EDITOR - It took six years to spot the unsupported conclusions
derived from faulty analyses reported by Pawlikowska and colleagues. (1)
Bland's criticism is directed at the referees, editors and those who have
cited this work. (2) The original paper and the first two of the three
groups mentioned by Bland are events that happened and cannot be changed.
There are two points to my letter. First, I believe that Bland's critique
should not be disposed of too readily but be used to revisit this never-
ending problem of unjustified generalisations we are bombarded with. The
second point is a related matter about the third group of people mentioned
by Bland.
Not long ago, a medical professional I encountered would refer to the
analysis section of papers as the 'dry bits' that we should not bother
about. On one occasion, a discussion touched on the concepts of type I and
II errors but this doctor quickly terminated the matter by stating that if
a result was statistically significant then these issues were irrelevant.
I do not believe that this doctor's views are exceptional. However, while
this attitude is somewhat 'harmless' for most of us, the mere mortals who
keep consuming information, this doctor was an important key figure in a
Cochrane collaboration group. I am very worried about the illusive links
between evidence-based medicine and sets of information such as the
Cochrane database. I have observed that many tend to accept without
reservation the dicta that crop up in the form of systematic reviews, in
particular those that contain meta-analytic techniques, as if they were
foolproof statements of truth. Not only are there the problems related to
the gathering and filtering of information (my anecdote partly illustrates
this point; Bland's letter seals it), but also there is the additional
fact that statistical techniques continue to evolve. For example, in the
world of systematic quantitative reviews, pooled numbers needed to treat
have been considered misleading quite recently. (3)
From where I sit, clinicians thirsty for valid and justified conclusions
to be used in practice, I share Professor Bland's concerns.
David Marchevsky consultant psychiatrist
Department of Psychiatry, Campbell Centre, Milton Keynes MK6 5NG
Competing interests: None declared.
Pawlikowska T, Chalder T, Hirsch SR, Wallace P, Wright DJM, Wesley
SC. Population based study of fatigue and psychological distress. BMJ
1994;308:763-6.
Bland M. Fatigue and psychological distress. Statistics are improbable.
BMJ 2000; 320:515-6.
Smeeth L, Haines A, and Ebrahim S. Numbers needed to treat derived from
meta-analyses - sometimes informative, usually misleading. BMJ 1999;
318:1548-1551.
Competing interests:
No competing interests
29 February 2000
David Marchevsky
consultant psychiatrist
Department of Psychiatry, Campbell Centre, Milton Keynes MK6 5NG
Rapid Response:
Statistics were improbable, so what?
EDITOR - It took six years to spot the unsupported conclusions
derived from faulty analyses reported by Pawlikowska and colleagues. (1)
Bland's criticism is directed at the referees, editors and those who have
cited this work. (2) The original paper and the first two of the three
groups mentioned by Bland are events that happened and cannot be changed.
There are two points to my letter. First, I believe that Bland's critique
should not be disposed of too readily but be used to revisit this never-
ending problem of unjustified generalisations we are bombarded with. The
second point is a related matter about the third group of people mentioned
by Bland.
Not long ago, a medical professional I encountered would refer to the
analysis section of papers as the 'dry bits' that we should not bother
about. On one occasion, a discussion touched on the concepts of type I and
II errors but this doctor quickly terminated the matter by stating that if
a result was statistically significant then these issues were irrelevant.
I do not believe that this doctor's views are exceptional. However, while
this attitude is somewhat 'harmless' for most of us, the mere mortals who
keep consuming information, this doctor was an important key figure in a
Cochrane collaboration group. I am very worried about the illusive links
between evidence-based medicine and sets of information such as the
Cochrane database. I have observed that many tend to accept without
reservation the dicta that crop up in the form of systematic reviews, in
particular those that contain meta-analytic techniques, as if they were
foolproof statements of truth. Not only are there the problems related to
the gathering and filtering of information (my anecdote partly illustrates
this point; Bland's letter seals it), but also there is the additional
fact that statistical techniques continue to evolve. For example, in the
world of systematic quantitative reviews, pooled numbers needed to treat
have been considered misleading quite recently. (3)
From where I sit, clinicians thirsty for valid and justified conclusions
to be used in practice, I share Professor Bland's concerns.
David Marchevsky
consultant psychiatrist
Department of Psychiatry, Campbell Centre, Milton Keynes MK6 5NG
Competing interests: None declared.
Pawlikowska T, Chalder T, Hirsch SR, Wallace P, Wright DJM, Wesley
SC. Population based study of fatigue and psychological distress. BMJ
1994;308:763-6.
Bland M. Fatigue and psychological distress. Statistics are improbable.
BMJ 2000; 320:515-6.
Smeeth L, Haines A, and Ebrahim S. Numbers needed to treat derived from
meta-analyses - sometimes informative, usually misleading. BMJ 1999;
318:1548-1551.
Competing interests: No competing interests