The study
A clinic network in Seattle developed a ten-question online screening tool for a condition that affects about 1 in 100 adults. In a validation study with 500 people who had a confirmed diagnosis and 500 who did not, the tool correctly flagged 90% of those with the condition and correctly cleared 90% of those without it. The network now offers the tool free on its website to the general public. Its information leaflet tells anyone who screens positive: "This test is 90% accurate, so a positive result means there is a 90% chance you have the condition. Please book an assessment."
What is the flaw in this study?
- A Base-rate neglect The flaw
- B Order effects
- C Attrition: only completers analysed
- D Restriction of range
The concept: Positive predictive value and prevalence
This study is fictional, written for practice. No real study, institution or researcher is described.