[e-drug] Filariasis control: ethics, economics, and science (cont'd)

[John Urquhart <urquhart@ix.netcom.com>]

E-drug: Filariasis control: ethics, economics, and science (cont'd)
-------------------------------------------------------------------------

This note concerns the letter in yesterday's Lancet, describing the
problem of designing a suitably controlled, ethically acceptable
study to estimate the effectiveness of DEC in eradication of
filariasis.

The design and analysis of an ethical study that was used to good
effect on the value of vitamin A supplementation is described in:
Sommer A, Zeger SL. On Estimating Efficacy from Clinical Trials.
Statistics in Medicine 10: 45?52, 1991. The design involved the
provision of treatment to some villages, and not to others, the latter
designated as controls. A key problem in estimating efficacy was
that some patients in 'treatment' villages did not receive the
treatment, because they were not present at the right time, in the
right place. These people tended to be, within the socio-economic
strata of their respective villages, the poorest & least informed, and
also most likely to be malnourished and thus susceptible to
complications of malnutrition, some of which were the endpoints of
the trial. Thus, their characteristics differed substantially from the
overall averages in the untreated villages. It was therefore essential
to formulate a reliable, unbiased estimate of events within this
untreated, important minority, and to estimate its impact on the
estimates of efficacy, nominally taken as the difference between
treated and control villages. Sommer and Zeger utilized a special
approach, known in the biostatistical trade as "instrumental
variables", which is described in this seminal paper. The technique
has cropped up more or less independently in a number of other
situations where interventions have nonuniform application, e.g., it
is called "Bloom's variables" in educational studies, but the
Zeger-Sommer use is a first in biomedical applications.

By this means, Zeger & Sommer were able to convert variable
execution of a planned program of treatment into a reliably and very
usefully interpretable observational study -- in effect taking
advantage of a superficially undesirable deviation from the regimen
assignments called for by the protocol. The result was far more
informative than the conventional intention-to-treat analysis, which
censors information on actual receipt of treatment. The result also
provided quantitative data on the crucial importance of taking
special efforts to reach the people who, in the trial, went untreated.

John Urquhart, MD, FRCP(Edin)
Professor of Pharmaco-epidemiology
Maastricht University, Maastricht, NL
Chief Scientist, AARDEX Ltd/APREX Corp, Zug, CH & Union City,
CA, USA
Professor of Biopharmaceutical Sciences, UCSF, San Francisco
home office: 975 Hamilton Avenue, Palo Alto, CA 94301 USA
email: urquhart@ix.netcom.com


--
Send mail for the `E-Drug' conference to `e-drug@usa.healthnet.org'.
Information and archive http://satellife.healthnet.org/programs/edrug.html
Mail administrative requests to `majordomo@usa.healthnet.org'.
For additional assistance, send mail to:  `owner-e-drug@usa.healthnet.org'.