Identifying work related injuries: comparison of methods for interrogating text fields

Table 3 Summary of case identification results using each method

Approach	Number of true cases identified	Number of false positives	Sensitivity	Specificity	PPV
Text Search Approaches
Basic keyword search	12,457	1,916	0.58	0.99	0.87
Index search	13,252	23,434	0.62	0.87	0.36
Keyword OR index	17,004	24,416	0.79	0.87	0.41
Content Analytic Approaches
Binary classification	17,299	13,894	0.81	0.93	0.55
Adjusted probability classification	16,424	8,699	0.77	0.95	0.65

ISSN: 1472-6947