Diagnostic Tests Flashcards
What is a test ?
- biological measure I.e HB , Ca2+
- images - ultrasound , mammogram
- questions - CAGE questions ( craving , anger , guilt , eye opener )
Examination - tactile vocal fremitus , liver edges
Diagnostic tests
Always a possibility that test results are inaccurate and diagnosis is wrong so need to have a certain threshold before we decide to treat and below threshold no treatment - threshold depends on disease , potential harm and benefits or treat or no treat - need to prevent false positives ( treat patients unnecessarily ) and false negatives ( fail to treat others adequately )
Disease and test result
Disease present and test result + then True positive - maximise
If disease is absent and test result is + then False positive - minimise
If disease present and test result is - then False negative - minimises
If disease is absent and test result is negative then True negative - maximise
What is sensitivity ?
Tests with high sensitivity correctly classify a high proportion of people who really have disease
What is specificity ?
Tests with high specificity correctly classify a high proportion of people who really don’t have a disease
Positive predictive value
= no true positives : All those test positive
Chance of having disease if your test is positive
Negative predictive value
No true negatives / All those test negative
Chance of not having disease if your test is negative
Summary
Sensitivity - no true positives - all those with disease
Specificity - no true negatives - all those without disease
PPV - no true positives - all those test positive
NPV - No true negatives - all those test negatives
Prevalence
Some aspects of test performance are strongly affected by prevalence
Sensivity and specificity of a test often stays constant
Predictive values can change
As prevalence rises
- negative predictive value fails
- the positive predictive value rises
And vice versa
Situations where prevalence Changes
- between primary care and secondary care
- across age groups
- between countries
Likelihood ratios
Another way of summarising performance of tests
A test with 2 outcomes ( positive/negative ) has two likelihood ratios :
- likelihood ratio for a positive test ( LR+)
- likelihood ratio for a negative test result ( LR-)
Approach can be extended for continuous results (e.g on enzyme level )
Likelihood ratios can help you assess how the chance of disease have changed after a test
Likelihood ratios
The larger the LR+ve greater chance have disease if your test is positive
The smaller the LR-ve less chance have disease if your test is negative
Basically
Chance of disease before test x LR - chance of disease after test
