Involvement of the consumer voice

It is probably true to say that no two health‐care decisions are the same. Every one of them will reflect the individual circumstances, opinion and knowledge of the potential recipient of the intervention – be it for treatment or prevention of disease – the health practitioners and, sometimes, other people as well, such as parents of a newborn baby. In order to arrive at a satisfactory and satisfying decision, good quality information based on the best current reliable evidence will be needed, so that, for example, the potential recipient can factor in their personal preferences, which will vary from one person to another. ¹ A thorough, and common, understanding of words in everyday use, such as benefit, harm, and risk is essential to meaningful discussion about the decisions to be made. A guest editorial ² clarifying these research terms is a welcome prelude to understanding fully the papers about screening; self‐management; decision making; appreciating preferences; and consumer involvement in research that follow in that issue of Health Expectations.

We must, however, challenge the example used to illustrate how an individual's personal priorities influence their attitudes to particular treatments, in this case tamoxifen for breast cancer. We are concerned, in particular, about the apparent claims for the strength of the evidence concerning this drug's effect on the voice, and the accusations about the efforts that have been made to investigate this reported side‐effect.

Herxheimer, as an Emeritus Fellow of the UK Cochrane Centre, rightly emphasizes the importance of up‐to‐date evidence obtained from the Cochrane Database of Systematic Reviews (http://www.thecochranelibrary.com). To achieve truly effective reviews requires direct involvement of consumers’ input and influence. ³ The consumer in the case of tamoxifen and breast cancer has been repeatedly urged by Cochrane researchers to use The Cochrane Library's feedback system to disseminate her criticisms of the relevant overview, ⁴ and to cite the ‘good independent evidence’ about side‐effects to which she has referred, but has not done so.

It is important that anyone involved in The Cochrane Collaboration, holding office within in it, and advocating its reviews, should be accurate and committed to the principles that underpin the desire to reduce uncertainty about the effects of treatments, and to minimize bias. We would like to take this opportunity to redress this, following Herxheimer's editorial.

Thousands of patients have been studied in the NATO, CRC, IBIS and ATAC trials of treatments for women with breast cancer, which involved the use of tamoxifen. The ATAC trial with over 9000 patients set up for registration of anastrozole with the FDA and with monitors employed for ‘good clinical practice’ (GCP) is a good example of how clinical researchers leave no stone unturned in their search for both anticipated or unexpected adverse effects of drugs. The protocol of this study allowed for any ‘adverse event’ (AE) described anecdotally in the case records to be captured blinded to treatment allocation. ⁵ To suggest, therefore, that the researchers were insensitive to any harm of importance to the individual participants in the trial, whether it was expected or unexpected, is absurd. Mechanisms were available to capture any information on harms that any woman in the trial reported. A detrimental effect on the singing voice was not reported. Furthermore, even within the detailed quality of life study carried out with 1000 patients in the ATAC trial, this problem was not identified.

We would also like to point out that investigating whether a patient has suffered an AE because of a treatment such as tamoxifen needs to take proper account of confounding factors such as concurrent therapy and the psychological stress of the diagnosis and surgical treatment of the cancer. Aside from this, it is well known that lowering of both the singing and speaking female voice can occur at the time of the natural menopause, which frequently coincides with the diagnosis of breast cancer. Finally, statements which imply that if researchers do not do all the extra tests and data collection necessary to collect every outcome that might be deemed relevant by every patient in a trial, this is because they deem these outcomes to be trivial and unimportant does a great disservice to researchers and the people who work with them to design and conduct randomized trials.