Reference Ranges Completed Confidence Ex. Reading Module MBC BY3BC1 = Things to do: · Type Lecture Learning Objective: · How does population screening work? · What is therapeutic drug monitoring? Possible Test Results . nothing is perfect in reality and there is always scope for error · True Positive (TP) - positive test and subject has disease · False Positive (FP) - positive test and subject does not have disease · True negative (TN) - negative test and subject does not have disease · False Negative (FN) - negative test and subject has disease Sensitivity · proportion of patients who have the disease who give a positive test · how good the test is at correctly identifying people with the disease ·× 100 True positives + False negatives True positives Reference Ranges 1
Specificity Normal population Negative test False negatives population Positive test False positives · ideally each would be around 100% · almost impossible · so should aim for >90% . raising the cut off will increase the specificity · false positives will reduce but there may be many false negatives and the sensitivity will fall . the mean is the middle of the curve, . anyone above this mean will have the disease · anyone below will not have the disease but may have other health indications . the pink line = threshold value . the value will be indicative of the disease · anyone above this line, does have this disease · anyone immediately before it = pre-disease (very close) · anyone before it = does not have the disease Population-based reference ranges Reference Ranges 2
. Most common means establishing a reference range · essential to use local population · eliminates some social, genetic and environmental variables · age · gender . race · diet · genetics · we have to exclude people who are already ill also people: · risk factors: e.g. obesity, hypertension · Drugs: e.g. alcohol, oral contraceptives · Physiology: e.g. pregnancy, stress, excessive exercise Reference range production · Gather values: · patients consent · sample collection · specimen storage and analysis · partition if necessary · Inspect and apply data: · easiest if data are normally distributed Reference Ranges 3