Table 2.
Performance of the text mining algorithm for diagnosis
| Outcome | n |
|---|---|
| Complete match | 71 |
| Algorithm reported overlapping diagnosis of LLD/CLE in context of SLE, whereas immunologist did not report overlap | 9 |
| Algorithm missed an overlapping diagnosis | 4 a |
| Algorithm assigned an extra overlapping diagnosis | 4 a |
| Diagnosis wrongly assigned by algorithm | 5 |
| Diagnosis missed by algorithm | 8 |
Comparison of 100 randomly selected patients: diagnosis assessed by text mining algorithm compared to assessment of a clinical immunologist.
Abbreviations: CLE, cutaneous lupus erythematosus; LLD, lupus‐like disease; SLE, systemic lupus erythematosus.
In one case, the text mining algorithm missed one overlapping diagnosis and misdiagnosed another overlapping diagnosis. This case is represented in both rows.