Indiscriminate investigations have adverse effects

EDITOR—The application of evidence based medicine is leading to better treatments by thorough evaluation of treatments based on analyses of risks and benefits. These balance the beneficial clinical gains against the adverse pharmacological and medical effects, using information derived from randomised controlled trials and cost effectiveness studies. In contrast, no such critical approach has been taken for diagnostic tests nor have the consequences and adverse effects of inappropriate investigations been explored. The debate around diagnostic tests has centred largely on minimising the unit costs of the delivery of tests in the light of the enormous increase in the demand for investigations without an obvious and proportionate improvement in health status.1

The case report by Krishnan et al highlights an adverse effect of an inappropriate investigation in a woman with hypothyroid induced ascites.2 The published literature is clear that ascites, and any serous effusion of any aetiology, is associated with raised CA125 concentration.3 Yet despite this evidence, the interpretation of a false positive result triggered a number of adverse effects and consequences—namely, a clinical consultation by an oncologist, computed tomography of the abdomen, diagnostic laparoscopy, mammography, and oral gastroduodenoscopy. These inappropriate secondary investigations carry considerable physical, emotional, and financial cost.

What can we do to improve the appropriate use of laboratory and radiological investigations? Previous attempts at educating clinical staff have shown only short lived improvements.4 We need better solutions because there is a vicious amplification cycle in which increases in investigations are mirrored by increases in operative procedures,5 justified on the basis of the investigations which themselves generate investigations. This increase in test volume increases the probability of error and harm to patients. The discipline of evidence based diagnostics may not exist because we do not know what questions to ask in relation to investigation strategies or because there are no hard end points (such as death or cure) to judge success as in pharmacological studies. That should not be an excuse to ignore a significant problem. Where the definition of a disease is made by laboratory and radiological investigations, it is mandatory that the error rate and interferences in the tests are recognised.