata mydata /replace

• Delete graphs: erasepng /noconfirm

• Close programme: exit