Morgan Schellenberg

@nieman_dylan That's a great thought. We retained those patients because we wanted our baseline population broadly defined by gunshot wounds to the liver. However, your point is very well taken that it may have been tighter to only include stable patients. #EASTjc

@nieman_dylan @JenConeMD For sure. Despite adjusting for potential confounders, the likelihood of selection bias is high. What we hoped to convey with the paper was not that NOM is always the right choice, but that in carefully selected patients, it can reduce mortality and complications. #EASTjc

@runwithscalpels @nieman_dylan @SchellenbergM As for AE data, there are procedure codes but they are not associated with specific injuries (e.g. AE for pelvic fracture bleeding is indistinguishable from AE for liver injury). Sadly a limitation of registry-based research. We could do a multicenter study in the future! #EASTjc

@runwithscalpels @nieman_dylan @SchellenbergM These are excellent questions that unfortunately are unanswered by the dataset. With the NTDB, there is no way to conclusively determine the contribution of each injury on outcomes, although the incidence of associated solid organ injury was low in this population. #EASTjc

@nieman_dylan Agreed, this is such an important question! There were >300 centers captured and each contributed <3% of the patients. We examined outcomes by level of trauma center (I vs. II-IV) and found no differences in failure of SNOM, complications, or mortality on multivariate analysis.

@nieman_dylan @SchellenbergM Absolutely! One of the important pieces of SNOM is quantifying that "eyeball test" - i.e. patients with a GSW through the RUQ with hemodynamic instability, peritonitis, evisceration, and those who are unevaluable go immediately to the OR and are not candidates for SNOM. #EASTjc