Claims for malpractice and medical negligence are a potentially important source of information on the causes of harm to patients and have provided valuable lessons in the past. However today, with many additional sources of information and methods of analysis, the role of claims analysis needs to be reappraised. We consider the role of claims analysis in relation to other methods of studying adverse outcomes, review previous studies of claims and summarize the findings of four recent British specialty claims reviews. Claims analysis has a number of inherent limitations. We suggest that there is now no case for ad hoc claims reviews which rely on data that have been assembled for legal purposes only. Claims review is still potentially useful for rare events or in cases where other sources of data are not available. However, future claims reviews need to meet basic criteria before being undertaken; these include prospective identification of the relevant questions and variables, adequacy and completeness of the data set, availability of expert reviewers and clear protocols for review.