Sensitivity and Specificity SpPin When a sign, test or symptom has an extremely high specificity (say, over 95%),

a positive result tends to rule in the diagnosis. For example, the specificity of 3 or more positive responses on a !"# $uestionnaire in diagnosing alcoholism is %99% among patients. &herefore, if a person does ans'er (yes( to 3 or ) of the !"# $uestions, it rules in the diagnosis of alcohol dependency. SnNout When a sign, test or symptom has a high sensitivity, a negative result rules out the diagnosis. For example, the sensitivity of the loss of retinal vein pulsation in diagnosing high intracranial pressure is *++ per cent. &herefore, if a person displays retinal vein pulsation, it rules out important increases in intracranial pressure. &hese terms are closely related to the measures of, -ositive -redictive .alue, &he proportion of people 'ith a positive test 'ho have the target disorder/ and

0egative -redictive .alue&he proportion of people 'ith a negative test 'ho do not have the target disorder. 1ften the 2est place to loo3 for 4p-ins and 4n0outs is at the highest (for 4p-ins) and lo'est (for 4n0outs) levels of multilevel li3elihood ratios. Calculations &hese can 2e calculated thus,

sensitivity 5 a6(a7c) specificity 5 d6(27d) li3elihood ratio (897) 5 sensitivity 6 (*:specificity) 5 (a6(a7c)) 6 (26(27d)) li3elihood ratio (89:) 5 (*:sensitivity) 6 specificity 5 (c6(a7c)) 6 (d6(27d)) positive predictive value 5 a6(a72) negative predictive value 5 d6(c7d) Example Table: Retinal veins, pulsation of, and increased intracranial pressure ;<"; <ntracranial -ressure !>4#0 # of retinal vein pulsation -9#4#0 # of retinal vein pulsation total 43 0 )3 019=!8 <ntracranial -ressure 18 128 *)? &otal ?* *@A *A9

;<"; 2y lum2ar puncture (%*9+ mm ;@+), surgery, or evidence of herniation. 019=!8 2y the a2sence of signs, symptoms, or suspicion of high pressure. (4ensitivity and the loss of 49.- 5 *++% 5 4n0outB/ presence of 49.- in normals 5 4pecificity 5 *@A6*)?5AA%) C8evin >#, &he clinical significance of spontaneous pulsations of the retinal vein. !rch 0eurol *9DA/35,3D:)+E Walsh et al chec3ed this out 2y 'atching 9.- during the Fuec3enstedt manoeuvre among a gra2 sample of 9 neurology patients 'ho had normal pulsation prior to the 8-, disappeared 'hen spinal fluid pressure rose past @+) (76: *D) mm ;@1

reappeared 'hen spinal fluid pressure fell past @+* (76: *D) mm ;@1

CWalsh &G, "arden GW, "allacher >, 12literation of retinal vein pulsations during elevation of cere2rospinal:fluid pressure. !mer G 1pthalmology *9?9/?D,95):?.E