Editorial policy for biomedical journals increasingly calls on authors to affirm that approval was received from an Institutional Review Board (IRB) (or equivalent) prior to initiating any human subjects research presented in a manuscript submitted for publication. For most investigations and investigators this does not present any problem. However, when research is carried out in a setting where there is no IRB, should editors consider a report of such a study based on the merits alone? There is no simple answer to this question. This commentary explores aspects of the question and presents issues to be considered in developing an answer. Am. J. Ind. Med. 52:172-175, 2009. (c) 2008 Wiley-Liss, Inc.

Systems used in complex, multi-agency environments have a number of inherent problems and challenges that in many cases, lead to systems failure. The area of system failure has been the subject of extensive research in the past. It has been well documented in a number of places that information systems are difficult to build and are prone to failure. This paper examines system failure from a social perspective in the context of building systems in complex environments. Through a case study modelling approach, a framework has been developed to assist in building information systems in such complex multi-agency environments. The paper introduces this framework through a study of a number of challenges confronting the builder of information systems in complex, social, multi-agency environments. We examine the role of task accountability and problems identifying authority as key reasons why systems are rejected by users. We introduce a task accountability model to help understand this problem. We conclude by identifying work still to be carried out within the study.