Pre-randomisation consent (Zelen

1
Pre-randomisation consent(Zelens method)
2
Background

• Most individually randomised trials ask patient
  CONSENT before randomisation.
• There are problems with asking for consent some
  scientific others ethical.

3
Consent problems - Scientific

• Asking consent means refusal by a proportion of
  the population, which limits the generalisability
  of study. Consent also alerts patient to trial
  potentially leading to Hawthorne effects or
  resentful demoralisation.
• Sometimes getting consent is simply not possible.

4
Pre-randomisation consent

• Two different types of pre-randomisation consent.
• Some trials cannot refuse allocated intervention.
• Other trials use Zelens method where consent to
  treatment can be withheld.

5
No Consent

• Trial of income sanctions for childhood
  immunisation. Families were randomised to have
  income supplements reduced if they did not get
  their children immunised. Cannot refuse consent
  to the sanction, which is the intervention.
• Similarly for cluster trials.

6
Zelens method

• Zelens method allows refusal of the
  intervention.
• This introduces DILUTION effects as people cross
  over to the other treatment.

7
Consent problems -Ethical

• Whilst NOT asking consent is usually seen as
  UNETHICAL sometimes it has been justified as
  being more ethical.
• Example - an early ECMO trial used Zelens method
  because it was felt to be unethical to ask very
  distressed parents consent.

ORouke. Pediatrics 198984957.
8
Consent problems - feasibility

• Sometimes it is not feasible to ask consent at
  the time randomisation is required.
• Example, a trial of allocating confused elderly
  used Zelens method as relatives were unavailable
  to give proxy consent.

9
Consent problems - administrative

• The requirement for doctors to ask for consent
  can reduce or slow recruitment to trials.
• Also Zelen felt a clinicians might compromise
  patient doctor relationships by admitting that
  they did not know which treatment to recommend.
• Because of these issues Marvin Zelen proposed the
  randomised consent method.

10
Randomised Consent

• Zelen argued that for patients allocated to
  conventional therapy there was no need to get
  consent as patients would get standard therapy
  anyway. Consent for treatment only is required.
• Consent from patients allocated to the new
  treatment only and this is consent for treatment.

11
Two versions

• There are two versions of Zelens method - the
  original method, known as the single consent
  design or later the double consent approach.
• In the first only patients allocated to the novel
  treatment are asked for consent to treatment,
  which can be refused. In the second both groups
  are asked for consent to treatment.

12
Zelens single consent method
13
Applications

• Zelens method has commonly been used in
  screening programmes where people are randomised
  without consent to be invited to be screened or
  not.
• Follow-up is often done at a distance through,
  for example, cancer registeries.

14
Colorectal Cancer Screening

• Harcastle et al. Randomised nearly all the
  middle aged male population (45-74) in Nottingham
  to be screened for early signs of colorectal
  cancer or act as controls.
• Zelens method was used in order to gain
  pragmatic estimates of the effect of screening
  on a population.

Hardcastle Lancet 19963481472-77.
15
Colorectal Cancer

• In the Harcastle trial patients in the control
  arm were allowed unlimited access to diagnostic
  and treatment facillities for bowl cancer at
  their own or GPs request.
• Of 76,000 men offered screening 60 accepted.
  This lead to a 15 reduction in bowl cancer
  deaths.

16
Advantages or Disadvantages?

• Had Hardcastle used a standard method of RCT
  then this would have resulted in less dilution
  bias BUT would not have given a pragmatic
  estimate of the effects of cancer screening.
  Also some participants would have excluded
  themselves from the control group, which may have
  made the trial less generalisable.

17
Screening What is the question?

• If the question to be answered is
• What is the effect of introducing a bowl cancer
  screening programme? Zelens method gives the
  best answer.
• If the question is
• What is the effect of giving bowl cancer
  screening to people who want screening? Normal
  consent methods are best.

18
Practical Considerations

• Dey et al used Zelens method because of
  practical reasons. They wished to evaluate a
  one stop breast clinic. To ask consent before
  allocation would have been difficult as women
  would already be at the clinic and the clinicians
  required notice for diagnostic tests.

Dey et al. BMJ 20023241-5.
19
Disadvantages - Ethical

• Many ethics committees and researchers are
  opposed to Zelens method because of non-consent.
  
• BMJ debated this issue and decided to still
  publish Zelens trials despite some ethicists
  recommending not to publish.
• As in all things Zelens method is ethical and
  unethical depending on the circumstances.

20
Ethics

• If consent is impracticable Is it more or less
  ethical to offer an unevaluated treatment than
  offering a treatment without consent where the
  patient has at least a 50 chance of getting the
  safer treatment.

21
Disadvantages - Scientific

• Major scientific problem is through dilution
  bias, particularly with the double consent
  method.
• Cross over into the opposing group will dilute
  any treatment effect and make it harder to
  observe a difference.
• A review of Zelens trials in cancer treatments
  concluded cross-overs were a serious problem in
  many trials.

Altman et al. Eur J of Cancer 199531A1934-44.
22
Disadvantages

• Because of the threat of cross over one may need
  larger sample sizes to cope with dilution
  effects, which can increase the cost of duration
  of the trial.
• Outcome collection may alert patients to the
  trial which could result in WORSE demoralisation.

23
Systematic review

• Adamson et al, have reviewed 43 trials published
  since 1990 using the Zelen approach.
• Most of the trials used the method to avoid the
  introduction of bias (e.g., Hawthorne effect) NOT
  to enhance recruitment.
• Low cross-over (median 6).

Adamson J, Cockayne S, Puffer S, Torgerson DJ.
Submitted.
24
Future of Zelens method

• Probably useful for pragmatic trials of
  screening.
• May be useful for other treatments on occasions.
  May not be worth, generally, overcoming the
  ethical barriers to its use.
• HOWEVER, cross over is relatively low.

25
Summary

• Pre-randomisation consent is useful in some
  circumstances.
• Zelens method offers a number of potential
  advantages, particularly for screening type
  trials.
• It shouldnt be undertaken if there is the
  potential for high cross over rates.