Reducing error, improving safety

[...]the need to take a British study beyond the pilot phase may not be supported. 4 Be that as it may, an important issue was not addressed in the BMJ. When several units can be compared over time, a benchmark might be established for the optimal level of reporting, which will never be the lowest. [...]a benchmark is established, anybody publishing such information should explain prominently in the introduction that there is no standard yet, and the contrary effects on the apparent frequency of incidents of honest reporting and good practice mean that no league tables can be construed.