Sensitivity and specificity can both improve as more investigations are used

BMJ 1998; 316 doi: http://dx.doi.org/10.1136/bmj.316.7124.69 (Published 03 January 1998) Cite this as: BMJ 1998;316:69
  1. Neil Gillespie, Lecturer on aging and healtha,
  2. Stuart Pringle, Consultant cardiologista,
  3. Allan Struthers, Professor of clinical pharmacologya
  1. a Ninewells Hospital and Medical School, Dundee DD1 9SY

    Editor—We wish to reply to Davie and McMurray's comments about our article on the assessment of acutely dyspnoeic patients with suspected heart failure; we did not see it before publication.1 We too were surprised by the excellent sensitivity of clinical examination in our study. We suspect that all our patients with systolic dysfunction had advanced disease and that assessment by general practitioners before admission resulted in referral of only those with New York Heart Association grade III or IV heart failure.

    Davie and McMurray criticise the sensitivities obtained when combinations of investigations were used in addition to clinical examination. The apparent misunderstanding relates to the fact that we used clinical examination and/or the various investigations (chest radiography, electrocardiography) in the detection of systolic dysfunction. The left hand column in table 2 in our paper shows the collection of investigations that were used to reach the diagnosis in each case, but the final diagnosis was based on the appropriate abnormality on clinical examination and chest radiography or electrocardiography. This is why sensitivity and specificity can both improve as more investigations are considered. We apologise if this was unclear from our table 2.


    1. 1.
    View Abstract