Table 3.
Number of cases identified by natural language processing (NLP) in the base study population (n=53,585).
| Natural language processing–identified cases | n | %a (n=53,585) | %b (n=46,086) | ||||
| Shoulder injury identified | 46,086 | 86 | —c | ||||
| Anatomic site | |||||||
|
|
Laterality identified | 44,488 | 83 | 96.5 | |||
|
|
Laterality mismatch | 1220 | 2.3 | 2.6 | |||
| Causality | |||||||
|
|
Cause identifiedd | 25,325 | 47.3 | 55.0 | |||
|
|
Cause identifiede | 19,039 | 35.5 | 41.3 | |||
| Temporality | |||||||
|
|
Onset identified | 45,252 | 84.4 | 98.2 | |||
|
|
Symptom duration >30 days postvaccination | 35,135 | 65.6 | 76.2 | |||
| SIRVAf cases | 467 | 0.9 | 1 | ||||
aPercentage of cases among the number of cases with shoulder injury diagnosis code (n=53,585).
bPercentage of cases among the number of natural language processing–identified shoulder injury cases (n=46,086).
cNot applicable.
dIncludes unknown cause stated in the clinical notes.
eExcludes unknown cause stated in the clinical notes.
fSIRVA: shoulder injury related to vaccine administration.