More (unnecessary and potentially harmful) regulation
David Armstrong, King's College London
2 October 2009
Increasing regulation seems to be a key feature of ‘the audit society’ 1. Undesirable outcomes, from food poisoning to bank insolvency, can be prevented, or so it is argued, by codified rules and monitoring processes. Regulation has also spread rapidly within biomedical research mainly to protect patients from experimental harms and from the dangers of health care interventions based on misleading or erroneous research findings. Yet regulation is not necessarily a benign activity. The ethical regulation of medical research, for example, may at times be so unnecessarily rigid as to block or delay advances which could have significant patient benefit 2.
In recent years there has been growth in the regulation of what can and cannot be reported in medical journals, ostensibly in an attempt to ensure that misleading results – which might be translated into patient harm – are not published. The CONSORT statement, for example, ensures that scientific reports account for all participants in a trial so as to avoid making biased interpretations based on a limited sub-set of the data. The proposed requirement that researchers describe behavioural interventions in detail 3 is another example of the extension of regulatory reach in scientific reporting: but will it have beneficial effects and, more importantly, could it be harmful?
A poorly specified behavioural intervention is highly unlikely to cause direct patient harm as its implementation would be well nigh impossible; reports that ‘spatial counselling' has been shown to be highly effective cannot harm anyone as this finding simply invites the rejoinder: 'Exactly what type of counselling has these effects? '. Of course, those reading the report might find it frustrating that ‘spatial counselling’ was not well described – though presumably the level of description satisfied the paper’s referees as they might have thought the novelty of the idea outweighed its lack of precision – but it is the creativity implied by the demand for more clarity which will be lost once these new rules come into force.
The philosopher Richard Rorty has argued that science is characterised by the literalisation of metaphor 4. New and powerful scientific ideas, he argues, are usually expressed as metaphors (‘natural selection’, ‘stem cells’, ‘attitudes’, ‘cell regulation’) which owe their fertility to the very fact that as metaphors they are poorly defined. In the metaphor ‘no man is an island’ it would be completely facile to demand that the type and size of island was clearly specified; scientific metaphors, however, do become literalised and precisely defined over time but thereby lose their power to incite new interpretations and new lines of development. In similar fashion, if a new intervention such as ‘spatial counselling’ was reported as being effective it would seem desirable, though contrary to the new editorial policy, that it was as weakly specified as possible (presumably just sufficient to get through the review process) so as to leave sufficient ambiguity for later researchers to refine and explore its potential as a basis for pragmatic interventions.
Clearly there are some instances when it is highly desirable (such as a pragmatic trial) that a behavioural intervention is defined sufficiently clearly (though not necessarily to the same level of detail described in the new editorial policy) so as to enable it to be implemented elsewhere should it be found effective. Equally, there seem some situations, particularly when less applied ‘discovery science’ is being reported, when specificity might be harmful; it seems naive to expect new behavioural interventions to be born fully-formed. At the moment we have the good sense of reviewers to decide the appropriate level of specificity, one that balances the risks of various harms. In its place we have a policy diktat from an editorial group which is introducing a new restrictive regime (so much for reviewer’s judgements) which can only mean that many interesting and creative papers have to seek a home elsewhere. This is not what journals should be about; this is not what journals should be doing.
1 Power, M. The Audit Society: rituals of verification. Oxford: Oxford University Press, 1997. 2 Glasziou P, Chalmers I. Ethics review roulette: what can we learn? BMJ 2004;328: 121-2. 3 Michie, S. Fixsen, D. Grimshaw, JM. Eccles, MP. Specifying and reporting complex behaviour change interventions: the need for a scientific method. Implementation Science 2009, 4:40. 4 Rorty, R. Contingency, Irony, and Solidarity. Cambridge: Cambridge University Press, 1989.
Implementation Science
More (unnecessary and potentially harmful) regulation
2 October 2009
Increasing regulation seems to be a key feature of ‘the audit society’ 1. Undesirable outcomes, from food poisoning to bank insolvency, can be prevented, or so it is argued, by codified rules and monitoring processes. Regulation has also spread rapidly within biomedical research mainly to protect patients from experimental harms and from the dangers of health care interventions based on misleading or erroneous research findings. Yet regulation is not necessarily a benign activity. The ethical regulation of medical research, for example, may at times be so unnecessarily rigid as to block or delay advances which could have significant patient benefit 2.
In recent years there has been growth in the regulation of what can and cannot be reported in medical journals, ostensibly in an attempt to ensure that misleading results – which might be translated into patient harm – are not published. The CONSORT statement, for example, ensures that scientific reports account for all participants in a trial so as to avoid making biased interpretations based on a limited sub-set of the data. The proposed requirement that researchers describe behavioural interventions in detail 3 is another example of the extension of regulatory reach in scientific reporting: but will it have beneficial effects and, more importantly, could it be harmful?
A poorly specified behavioural intervention is highly unlikely to cause direct patient harm as its implementation would be well nigh impossible; reports that ‘spatial counselling' has been shown to be highly effective cannot harm anyone as this finding simply invites the rejoinder: 'Exactly what type of counselling has these effects? '. Of course, those reading the report might find it frustrating that ‘spatial counselling’ was not well described – though presumably the level of description satisfied the paper’s referees as they might have thought the novelty of the idea outweighed its lack of precision – but it is the creativity implied by the demand for more clarity which will be lost once these new rules come into force.
The philosopher Richard Rorty has argued that science is characterised by the literalisation of metaphor 4. New and powerful scientific ideas, he argues, are usually expressed as metaphors (‘natural selection’, ‘stem cells’, ‘attitudes’, ‘cell regulation’) which owe their fertility to the very fact that as metaphors they are poorly defined. In the metaphor ‘no man is an island’ it would be completely facile to demand that the type and size of island was clearly specified; scientific metaphors, however, do become literalised and precisely defined over time but thereby lose their power to incite new interpretations and new lines of development. In similar fashion, if a new intervention such as ‘spatial counselling’ was reported as being effective it would seem desirable, though contrary to the new editorial policy, that it was as weakly specified as possible (presumably just sufficient to get through the review process) so as to leave sufficient ambiguity for later researchers to refine and explore its potential as a basis for pragmatic interventions.
Clearly there are some instances when it is highly desirable (such as a pragmatic trial) that a behavioural intervention is defined sufficiently clearly (though not necessarily to the same level of detail described in the new editorial policy) so as to enable it to be implemented elsewhere should it be found effective. Equally, there seem some situations, particularly when less applied ‘discovery science’ is being reported, when specificity might be harmful; it seems naive to expect new behavioural interventions to be born fully-formed. At the moment we have the good sense of reviewers to decide the appropriate level of specificity, one that balances the risks of various harms. In its place we have a policy diktat from an editorial group which is introducing a new restrictive regime (so much for reviewer’s judgements) which can only mean that many interesting and creative papers have to seek a home elsewhere. This is not what journals should be about; this is not what journals should be doing.
1 Power, M. The Audit Society: rituals of verification. Oxford: Oxford University Press, 1997.
2 Glasziou P, Chalmers I. Ethics review roulette: what can we learn? BMJ 2004;328: 121-2.
3 Michie, S. Fixsen, D. Grimshaw, JM. Eccles, MP. Specifying and reporting complex behaviour change interventions: the need for a scientific method. Implementation Science 2009, 4:40.
4 Rorty, R. Contingency, Irony, and Solidarity. Cambridge: Cambridge University Press, 1989.
Competing interests
None.