Authors' reply

BMJ 1994; 308 doi: https://doi.org/10.1136/bmj.308.6923.274c (Published 22 January 1994) Cite this as: BMJ 1994;308:274
  1. T Nguyen,
  2. G Jones,
  3. P Sambrook,
  4. P Kelly,
  5. S Lord,
  6. J Freund,
  7. J Eisman
  1. Garvan Institute of Medical Research, St Vincent's Hospital, Darlinghurst, NSW 2010, Australia.

    EDITOR, - The reference cited by Malcolm Duncan1 in no way suggests that our modelling approach was erroneous; in fact, in principle it indicates that our use of this technique was totally appropriate. Sensitivity and specificity (that is, receiver operating characteristic curves) can be calculated from a single variable by treating the variable as either continuous or categorical data. In a more complex situation several variables need to be …

    View Full Text

    Log in

    Log in through your institution


    * For online subscription