Intended for healthcare professionals

Education And Debate

Modified informed consent procedure: consent to postponed information

BMJ 2003; 327 doi: (Published 31 July 2003) Cite this as: BMJ 2003;327:284
  1. Han Boter, junior researcher (h.boter{at},
  2. Johannes J M van Delden, professor of medical ethics2,
  3. Rob J de Haan, professor of clinical epidemiology3,
  4. Gabriël J E Rinkel, professor of neurology1
  1. 1Department of Neurology, Rudolf Magnus Institute of Neuroscience, University Medical Centre Utrecht, PO Box 85500, 3508 GA Utrecht, Netherlands
  2. 2Julius Centre for Health Sciences and Primary Care, University Medical Centre Utrecht
  3. 3Department of Clinical Epidemiology and Biostatistics, Academic Medical Centre, PO Box 22660, 1100 DD Amsterdam, Netherlands
  1. Correspondence to: H Boter
  • Accepted 19 May 2003

How do you obtain a valid assessment of subjective outcomes in a trial in which the participants cannot be blinded to the intervention? Bias is inevitable from unblinded patients, but trials that have not told patients about treatment in all arms have been heavily criticised. Asking participants to consent to postponed information could be a solution

The most powerful tool for studying the effectiveness of a medical treatment is a randomised controlled clinical trial with blinded assessment of outcomes. However, blinding is not always possible—for example, in a trial comparing a surgical intervention with non-surgical treatment or the effectiveness of supplemental care compared with conventional care. Blinding needs special attention in such studies in order to prevent bias. When outcomes are assessed by doctors, it is often easy for the assessment to be done by doctors other than those who performed the procedure. However, when studies measure patients' assessment of outcomes, blinding is much more complicated or even impossible.

Embedded Image


Dealing with unblinded participants

Unblinded patients who assess outcomes after being informed about the different treatment options during recruitment might bias the results of a study. The likelihood of bias increases when patients have a preference for one of the treatment options. For example, patients may have been told during recruitment that the new or supplemental strategy has been developed because the current strategy has disadvantages. The best way to obtain valid assessments in such studies remains unclear.

Intense debate was generated when researchers tried to mask patients in a trial on the effectiveness of additional care from a stroke family care worker by keeping them ignorant of the treatment options.110 The researchers used a “two-stage randomised consent design,”11 12 in which they sought patients' consent for follow up but not for randomisation. Patients were randomised after they gave consent. The researchers then sought additional consent from patients randomised to non-standard treatment but not from patients randomised to receive standard care.1 11

Modified informed consent procedure

We developed a modified consent procedure for a trial on the effectiveness of an outreach nursing care programme for patients who returned home after being admitted for a stroke. Because many stroke patients experience reductions in quality of life or are dissatisfied with the care received after discharge from hospital,13 14 we intended using self reported quality of life and satisfaction with care as primary outcomes. However, we were concerned that unblinded patients could lead to biased results. Firstly, patients in the control group might be dissatisfied because they knew that other patients were receiving outreach care. Secondly, patients who received outreach care might make a more favourable assessment out of loyalty to the programme's staff. We therefore asked patients' consent to follow up and to postpone information on the study until the end of the follow up (six months after discharge).10 15

The box shows the oral and written information that study nurses gave to eligible patients before discharge. Additionally, patients were given telephone numbers for one of the authors (HB) and the treating neurologist, in case they had questions or complaints about the study.

Patients who gave consent were randomised. Those in the intervention strategy were then informed about the outreach care programme and asked to participate but were not told that the study was assessing the effectiveness of the programme. Patients who were randomised to the control strategy received no further information at that time. After six months, we sent a letter with the postponed information to all recruited patients who were known to be alive at follow up. It contained the postponed information on the additional research question, the randomisation, the reasons why we did not inform the patient about the additional question and the randomisation during recruitment, and telephone numbers of HB and the treating neurologist. We wrote this letter in close cooperation with the patient service office and the ethics committee of one of the university hospitals and paid special attention to its clarity.

Is modified consent ethical?

The ethical acceptability of our modified procedure for informed consent can be questioned. At least three arguments have been raised against modifications of this kind.1 4 16 The first is that by not providing patients with full information during recruitment we are not treating them with respect. The ethical basis for informed consent is not that informing patients enables them to make informed decisions about participation in the study. Instead, the basis for informed consent is the respect that we owe them as moral actors, irrespective of how they evaluate being informed or the information as such. A second argument holds that a modified procedure might evoke negative responses in patients, leading to a decreased willingness to participate in future research. A third, closely linked criticism is that it may reduce patients' trust in their doctors.

However, questions can also be raised about using costly services that have not been evaluated appropriately. In some instances unbiased evidence can be gathered only by using modified procedures for informed consent. We think that, on balance, there can be good reasons for using such a modified procedure. We also feel that our modified procedure exceeds the two stage randomised consent design used previously.1 Our participants knew that they were not fully informed during recruitment and gave their consent. Admittedly this is not the same as being fully informed, but by giving patients a choice we took them seriously as moral actors. Additionally, all enrolled patients, including non-respondents, who survived until follow up were eventually informed about the intervention and the randomisation.

Information given to patients during recruitment

  • We are studying patients' needs six months after discharge and their satisfaction with services received after discharge

  • We cannot inform you about an additional research question since that would affect the results

  • You will be informed about this question after assessment of the outcome

  • This additional research question entails no risk

  • The ethics committee approved this study

Summary points

Blinding of patients in trials that use self reported assessment of outcomes is crucial but can be difficult in some circumstances

The ethics of blinding patients by asking consent to follow up but not to randomisation have been intensely debated

An alternative is to ask consent to follow up and postponed information

Although patients know that they will be fully informed after follow up, some people may still believe that the procedure does not treat patients with enough respect

Although conventional informed consent procedure should be the first choice, our modified procedure with postponed information deserves consideration when subjective outcomes are used as the primary outcome measurements, the outcome assessors cannot be blinded, the additional treatment entails no risk, and the intervention seems attractive to patients.

Embedded Image

Members of the steering committee are lised on


We thank A Algra, J van Gijn, and L J Kappelle for their comments on a previous version of this manuscript.


  • Contributors and sources HB is a nurse researcher with experience in healthcare research. JJMvD is a nursing home physician and medical ethicist. RJdeH has expertise in clinimetrics. GJER has experience in stroke research. All authors were involved in the development of the informed consent procedure.

  • Funding The trial was supported by an established clinical investigator grant from the Netherlands Heart Foundation to G J E Rinkel (grant D98.014), by a grant from the Netherlands Heart Foundation and the Netherlands Organisation for Scientific Research (940-32-014), and by a grant from the University Medical Centre Utrecht.

  • Competing interests None declared.


View Abstract