Adam Adika, MS4
Article: Becker, B. A., et al. (2019). A Prospective, Multicenter Evaluation of Point‐of‐care Ultrasound for Small‐bowel Obstruction in the Emergency Department. Academic Emergency Medicine.
Idea:
300,000 adults hospitalized annually for SBO in US
H&P is unreliable, so imaging is important à but X-rays are increasingly known as horrible way to diagnose as CT is much more sensitive and specific à however, expensive, can increase throughput times, ionizing radiation à what about SBO: cheap and very fast
*case series in 1970s showed that it can sensitive and specific
*meta-analysis of POCUS for SBO promising (average sensitivity and specificity of 90% and 91%, respectively, with incidence of 40%), but only two studies were studies based on emergency medicine physicians (Jang et al in the Emergency Medicine Journal and Unluer et al in the European Journal of Emergency Medicine)
Study
Prospective, multicenter, observational study examining accuracy of POCUS for suspected SBO 217 patients – July 2014 to May 2017, overall SBO prevalence 42.9%
Inclusion criteria: 18+, consentable in English, no prior imaging, symptoms concerning for SBO (not specifically defined)
Exclusion criteria: if ultimately didn’t receive CT, pregnant
Overall study flow: Initial POCUS performed at bedside by physician blinded to labs/imaging/CT, then respectively reviewed by “expert” EM ultrasound fellowship-trained physician à then compared with CT
*The initial POCUS performed was either by an attending ED physician, current EM US fellow, or a PGY2/3 EM resident (must have performed at least 50 POCUS exams, and finished 1 US rotation) – these individuals were given a 30-min lecture on SBO + hands-on practice on a normal, non-SBO individual
5 specific POCUS parameters were defined for SBO: SBO dilation > 25 mm; abnormal peristalsis (to-and-fro of contents), transition point, intraperitoneal free fluid (anechoic collections extraluminally b/w bowel loops), bowel wall edema (keyboard sign – plicae circulares projected into the bowel lumen)
Primary outcome: POCUS-mediated diagnosis of SBO, confirmed by CT
Secondary outcome: SBO diagnosis by expert POCUS reviewer, and diagnostic accuracy of each of the 5 specific US parameters
*any indeterminate/equivocal POCUS or CT interpretation was considered positive for SBO à this is typical ED practice b/c unequivocal studies are usually followed by further workup
Results
SBO incidence = 42.9%
For diagnosis of SBO, sensitivity, specificity, positive likelihood ratio, and negative likelihood ratio were all calculated for the physicians performing the “initial ultrasound”
Sensitivity = 0.88 (95% CI = 0.80 to 0.94)
Specificity = 0.54 (0.45 to 0.63)
Positive likelihood ratio = 1.92 (1.56 to 2.35)
Negative likelihood ratio = 0.22 (0.12 to 0.39)
Stratified by level of education:
-Residents/fellows: sensitivity of 0.91, specificity of 0.51
-Attendings: sensitivity of 0.85, specificity of 0.61
Data for the expert reviewer:
Sensitivity = 0.89 (0.81 to 0.95)
Specificity = 0.82 (0.74 to 0.88)
Accuracy = 0.85 (0.80 to 0.90) compared to 0.69 (0.62 to 0.75)
Percentage agreement b/w initial POCUS and expert reviewer = 68.1%
More sensitive parameters: SBO dilation > 25 mm; abnormal peristalsis
More specific parameters: transition point, intraperitoneal free fluid, bowel wall edema
*However, all parameters were significantly correlated
Limitations: all 3 centers did not routinely use POCUS for SBO evaluation prior to this study à this may have heightened the gap b/w experienced (i.e. fellowship-trained) and inexperienced (non-fellowship trained providers)
-1 of the 3 centers only sampled two patients
-Reviewers had an indeterminate option – so initially I thought that could have accounted for the low specificity of 54% à however, when these were excluded, specificity only increased to 66% (still very low compared to other studies)
-Convenience sampling
Takeaways
-Within this study, residents/fellows and attendings seemed to have similar performance à attendings had higher specify but slightly lower sensitivity
-US fellowship-trained physicians had dramatically better performance than non-fellowship trained, especially in specificity and accuracy
-patterns -- Is it just more experience and exposure that these fellowship-trained physicians have? The added education? Both?
-in any case, this contemplation further strengthens the assertion that this fellowship is extremely important, even when compared to EM physicians that have prior POCUS experience à it adds to the argument that more fellowship positions should be created albeit based on this one specific disease process. It also supports the general trend in medicine to incorporate ultrasound training to those at all levels of medicine, even medical students such as us
-Furthermore, this study had significantly lower sensitivity and specificity than the 2 other SBO POCUS ED studies that were mentioned earlier (0.91, 0.98) and (0.84, 9.93), respectively with a very similar prevalence (0.84 and 0.93) and CT confirmation
-Jang et al – each had to be signed off with 5 SBO-+ POCUS exams
-Unluer et al – 6 hours of training
-This study: even though the physicians had prior experience of at least 50 POCUS exams and finished 1 US rotation, they only received 30 minutes of training and the hands-on practice was on a healthy, non-SBO individual
Further ideas
-Since SBO is pretty common in the US, might want to explore this education aspect further à maybe repeat the same study but have one group that has bolstered education/experience similar to the prior 2 trials (e.g. being signed off on a certain number of SBO-positive cases and/or more hours of training)
-Perform studies on other disease processes and gauge if a similar phenomenon exists b/w fellowship-trained and non-fellowship trained physicians exists
-Greater collaboration b/w fellowship-trained and non-fellowship trained physicians
-Greater requirements for education by hospitals for physicians who use POCUS but are not fellowship trained
-If the above directives were initiated and were successful in improving performance, namely sensitivity, this could possibly reduce CT use down the road in terms of SBO diagnoses à as such, reducing health care expenditures, decreasing throughput times, and reducing patient exposure to ionizing radiation