Decision analysis and the implementation of research findings
BMJ 1998; 317 doi: https://doi.org/10.1136/bmj.317.7155.405 (Published 08 August 1998) Cite this as: BMJ 1998;317:405All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Lilford et al claim decision analysis reconciles evidence based medicine with patients' preferences 1. This it does, but at the cost of equating the myriad human values that are patients' preferences, with an arbitrary numerical value that represents the trade off between the disadvantages of one (medically chosen) outcome and the advantages of another. The assumption is that the clinician knows what the important trade off for the patient will be. Decision analysis only allows a decision to be rational from a medical perspective, which may not be a rational decision from the patient's perspective 2; 3. As Dostoyevsky knew, people are so very different, and a respect for their autonomy means respecting an approach that may not always appear rational. This alternative approach is what Keats called "Negative Capability, that is, when a man is capable of being in uncertainties, mysteries, doubts, without any irritable reaching after fact and reason" 4. The uncertainties, mysteries and doubts that are human values can be replaced by a number between 1.0 and 0, but we risk losing much in the translation.
References
1. Lilford RJ, Pauker SG, Braunholtz DA, Chard J. Decision analysis and the implementation of research findings. BMJ 1998;317:405-9.
2. Donovan JL, Blake DR. Patient non-compliance: deviance or reasoned decision-making? Soc Sci Med 1992;34:507-513.
3. Brett AS. Ethical issues in risk factor intervention. Am J Med 1984;76:557-561.
4. Forman M. The letters of John Keats. London, Oxford University Press, 1931; p77. Quoted in: Dowie J, Elstein A. Professional judgement: a reader in clinical decision making. Cambridge: Cambridge UP, 1988, p562.
Competing interests: No competing interests
Re: Decision analysis and the implementation of research findings
Editor
Decision analysis has a role in the evaluation of diagnostic tests as well as in the choice of treatment. An example is the evaluation of clinicians and computer programs interpreting the ECG. While there are numerous statements for interpreting the ECG, the only ones that can be validated by non-electrocardiographic means are left and right ventricular hypertrophy, myocardial infarction, combinations of these three, and the absence of any of them ('normal').
With more than two diagnostic categories, those interested in the performance of clinicians or computer programs in interpreting the ECG will want to know firstly how well abnormal is differentiated from normal (if the true diagnosis is left ventricular hypertrophy (LVH) but the interpretation is myocardial infarction (MI), the interpretation of abnormal is correct), and secondly how well the abnormalities are diagnosed (if the true diagnosis is LVH but the interpretation is MI, the interpretation of abnormal is incorrect). It is in the second instance that utilities have an important role: is a missed myocardial infarction a more serious error than a false positive prediction of myocardial infarction? what credit is given to a diagnosis of LVH when the patient has RVH as well as LVH ? is twice the credit given if a correct diagnosis of biventricular hypertrophy (BVH) is made? Using the patient as the functional unit such that no interpretation could score more than 3 points per patient, the utilities developed in reference 2 were:
True Predicted
LVH RVH MI Other BVH LVHMI BVHMI LVH +3 -3 -3 -3 0 0 -1 RVH -3 +3 -3 -3 0 -3 -1 MI -3 -3 +3 -3 -3 0 -1 Normal -3 -3 -3 +3 -3 -3 -3 BVH 0 0 -3 -3 +3 0 +1 LVHMI 0 -3 0 -3 0 +3 +1 BVHMI -1 -1 -1 -3 +1 +1 +3
Normal = normal clinico-pathological data, Other = an ECG interpretation that does not include either LVH or RVH or MI.
While summing the points for the ECG interpretation of each patient is adequate for comparison purposes, to compute the merit of a diagnostic procedure the prevalence of the diagnostic categories in the population under study need to be known. With this information and a table of utilities like that above, it is possible to determine how far the diagnostic procedure is from the ideal.
No conflict of interest
Peter Bourdillon Consultant in Cardiovascular Disease Hammersmith Hospital London W12 0HS pbourdillon@msn.com
Competing interests: No competing interests