The use of likelihood ratios can over come some of this. However, most areas of diagnosis and prognosis, still suffer from small number problems in their development. Perhaps, could explain the range of CI to patients, by stating that as we use the test, treatment, etc. the CI should shrink as we accumulate data. Of course, it would behoove us all to do this on a local level, so that we can generate CI's for our patients.
Best regards,
Dan Sontheimer, M.D., Asst. Prof. Of Family Medicine
Spartanburg Family Medicine Residency Program
853 N. Church Street Suite 510
Spartanburg, SC 29303
Email: [log in to unmask]
-----Original Message-----
From: Jim Wang [SMTP:[log in to unmask]]
Sent: Monday, December 13, 1999 6:34 PM
To: [log in to unmask]
Subject: Confidence interval
Dear All,
In counselling/discussing with patients about their individual risk is
confidential interval useful? I feel it is not since no one can really
understand the true meaning of " your risk is 30%, but the confidential
interval can be as low as 15% and as high as 45% while the actual final
outcome is you either get it or not get it". But CI is certainly useful in
a group situation to assess the accuracy of your estimation or your
treatment efficacy etc. Any comments?
Regards and Happy New Year to all of you.