 |
 |
Dear all (please reply to sender, [log in to unmask]) I wonder whether somebody can help me. I have estimated a cut point corresponding to a point where sensitivity is equal to specificity for a specific test in a number of study populations. Thus the waist measurement for optimum prediction of diabetes is 84cm in group A (n=300), 89cm group B (n=350), 94cm in group C (n=375), etc. I am trying to confirm that this measurment behaves differently in these different populations. That is that I want the difference in cut-point between groups A and B and C to be significant. Ways to do this would be to: 1. Formally test the differnces in cut-point between the groups. 2. Generate 95% confidence intervals for the cut-points. I thought about 1. translating the cutpoint into a centile 2. calculating the 95%ile for that cut-point However, my thoughts get a little fuzzy and I cannot get my mind around how to do this... Sorry. Can anyone put me out of my misery? Cheers Terry ********************************************* Dr Terry Aspray Department of Medicine The Medical School Newcastle upon Tyne NE2 4HH tel +44 (0)191 222 5407 pager +44 (0)4325 241613 Sheila Patel, Research Associate, Department of Epidemiology & Public Health University of Newcastle upon Tyne Wellcome Research Laboratories Royal Victoria Infirmary Newcastle upon Tyne NE1 4LP England UK Tel: (0191) 222 7154 Fax: (0191) 261 1763 E-mail: [log in to unmask] %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%