When becomes more in fact less?

The principles ‘in dubio abstine’ and ‘primum non nocere’ are as least as old as the language they are written in, and have been made very concrete in criticisms on iatrogenesis, medicalization, polypharmacy and polypragmasy. Clinical decision analysis [1], health technology assessment, and critical appraisal of clinical evidence [2] have provided useful methods to translate these principles in better health care by avoiding useless and potentially harmful intervention. To day, these principles are also applied in research practice, as research resources are scarce, and superfluous research is not only a waste but may also push valuable research aside[3e5]. Moreover, useless research implies exposing research subjects to the burden and sometimes risk of being studied without justification. We must therefore also make progress in designing methods to recognize and avoid research studies without added value. These issues are addressed in various contributions. Stalpers et al. give an update of methods that can help clinicians to reduce diagnostic testing that has no added value and may even be harmful, and further elaborated the concepts ‘threshold approach to clinical decision making’ [6,7] and ‘indication area of a diagnostic test’[8]. They developed a decision analytical model in which ‘testing’ and ‘no testing’ were compared, in order to identify the indication area of prior probabilities wherein use of a diagnostic test results in a net gain. They illustrate their approach, using a freely available website-based calculator with graphical representation developed for his purpose, with various clinical examples. In a second paper, Stalpers et al. investigate whether indication area and maximum diagnostic gain are robust measures in relation to test dependence [9], alternative physician’s heuristics, and varying patient’s utilities. Based on mathematical, decision heuristic, and utility analysis and again supported by examplesethe authors conclude that indication area and the maximum diagnostic gain are indeed robust measures of test performance. In research decision making, a key question is whether adding a new study to the already available evidence, would change conclusions. Chevance c.s. investigated whether robustness of an existing meta-analysis can justify decisions on whether to conduct an additional study on the same research question. They present a contour plot assessment of meta-analysis, using the example of the potential impact of an additional study on the effect of statin use