Figure 3

A formula to express the relationship between the parameters of CRR scenarios and their numbers of detected harm incidents, which is associated with the precision and power of estimated harm rates. ^$We specified a three month period of review in each record for the purposes of this study. In our examples this is expressed as 0.25 years. Increasing the review period from three to twelve months would have resulted in a fourfold reduction in the number of records each practice had to review. *The levels of precision and power we selected for the purposes of this study. **The values of rHR, nPrac and nRec are taken from the ‘lines’ in Figures 1 and 2.