Selecting new health technologies for evaluation: Can clinical experts predict which new anticancer drugs will impact Danish health care?

Several countries have systems in place to support the managed entry of new health technologies. The big challenge for these so-called horizon-scanning systems is to select those technologies that require decision support by means of an early evaluation. Clinical experts are considered a valuable source of information on new health technologies, but research on the relevance of their input is scarce. In 2000, we asked six Danish expert oncologists to predict whether a sample of 19 new anticancer drugs would impact Danish health care over the next 5 years. In 2005, we assessed the accuracy of these predictions in a delayed type cross-sectional study. The specificity of the Danish experts' prediction was 1 (95% confidence interval 0.74-1.00) and the sensitivity was 0.63 (0.31-0.86). The negative predictive value was 0.79 (0.52-0.92) and the positive predictive value was 1 (0.57-1.00). This indicates that clinical experts have the ability to predict which new anticancer drugs are unlikely to have an impact. This information can be used to increase the efficiency of selecting new technologies for evaluation. As the experts missed 37% of drugs that are in need of guidance, they should not be relied upon to select drugs relevant for evaluation. (c) 2006 Elsevier Ltd. All rights reserved.