Data collection principles

Transcript

1. Produced by NIHR Global Surgery Unit - Licenced under Creative Commons for education use • Good and bad variable names • International data formats • Unknown and Missing values Defining variables Data Collection Principles Photo by National Cancer Institute on Unsplash
2. Produced by NIHR Global Surgery Unit - Licenced under Creative Commons for education use Good variable names: for analysis and data sharing Produced by NIHR Global Surgery Unit - Licenced under Creative Commons for education use, Photo by Austin Kirk on Unsplash All Lowercase • age, not Age • bmi, not BMI Separated with underscores • cancer_type, not cancer-type • No other special characters or spaces, avoid (- : ; & etc.) Consistent with previous projects • Make sure to use same variable names as your previous projects Short, but meaningful • blood_ox, not meas1 • frailty, not fr
3. Produced by NIHR Global Surgery Unit - Licenced under Creative Commons for education use Fix these variable names ❌ bad names ✅ good names EXERCISE • Body mass index • smoking-status • distanceToHome • Endoscopic/colonoscopic biopsy • size: tumour • bmi • smoking_status • distance_to_home • biopsy_scope • size_tumor
4. Produced by NIHR Global Surgery Unit - Licenced under Creative Commons for education use YYYY-MM-DD: 2024-04-03 Data standards: there is only one international standard for dates Produced by NIHR Global Surgery Unit - Licenced under Creative Commons for education use Photo by Towfiqu barbhuiya on Unsplash
5. Produced by NIHR Global Surgery Unit - Licenced under Creative Commons for education use 0.5 0,5 Data standards: there are two decimal separators Produced by NIHR Global Surgery Unit - Licenced under Creative Commons for education use Map by NuclearVacuum - File:BlankMap-World6.svg, CC BY-SA 3.0, https://commons.wikimedia.org/w/index.php?curid=10843055
6. Produced by NIHR Global Surgery Unit - Licenced under Creative Commons for education use 1 NA Stands for both Not Available as well as Not Applicable, it’s a generic way to record both Unknown and Missing values 2 Missing The value could be known, but it has not been entered into the database, for example the study staff ran out of data time to enter data 3 Unknown The value of the variable can’t be known, for example the patient has not been tested for the disease the variable is asking about Missing or Unknown data: make it explicit Produced by NIHR Global Surgery Unit - Licenced under Creative Commons for education use
7. Produced by NIHR Global Surgery Unit - Licenced under Creative Commons for education use • Good variable names • Common formats/standards • Unknown and Missing values Summary Photo by National Cancer Institute on Unsplash