Table 1

Similarity, number of confirmed errors, positive predictive value (PPV) and alert yield

Similarity levels	Similarity range	Total cases	Cases reviewed	Confirmed errors	PPV (%)	SE of PPV (%)	Yield
1	0.38–0.42	1110	128	3	2.3	1.3	26
2	0.42–0.46	657	100	4	4.0	2.0	26
3	0.46–0.49	338	57	7	12.3	4.4	42
4	0.49–0.53	171	94	1	1.1	1.1	2
5	0.53–0.57	31	31	2	6.5	3.2	2
6	0.57–0.61	39	39	4	10.3	5.3	4
7	0.61–0.65	8	8	5	62.5	17.1	5
8	0.65–0.68	39	38	25	65.8	7.8	26
9	0.68–0.72	2	2	1	50.0	35.4	1
10	0.72–0.76	9	9	9	100.0	0.0	9
Total	0.38–0.76	2404	506	61	12.1	1.4	290

Higher similarity score indicates greater similarity. The range is 0 to 1.