We thank our correspondents for pointing out an important issue that we need to address more assiduously in our reviews of papers. We agree fully that the British Journal of Psychiatry needs to ensure that a greater policy of openness towards low- and middle-income countries is not accompanied by any lowering of ethical standards.

However, there are clear divisions of opinion here. When the protagonists for each of these make their eloquent arguments, it may seem strange that any should remain rather uncomfortably on a rickety fence when the alternative certainties are so much more inviting. Well, we are still wobbling because we feel it is right to wobble. The two sides of this argument, put crudely, are (a) it is unethical to exploit patients in low-income countries for studies that would never be allowed to proceed in rich countries, and (b) research performed for a global scientific community has to provide general evidence, not specific to one group or country, and so worldwide efficacy studies are necessary.

Drs Murtagh & Murphy, Basil et al, and Srinivasan et al all allege, directly or indirectly, that the patients in India have been selectively exploited for research purposes and this is fundamentally unethical. Patel (Reference Patel2006) also asks whether there is a personal financial aspect to the trial that has been undeclared. The allegation that ‘this trial could not have been conducted in a high-income country but may have been conducted in India because regulatory requirements could be fulfilled there’ (Srinavasan et al) is a serious charge.

However, the case for the trial is also strong. Although Basil and his colleagues suggest that ‘all future trials concerning the efficacy of a medication for acute mania should use an arm with one of the proven medications as a comparator’, regulatory bodies such as the Food and Drug Administration insist on at least two placebo-controlled studies that demonstrate superiority of the index drug over placebo in order to get a licence approved. Although one may criticise the Administration for this requirement, it is scientifically unimpeachable and is a general one for drug treatments. A very similar trial has also been carried out in the USA in which risperidone was also compared with placebo treatment (Reference Hirschfeld, Keck and KramerHirschfeld et al, 2004) (and which should have been disclosed with the paper of Reference Khanna, Vieta and LyonsKhanna et al, 2005). The findings suggest that when risperidone is licensed for the treatment of mania it is possible to argue that both these positive trials represent an advance in patient care. A subsidiary argument, a practical one not always well-received in ethical circles, is that participation in a research study can, and should be, a proper and ethical way of providing good patient care, exemplified by the recent comments of Phillips et al (Reference Phillips, McGorry and Yung2005): ‘the clinical treatment of young people identified as being at high risk of developing a psychotic disorder, particularly the use of neuroleptics, should be provided only in the context of a research trial, where standards of informed consent and monitoring are highest’.

Nevertheless, there remain worries about trials in poorer countries. Ethical committees often do not have the same level of independence as they do elsewhere, financial inducements may lead to covert or overt pressures, and there is even sometimes a nationalistic element (e.g. if country X can recruit 100 patients, we must not recruit fewer than 200). This somewhat macho mentality may be behind comments such as that by Khanna et al (Reference Khanna, Vieta and Lyons2005) that the symptoms of mania in the patients seen were ‘substantially more severe than those of patients with bipolar disorder participating in trials elsewhere’, implying that only countries that can be successful in persuading these ‘difficult’ patients to take part should be chosen.

We note that the Indian Council of Medical Research has now decided to audit clinical trials systematically to ensure that national recommendations are followed (Reference MudurMudur, 2005) and the outcome of this will be followed closely. For our part, we have made changes to our refereeing procedure, and have been asking assessors to examine more closely the ethical aspects of papers that are submitted. We shall also be using our new group of international editors (in the case of India this will be Dr Vikram Patel) to advise on ethics both generally and with regard to specific papers, attempting as much as possible to take account of the need for ‘autonomy, beneficence, non-maleficence and justice… and care ethics’ summarised by Bloch & Green's (Reference Bloch and Green2006) recent paper.