Title: Ethical issues in the use of placebo control groups in water intervention trials UC Berkeley School
1Ethical issues in the use of placebo control
groups in water intervention trials UC Berkeley
School of Public HealthJack Colford, MD PhDBen
Arnold, MPHSwiss Tropical Institute
University of BaselDaniel Maeusezahl,
PhDBerkeley Water CenterNovember 9, 2006
2Motivation
- Design crisis / by the numbers
- Has our field reached a point at which it is no
longer ethical in trials of drinking water
interventions for comparison group subjects to be
assigned to receive no active intervention? - No active intervention here implies groups
randomized to receive either placebos OR groups
randomized to continue current drinking water
practices.
3Choices
- What should future trials of drinking water
interventions choose as comparison groups? - No active intervention (conditional on first
question) - Placebos
- Current drinking water practices (traditional
use) - By randomization
- By natural experiment
- Active interventions
- Equivalence trials
- Stepped wedge design
4Outline
- The Berkeley WET studies
- PilotWET (CDC,Emerg Infect Dis 8(1) 29-36)
- BigWET (CDC, Am J Epidemiol 161(5) 472-82)
- HIVWET (CDC/UC, J Water Health 3(2) 173-84)
- Sonoma-WET (NIH, analysis beginning)
- Bolivia-WET (NIH, analysis beginning)
- Rec-WET (Calif Wat Brd, in press, Epidemiology,
Jan 07) - Rec-WET II (NIH, begins June 2007)
- Placebos in domestic drinking water studies
- Role of water in the history of the development
of randomized, blinded, placebo-controlled
trials? - The vertical transmission prevention debate and
its analogy
5Outline (2)
- Choices for comparison arms in water trials
- Superiority designs
- Placebo groups
- Participants current standard of drinking water
(traditional use) - Equivalence / (non-inferiority) designs
- Active device (with previously proven efficacy)
- Conclusions and suggestions
- Your assignment on leaving the room today
6Blind IRBs
7Bolivia-WET
- Natural experiment taking advantage of NGO
roll-out of SODIS near Cochabamba - Funding from NIH (R01)
- 22 villages (660 children lt5y) randomized 5050
to SODIS vs. current practices in a natural
experiment - Villages matched on baseline diarrhea incidence
- NGO roll out staggered
- Community-based participatory training followed
by monthly motivational visit
8SODIS -- Design
8-12
9Bingo!
10Declaration of Helsinki
- The benefits, risks, burdens, and effectiveness
of a new method should be tested against those of
the best current prophylactic, diagnostic, and
therapeutic methods. This does not exclude the
use of placebo, or no treatment, in studies where
no proven prophylactic, diagnostic, or
therapeutic method exists
11HIV Vertical Transmission Trial Debate
- Long course (076) AZT treatment was proven
method known to reduce vertical transmission - Proposed short courses (either short AZT or
single dose nevirapine in peripartum period) were
compared to placebos in most studies - One side argued that the short course should have
been compared to long course AZT therapy and not
to a placebo - The study investigators argued that it was
unethical to conduct the study without a placebo
12Vertical transmission arguments
- A ban on the use of less than the best worldwide
methods would prevent evaluation of less
effective, but practical methods - 75 of vertical transmission occurs during
delivery and post-partum 25 occurs before
labortherefore short course automatically will
leave some women/children at risk - Nevirapine (one dose) was believed to be
effective during delivery (but such a regimen
would have no effect on the 25 of cases
transmitted earlier in pregnancy) - Testing nevirpaine would necessarily result in
more cases of transmission than the standard long
course
13Wendler, D., E. J. Emanuel, et al. (2004). "The
standard of care debate can research in
developing countries be both ethical and
responsive to those countries' health needs?" Am
J Public Health 94(6) 923-8.
14Analogy to waterborne disease trials
- Are waterborne diseases as great a public health
burden as HIV/AIDS? - Are waterborne disease interventions as effective
at reducing mortality/serious illness as HIV
treatments are at reducing HIV transmission in
pregnant, HIV-infected women? - If the ethics of placebo (or not treatment)
controlled trials are being debated in the
setting of vertical transmission of HIV, should
the use of placebos (or no treatment) trials be
re-considered in trials of waterborne disease
interventions?
15WHO standard
- The ethical standards applied should be no less
exacting than they would be in the case of
research carried out in the sponsoring country - WHO. International ethical guidelines for
biomedical research involving human subjects.
Geneva Council for International Organizations
of Medical Sciences, 1993.
16The comparison group debate
- The sole point of disagreement is the best
comparison group to use in assessing the
effectiveness of less-expensive interventions
once an effective intervention has been
identified. The researchers conducting the
placebo-controlled trials assert that such trials
represent the only appropriate research design,
implying that they answer the question Is the
shorter regimen better than nothing?
Lurie, P. and S. M. Wolfe (1997). "Unethical
trials of interventions to reduce perinatal
transmission of the human immunodeficiency virus
in developing countries." N Engl J Med 337(12)
853-6.
17The church of placebo orthodoxy
- Placebo orthodox
- Placebos (when feasible) mandatory on the basis
of methodologic concerns - Problems
- Criteria for ethical use not formally stated
- no permanent adverse consequences or
- only temporary discomfort or
- will not be harmed
18Placebo advocates
- Placebos are necessary to ensure validity
- Without placebos a finding of no difference is
difficult to interpret since the active control
may have been no better than placebo - All agree that placebos are unethical in
life-threatening or serious morbidity situations
19Active control advocates
- Claim that placebo orthodoxy sacrifices ethics
and patient welfare - Helsinki seems to embrace this
- Argues that when effective therapies exist they
must be used - States that the clinically relevant question is
whether the new treatment is better than (or
equivalent to) the existing treatment
20Examples
- Traditional use
- Clasen, T., G. Garcia Parra, et al. (2005).
"Household-based ceramic water filters for the
prevention of diarrhea a randomized, controlled
trial of a pilot program in Colombia." Am J Trop
Med Hyg 73(4) 790-5. - Active group received a ceramic water filter
system designed for use at the household level. - Control group continued to use their
customary practices for collecting, storing, and
drawing drinking water. - Result Odds ratio 0.40 (95 CI 0.25 - 0.63)
21Examples
- Placebo use
- Conroy, R., Elmore-Meegan, M, et al. (1996).
Solar disinfection of drinking water and
diarrhoea in Maasai children a controlled field
trial. Lancet 3481695-97. - Active group told to fill the bottle with
water and leave it in full sunlight on the roof
of the hut - Control group instructed to keep their filled
bottles indoors in the shade - Result Odds ratio 0.66 (95 CI 0.50 - 0.87) for
treated vs. placebo
22Placebo use in trials of waterborne disease
intervention trials
23Comparison arms used in drinking water trials
Data extracted from Clasen T, et al.
Interventions to improve water quality for
preventing diarrhoea. (A Cochrane Review).
Cochrane Library, Issue 3, 2006
24Study design alternatives
- Alternate designs avoiding placebo use
- Superiority trials (either active control or
placebo) - Parallel
- Cross-over trials
- Natural experiments
- Equivalence and non-inferiority trials
- Active control
- Natural experiments
- Stepped wedge
25Superiority vs. equivalence trials
- Superiority trials
- Intended to determine if new treatment is
different from (better than) placebo or existing
treatment (active control). - Null hypothesis is that there is no difference
between treatments. - Alternative hypothesis is that the new treatment
is either different from (two-sided) or better
than (one-sided) control. - Equivalence trials
- Intended to determine that new treatment is no
worse than active control. - We can never assess absolute equivalence.
- We can only assess no difference within a
prescribed margin. - Null hypothesis and alternative hypotheses are
reversed. - Null hypothesis is that difference between
treatments is greater than X. - Alternative hypothesis is that difference between
treatments is less than X.
26Equivalence margin
27Related issues
- Sample size as an ethical issue
- Efficacy vs. effectiveness
- Marginal structural models and the death of the
trial - Stepped wedge design
28Stepped wedge
- Crossover cluster design in which clusters switch
treatmentsbut only in one direction - Clusters are randomized with respect to the time
of switch - Useful in situations with limited resources or
geographical challenges - All clusters eventually receive the intervention
Hussey, M. A. and J. P. Hughes (2006). "Design
and analysis of stepped wedge cluster randomized
trials." Contemp Clin Trials. 2006 Jul 6 Epub
ahead of print PMID 16829207
29Conclusions
- Our field needs to face the issue of comparison
group choice in trials - Greater pre-trial consideration should be given
to alternatives such as - Equivalence designs
- Natural experiments
- Stepped wedge design
- Examination of the role of newer, marginal
methods of inference in observational data - Your assignment