Circulation. 2000;101:e98
(Circulation. 2000;101:e98.)
© 2000 American Heart Association, Inc.
Circulation Electronic Pages |
Qualification of the Concepts of Unqualified Success and Unmitigated Failure
Kenneth M. Kessler, MD
Professor of Medicine University of Miami School of Medicine,
Miami, FL
 |
Introduction
|
|---|
To the Editor:
By mathematically compensating for adverse effects, Mancini and
Schulzer1 refine the concept of the numbers needed to be
treated (NNT) for a therapeutic success and introduce the use of the
terms: (1) the numbers needed to be treated for an unqualified success
(NNTUS) or (2) an unmitigated failure
(NNHUF). Nevertheless, there appear to be
important limitations to these concepts. First, the authors define an
unqualified success as "...a successful outcome unaccompanied by
treatment-related adverse events...". However, in the context of
practice, as one moves from the NNT to the NNTUS,
more adverse events will occur (eg, in the first study in the Table
shown in the article, the NNT of 58.82 would be associated with 1.18
adverse events, whereas the NNTUS of 60.02 would
be associated with a proportionate increase in adverse events to 1.20).
The only time a success would truly be unqualified would be if there
were no adverse events associated with a given therapy, but then the
calculation of the NNTUS would become
unnecessary. Second, neither the NNTUS nor the
NNHUF compensates for the adverse events that
occur in the proportion of the population who cannot benefit from
treatment. This is an essential concept since (1) these patients will
be harmed never having had the possibility of benefit (they are, in
essence, bystanders) and (2) the majority of adverse events will
emanate from this portion of the population whenever
p1 is <50%. Again using the authors’ first
example in the Table, 94.8% of adverse events would emanate from the
group of patients who could not have benefited from treatment.
Therefore, the number needed to be treated to have an adverse event in
this "bystander" group is 52.74, fewer patients than are needed for
a therapeutic success (NNT=58.82) and far fewer patients than the
NNHUF of 1428.57. Finally, the interpretation of
NNTUS and NNHUF is
confounded whenever the primary end point(s) of therapy and the
treatment-related adverse event(s) are not comparable. For example, the
NNTUS and NNHUF
mathematically equate the desired effect of the maintenance of
sinus rhythm at 12 months with the undesired effect of death (Table 1,
Reference 9) and equate the prevention of mortality with cough, taste
disturbance, and dizziness (Table 1, Reference 15). Therefore,
the interpretation of the absolute NNTUS and
NNHUF, as well as the size of the change between
the NNT and NNTUS or the NNH and
NNHUF, depends on the nature and comparability of
desired and undesired end points of therapy.
|
References
|
|---|
1.
Mancini GBJ, Schulzer M. Reporting risks and
benefits of therapy by use of the concepts of unqualified success and
unmitigated failure: application to highly cited trials in
cardiovascular medicine.
Circulation. 1999;99:377–383.
[Abstract/Free Full Text]
Top
Introduction
References