The core condition for clinical workplace learning is 'supported participation', the various outcomes of which are mutually reinforcing and also reinforce students' ability to participate in further practice. This synthesis has 2 important implications for contemporary medical education: any reduction in medical students' participation in clinical practice that results from the patient safety agenda and expanded numbers of medical students is likely to have an adverse effect on learning, and the construct of 'self-directed learning', which our respondents too often found synonymous with 'lack of support', should be applied with very great caution to medical students' learning in clinical workplaces.

This Guide provides an overview of educational theory relevant to learning from experience. It considers experience gained in clinical workplaces from early medical student days through qualification to continuing professional development. Three key assumptions underpin the Guide: learning is 'situated'; it can be viewed either as an individual or a collective process; and the learning relevant to this Guide is triggered by authentic practice-based experiences. We first provide an overview of the guiding principles of experiential learning and significant historical contributions to its development as a theoretical perspective. We then discuss socio-cultural perspectives on experiential learning, highlighting their key tenets and drawing together common threads between theories. The second part of the Guide provides examples of learning from experience in practice to show how theoretical stances apply to clinical workplaces. Early experience, student clerkships and residency training are discussed in turn. We end with a summary of the current state of understanding.

Prescribing errors affect patient safety throughout hospital practice. Previous reviews of studies have often targeted specific populations or settings, or did not adopt a systematic approach to reviewing the literature. Therefore, we set out to systematically review the prevalence, incidence and nature of prescribing errors in hospital inpatients. MEDLINE, EMBASE, CINAHL and International Pharmaceutical Abstracts (all from 1985 to October 2007) were searched for studies of prescriptions for adult or child hospital inpatients giving enough data to calculate an error rate. Electronic prescriptions and errors for single diseases, routes of administration or types of prescribing error were excluded, as were non-English language publications. Median error rate (interquartile range [IQR]) was 7% (2-14%) of medication orders, 52 (8-227) errors per 100 admissions and 24 (6-212) errors per 1000 patient days. Most studies (84%) were conducted in single hospitals and originated from the US or UK (72%). Most errors were intercepted and reported before they caused harm, although two studies reported adverse drug events. Errors were most common with antimicrobials and more common in adults (median 18% of orders [ten studies, IQR 7-25%]) than children (median 4% [six studies, IQR 2-17%]). Incorrect dosage was the most common error. Overall, it is clear that prescribing errors are a common occurrence, affecting 7% of medication orders, 2% of patient days and 50% of hospital admissions. However, the reported rates of prescribing errors varied greatly and this could be partly explained by variations in the definition of a prescribing error, the methods used to collect error data and the setting of the study. Furthermore, a lack of standardization between severity scales prevented any comparison of error severity across studies. Future research should address the wide disparity of data-collection methods and definitions that bedevils comparison of error rates or meta-analysis of different studies.