S05 Screening & Diagnostic Tests (FILLED IN)

False nega2ves in 1st test loss of net sensiuvity gain

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 75% Example (Parallel TesUng) Sn = 90% Sp = 75% - ve +ve 140 360 500 - ve 60 1440 1500 200 Test #1 +ve 1800 2000 Test #2 Sn = 70% Sp = 80% True Disease Status True Disease Status +ve - ve +ve 180 450 630 - ve 20 1350 1370 200 1800 2000 •  Net SensiUvity à༎ a)  b)  c)  d)  +ve on both tests: 140 TPtest1 * 90% Sntest2 = 126 +ve only on test 1: 140 TPtest1 – 126both(step a) = 14 +ve only on test 2: 180 TPtest2 – 126both(step a) = 54 Net Sn = (126 + 14 + 56) / 200 = 194 / 200 = 97% HIGHER Example (Parallel TesUng) Sn = 90% Sp = 75% - ve +ve 140 360 500 - ve 60 1440 1500 200 Test #1 +ve 1800 2000 Test #2 Sn = 70% Sp = 80% True Disease Status True Disease Status +ve - ve +ve 180 450 630 - ve 20 1350 1370 200 1800 2000 •  Net Specificity à༎ –  Recall: only considered disease –ve if –ve on both tests –  - ve on both tests: 1440 TN (test1) * 75% Sp (test2) = 1080 –  Net Sp = 1080 / 1800 = 60% LOWER MulUple Tests (Pros & Cons) •  Series TesUng –  Less expensive –  Less invasive (1st test) –  Aims to avoid false posi2ves (by bringing back for 2nd text) •  Will miss disease- posi2ve individuals who are false nega2ves in 1st test –  Loss of net sensiUvity –  Gain in net specificity •  Parallel TesUng –  More expensive –  Good for preven2ng false nega2ves (need - ve result from two tests!) –  Gain in net sensiUvity –  Loss of net specificity Validity (Sn & Sp) Relevance •  So far we have been interested in how good a test is at idenUfying individuals with the disease (Sn) and without the disease (Sp) –  Very useful for screening programs –...
View Full Document

This document was uploaded on 01/14/2014.

Ask a homework question - tutors are online