Is 15 per cent very common? Informing people about the risks of medication side effects

Objectives - To assess the general public's interpretation of the verbal descriptors for side effect frequency recommended for use in medicine information leaflets by a European Union (EU) guideline, and to examine the extent to which differences in interpretation affect people's perception of risk and their judgments of intention to comply with the prescribed treatment. Method - Two studies used a controlled empirical methodology in which people were presented with a hypothetical, but realistic, scenario about visiting their general practitioner and being prescribed medication. They were given an explanation that focused on the side effects of the medicine, together with information about the probability of occurrence using either numerical percentages or the corresponding EU verbal descriptors. Interpretation of the descriptors was assessed. In study 2, participants were also required to make various judgments, including risk to health and intention to comply. Key findings - In both studies, use of the EU recommended descriptors led to significant overestimations of the likelihood of particular side effects occurring. Study 2 further showed that the "overestimation" resulted in significantly increased ratings of perceived severity of side effects and risk to health, as well as significantly reduced ratings of intention to comply, compared with those for people who received the probability information in numerical form. Conclusion - While it is recognised that the current findings require replication in a clinical setting, the European and national authorities should suspend the use of the EU recommended terms until further research is available to allow the use of an evidence-based approach.

Testing Turing's five minutes, parallel-paired imitation game

Purpose – The purpose of this paper is to consider Turing's two tests for machine intelligence: the parallel-paired, three-participants game presented in his 1950 paper, and the “jury-service” one-to-one measure described two years later in a radio broadcast. Both versions were instantiated in practical Turing tests during the 18th Loebner Prize for artificial intelligence hosted at the University of Reading, UK, in October 2008. This involved jury-service tests in the preliminary phase and parallel-paired in the final phase. Design/methodology/approach – Almost 100 test results from the final have been evaluated and this paper reports some intriguing nuances which arose as a result of the unique contest. Findings – In the 2008 competition, Turing's 30 per cent pass rate is not achieved by any machine in the parallel-paired tests but Turing's modified prediction: “at least in a hundred years time” is remembered. Originality/value – The paper presents actual responses from “modern Elizas” to human interrogators during contest dialogues that show considerable improvement in artificial conversational entities (ACE). Unlike their ancestor – Weizenbaum's natural language understanding system – ACE are now able to recall, share information and disclose personal interests.

The heliospheric Hale cycle over the last 300 years and its implications for a “lost” late 18th century solar cycle

A Hale cycle, one complete magnetic cycle of the Sun, spans two complete Schwabe cycles (also referred to as sunspot and, more generally, solar cycles). The approximately 22-year Hale cycle is seen in magnetic polarities of both sunspots and polar fields, as well as in the intensity of galactic cosmic rays reaching Earth, with odd- and even-numbered solar cycles displaying qualitatively different waveforms. Correct numbering of solar cycles also underpins empirical cycle-to-cycle relations which are used as first-order tests of stellar dynamo models. There has been much debate about whether the unusually long solar cycle 4 (SC4), spanning- 1784–1799, was actually two shorter solar cycles combined as a result of poor data coverage in the original Wolf sunspot number record. Indeed, the group sunspot number does show a small increase around 1794–1799 and there is evidence of an increase in the mean latitude of sunspots at this time, suggesting the existence of a cycle ‘‘4b’’. In this study, we use cosmogenic radionuclide data and associated reconstructions of the heliospheric magnetic field (HMF) to show that the Hale cycle has persisted over the last 300 years and that data prior to 1800 are more consistent with cycle 4 being a single long cycle (the ‘‘no SC4b’’ scenario). We also investigate the effect of cycle 4b on the HMF using an open solar flux (OSF) continuity model, in which the OSF source term is related to sunspot number and the OSF loss term is determined by the heliospheric current sheet tilt, assumed to be a simple function of solar cycle phase. The results are surprising; Without SC4b, the HMF shows two distinct peaks in the 1784–1799 interval, while the addition of SC4b removes the secondary peak, as the OSF loss term acts in opposition to the later rise in sunspot number. The timing and magnitude of the main SC4 HMF peak is also significantly changed by the addition of SC4b. These results are compared with the cosmogenic isotope reconstructions of HMF and historical aurora records. These data marginally favour the existence of SC4b (the ‘‘SC4b’’ scenario), though the result is less certain than that based on the persistence of the Hale cycle. Thus while the current uncertainties in the observations preclude any definitive conclusions, the data favour the ‘‘no SC4b’’ scenario. Future improvements to cosmogenic isotope reconstructions of the HMF, through either improved modelling or additional ice cores from well-separated geographic locations, may enable questions of the existence of SC4b and the phase of Hale cycle prior to the Maunder minimum to be settled conclusively.