The issues of patient safety and quality of care have gained policy attention with a growing appreciation of the scale and impact of medical injury in health systems. While the focus is clearly on the prevention of iatrogenic injury, the question of patient compensation is now also considered important, if only because in fault-based tort systems the fear of litigation may itself be a barrier to the disclosure and open discussion of medical error. No-fault systems, by contrast, do not require proof of culpability, and thus may both reduce barriers to compensation and increase disclosure of error. Little evidence, however, is available on the performance of such systems. This article reports on the analysis of two data sources-a sample of hospital admissions and a complete set of compensation claims for medical injury. Both are for the same year and region of New Zealand, a country that has maintained a no-fault system of accident compensation for a quarter of a century. Just over 2 percent of hospital admissions were associated with an adverse event that was potentially compensable under scheme criteria. While the claims process was well targeted, the level of claims making and receipt was low, with the ratio of successful claims to potentially compensable events being approximately 1:30. Comparison of social and clinical characteristics of the two data sets revealed a degree of selectivity. Compared with the hospital events, the typical successful claimant was younger and female and was much more likely to have experienced a surgical adverse event that, while unexpected, was not due to substandard care. It is concluded that, in interpreting these results, account needs to be taken of a number of features unique to the New Zealand system. These include: the limited payoff for a compensation claim (no pain and suffering or lump sum, free hospital care); the relative complexity of the grounds for claim (either rarity and severity or practitioner error); and a history of limited litigation for medical error. This suggests that, while the New Zealand system is well targeted, cheap, and free of financial and legal barriers, a change in legal doctrine alone has not in itself been sufficient to remove completely the selective and low level of claims making traditionally associated with patient compensation under tort.