Patient safety and safer practices are central themes to many national strategies for accountability. The multinational Quality Indicator Project (QI Project) database is used to identify patterns of indicator use to measure safety of care in Asia, Europe, and the USA. The second objective is to assess, within the context of an indicator project, the usefulness of indicators to measure errors or mishaps. This descriptive study retrospectively analyses indicator use patterns among hospitals in Asia, Europe, and the USA. The QI Project database is used for the 1999-2002 period. Statistical testing (P-value) of the differences in use percentages across five countries is based on 'country' rather than 'hospital' as the unit of analysis. There was a significant increase in overall QI Project indicator use worldwide between 1999 and 2002. The average change in use was 6.8% for safety indicators and 4.2% for all other indicators. When analysed by country (USA, Austria, Belgium, UK, and Taiwan), the average increases in use percentage were highest in Taiwan and Belgium. When the country-specific differences were tested for significance, Taiwan showed the largest (and statistically significant) increase in safety indicator use between 1999 and 2002 (P<0.0001). In the USA, the rates of safety indicator use have decreased (P=0.0502) during the same time period. This paper identifies, perhaps for the first time, how traditional indicators of hospital performance are being used to understand a hospital's performance and associated safety of care. Although the study's time frame is limited to 3 years, the findings seem to suggest that the interest in using these traditional indicators as proxies for safer practice measures is increasing among the QI Project participants worldwide. The challenge of using inherently value-free indicators as indicators of safety (hence de facto labelled as 'error' focused) should be further studied.