Table 4 Characteristics of ultrasound skill assessment and overall findings for 23 included studies
From: Ensuring competency in focused cardiac ultrasound: a systematic review of training programs
| Â | SUBJECTS ASSESSED | Â | FINDINGS | |||
|---|---|---|---|---|---|---|
Views | Pathology Assessed | Comparator | ||||
Alexander 2004 [36] | 533 patients (mixed: ICU, intermediate care unit, clinic) | PLAX, PSAX, A4 (+ color doppler) | LV function (EF), MR, aortic valve mobility, pericardial effusion | TTE performed by experienced sonographer and interpreted by echocardiographer, as assessed by experienced observer | Acquisition: learners obtained satisfactory images to assess LV function in 98.7% of patients, MR in 92.7%, aortic valvular disease in 89.7%, and pericardial effusions in 97.9% | Interpretation: learners showed fair agreement with TTE for identifying MR (κ = 0.31), aortic valvular disease (κ = 0.31) and moderate agreement for pericardial effusion (κ = 0.51) and LV dysfunction (κ = 0.51) |
Beraud 2013 [37] | 5 simulated cases | PLAX, PSAX, A4, SC4, IVC | LV function, RV dilation, RV function, pericardial effusion | Simulator performance by novice medical students and expert sonographers | Acquisition: novices took longer to scan (358 ± 170s) than fellows (136 ± 63s), and fellows took longer than experts (38 ± 6s) | Interpretation: one diagnosis (RV dysfunction) was missed by one fellow (5.5%) + 3 novices (33%); experts were significantly faster (18 ± 7s) than fellows (72 ± 38s), who were significantly faster than novices (185 ± 86s) |
Caronia 2013 [38] | 102 patients (mixed: ICU, intermediate care unit) | PLAX, PSAX, A4, SC4 (+ color doppler) | LV function (EF, WMAs), RV strain or diastolic collapse, septal defects, valvular lesions, thrombi, aneurysm, pericardial effusion | Findings on TTE performed by sonographer and interpreted by cardiologist, as assessed by a trained fellow | Interpretation: learners correctly identified 95% of patients with systolic dysfunction (κ = 0.67), 85% of patients with pericardial effusion (κ = 0.60), but only 41% patients with RV strain (κ = 0.38); valvular pathologies were identified with moderate agreement (κ = 0.50 - 0.52) as were WMAs (κ = 0.49) | |
Carrie 2015 [39] | 180 ED patients | PLAX, PSAX, A4, SC4, IVC, lung, abdomen | LV size (LVH) + function, RV dilation, IVC size, pericardial effusion, pleural effusion, consolidations, interstitial fluid, abdominal pathology | Focused exam performed by experienced physician board-certified in ultrasound or echocardiography | Acquisition: learners took longer to perform exams than experts, but exam time decreased from 14.5 min (0-10 exams performed) to 10 min (10-20 exams) to 8 min (20-30 exams), while experts required 4-6 min for exams | Interpretation: agreement with experts improved when comparing learner performance over the first 10 exams performed (exams 0-10) with the last 10 exams performed (exams 20-30) for assessment of LV function (κ=0.77 to κ=0.92), LVH (κ=0.67 to κ=0.9), IVC dilation (κ=0.6 to κ=0.88) and collapse (κ=0.3 to κ=0.76), while agreement peaked after completion of 10-20 exams for assessing RV dilation (κ=0.78 to κ=0.83) and pericardial effusion (κ=0.74 to κ=0.9) |
Chisholm 2013 [40] | 1 healthy volunteer | PLAX, PSAX, A4, SC4, IVC | none | Evaluation of images by cardiologist board-certified in echocardiography | Acquisition: 85% were able to achieve acceptable PLAX + PSAX views within 120s, 70% and 57% were able to achieve SC4 and A4 views (respectively), and <50% obtained IVC view; SC4 and A4 views were more likely to be obtained by those who completed > 45 practice studies | |
Croft 2006 [41] | 72 clinic patients | PLAX, PSAX, A4, A2 (+ color doppler) | LV size + function (WMAs, LVH), valvular lesions, pericardial effusion | Focused exam performed by a level 3 certified echocardiographer | Acquisition: learners obtained diagnostic images in 94% of patients; A2 view was most difficult and was imaged adequately in 68% of patients while PLAX, PSAX, and A4 were imaged adequately in 96%, 92%, and 94%, respectively | Interpretation: learners correctly identified 93% of major findings (PPV = 93%, NPV = 99%) and 78% of minor findings (PPV = 97%, NPV = 93%) |
Farsi 2017 [42] | 205 ED patients | PLAX, PSAX, A4, SC4 | LV function (EF by EPSS or Quinones equation, WMAs), RV dilation, pericardial effusion | Findings on TTE performed by a cardiologist | Interpretation: all major pathologies were identified with >90% accuracy and with near perfect agreement to cardiologists, including low LVEF (κ=0.85), WMAs (κ=0.83), RV dilation (κ=0.86) and pressure overload (κ=1.00), + pericardial effusion (κ=0.83) | |
Ferrada 2011 [43] | 51 ICU patients | PLAX, PSAX, A4, SC4, IVC | LV function, IVC size, pericardial effusion | TTE performed and interpreted by a cardiologist | Acquisition: the A4 was the only view that could not be obtained in all patients (84.3%) | Interpretation: there was 100% correlation between learners and cardiologists on global heart function and contractility |
Gaudet 2016 [44] | 36 ICU patients | PLAX, PSAX, A4, SC4, IVC (+ M-mode) | none | Performance on 1st and 2nd exams, 10th and 11th exams, and 19th and 20th exams as assessed by a level 3 certified intensivist | Efficiency: efficiency improved incrementally after the first 10 studies (1.55 to 2.48) and by a greater extent than after 10 to 20 studies (2.48 to 2.61); efficiency was lower for trainees compared to experts at all intervals | Workload: mental and physical demand, time, effort, frustration, and anxiety decreased throughout all assessment intervals, with the greatest reduction after completing the first 10 studies |
Hellmann 2005 [45] | 229 floor patients | PLAX, PSAX, A4, A2 (+ color doppler) | LV size (thickness) + function, septal thickness, LA size, valvular lesions, pericardial effusion, aortic size | Review of TTE and focused exam by an experienced, level 3 certified cardiologist | Acquisition: image acquisition improved over time as more scans were completed (up to 22 scans); the A2 view had the slowest rate of learning while the PSAX was learned at the fastest rate | Interpretation: interpretation accuracy improved over time when performing up to 22 scans; measurement of LV diastolic size and identification of pericardial effusions were learned fastest while identifying AS, MR, and measuring septal wall thickness had the slowest rate of learning |
Johnson 2016 [46] | 178 patients (mixed: ICU, intermediate care unit, floor) | PLAX, PSAX, A4 or A5, A2, ALAX, SC4 | LV function | TTE performed by trained sonographer and interpreted by a level 2 or 3 certified cardiologist | Acquisition: 100% of learner-performed exams were adequate to characterize LV systolic function | Interpretation: learners identified impaired LV function with substantial agreement (κ=0.77, sens=0.91, spec=0.88) with experts, similar to the interobserver variability among echocardiographers within the study institution (κ=0.78); learners had the lowest sens and spec for identifying mild/moderate LV dysfunction (sens=0.70, spec=0.86) |
Labbe 2016 [47] | 115 ICU patients | PLAX, PSAX, A4, SC4, IVC (+ color, pw, cw doppler) | LV function (EF, WMAs, LV outflow tract velocity time integral, filling pressure, E/A + E/e' ratio), AV, MV, RV dilation, IVC size, pericardial effusion | Focused exam performed by a level 3 certified cardiologist, as assessed by two independent cardiologists | Acquisition: learners in the shorter training group obtained at least one optimal view in 83% of exams, compared to 91% by those who received additional training, while experts obtained at least one optimal view in 98% and 94% of exams; no difference in exam duration between the 2 learner groups (22±8 min, 22±10 min, respectively), but both groups took longer than experts to perform exams (12±6 min, 13±7 min) | Interpretation: learners in the shorter training group had more unanswered questions than experts (13% vs 7%), while those with additional training had a similar # of unanswered questions as experts (9% vs 8.5%); those who received additional training interpreted RV dilation, valvular pathology, and calculated aortic peak velocity, E/A ratio, and E/e' ratio with greater accuracy than those who received shorter training |
108 ICU patients | Above plus septal deviation | |||||
Lucas 2009 [48] | 314 patients (mixed: ICU, floor, short stay unit) | PLAX, PSAX, A4, A2, IVC | LV size + function (LVH), MR, LA size, IVC size, pericardial effusion | TTE performed by experienced sonographers and interpreted by a level 2 echocardiograpy certified cardiologist | Acquisition: leaners were able to obtain adequate imaging on 94-98% of intended assessments | Interpretation: learners were best at identifying severe MR (sens = 100%, spec = 83%), pericardial effusions (sens = 100%, spec = 95%), LA dilation (sens = 90%, spec = 74%), LV dysfunction (sens = 85%, spec = 88%), and less skilled at identifying LVH (sens = 70%, spec = 73%) and dilated IVC (sens = 56%, spec = 86%) |
Manasia 2005 [49] | 90 ICU patients | PLAX, PSAX, A2, A4 | LV function (WMAs), pericardial effusion | Focused exam performed by cardiologist, and assessment of learners’ images by a cardiologist | Acquisition: learners were able to successfully perform a TTE exam in 94% of patients | Interpretation: 84% of exams were correctly interpreted by learners |
Martin 2007 [50] | 354 floor patients | PLAX, PSAX, A4, SC4 (+ color, pw doppler) | LV size (wall thickness) + function (E/A ratio), LA size, valvular lesions, vegetations, pericardial effusion, aortic size | Image quality on focused TTE performed by echocardiography technician, assessed by an independent cardiologist and compared against TTE; interpretation of pre-recorded focused exams by cardiology fellows | Acquisition: learners obtained less optimal images than technicians across all views (80.6% agreement with TTE image quality vs 98.9% agreement), particularly on the A4 view (difference of 33.6%) | Interpretation: learner measurements were less accurate than technicians (80.2% vs 89.9% agreement); learners were best at assessing aortic root size + diastolic LV size (differences of 5.6% and 6.5%, respectively) and least accurate for systolic LV size (13.1% difference); cardiology fellows were no better than learners at interpreting E/A ratio, MR, and AS but were more skilled at interpreting LV function, AR, pericardial effusions, and vegetations |
Mjolstad 2013 [51] | 199 floor patients | PLAX, PSAX, A4, A2, ALAX, IVC, lung (+ color doppler) | LV function (WMAs), RV function (systolic excursion), RV dilation, septal flattening, valvular lesions, LA size, IVC size, pericardial effusion, aortic size | TTE performed by cardiologist or to findings on computed tomography imaging for detection of pleural effusion and aortic size | Acquisition: learners were able to assess LV function, pericardial effusion, + pleural effusions in >95% of patients, were able to assess RV function, LA size, MV, + AV in > 85% of patients, TV + IVC in > 75% of patients, and abdominal aorta + PV in only 50% and 49% of patients, respectively | Interpretation: learners interpreted LV function, pleural effusions, and pericardial effusions with very strong correlation with experts (r ≥ 0.83); AS, AR, + aortic aneurysms with strong correlation (r ≥ 0.67); WMAs, LA dilation, IVC size, TR, and MR with moderate correlation (r ≥ 0.53) |
Mozzini 2014 [52] | 15 floor patients | PLAX, PSAX, A4, SC4, IVC, supra-sternal | LV size + function, RV dilation, IVC size, pericardial effusion | TTE performed by sonographer and interpreted by either a cardiologist or a hospital certified in echocardiography | Acquisition: students who received 18 hours of training were more skilled at obtaining parasternal and apical views compared to those who received only 9 hours; learners required longer (7±1 min) to perform each exam during the first 3 days compared to the second 3 days (4±0.5 min) | Interpretation: students who received 18 hrs of training were more skilled at interpretation compared to those who received 9 hrs; after 18 hrs learners showed substantial agreement with experts on identifying pericardial effusions (κ=0.71), global LV function (κ=0.2 to 0.77), + atrial size (κ=0.66), moderate agreement for LV (κ=0.54) + RV (κ=0.56) enlargement and valvular pathology (κ=0.56), and fair agreement for IVC size (κ=0.35), WMAs (κ=0.35), + aortic size (κ=0.28) |
30 floor patients | Above + M-mode, color doppler | LV size + function (EF by Simpson's method, WMAs, MAPSE), RV size + function, valvular lesions, IVC size, pericardial effusion, aortic size | ||||
Ruddox 2013 [53] | 303 ICU or ED patients | A4, A2, ALAX | LV size + function (EF, WMAs), RV function, LA size, valvular lesions, aortic dilation, IVC size, pericardial effusion | Findings on focused exam performed by a level 3 certified echocardiographer | Interpretation: learners identified LVEF < 40% (κ = 0.53), LV dilation (κ = 0.43), WMAs (κ = 0.64), pericardial effusion (κ = 0.4), valvular abnormalities (avg κ = 0.43), + dilated IVC (κ = 0.17); overall accuracy improved over time from 0-9 exams performed (κ = 0.22) to 20-29 exams performed (κ = 0.38), and only marginally after 30-35 exams (κ= 0.41) | |
Ruddox 2017 [54] | 60 floor patients | PLAX, A4 (+ color doppler) | LV size + function (EF, WMAs), MR, AR, aortic dilation, pericardial effusion | Findings on focused exam performed by a level 3 certified echocardiographer | Interpretation: learners identified LVEF < 40% (κ = 0.7), LV dilation (κ = 0.75), and LA dilation (κ = 0.66) with substantial agreement with TTE, RV pathology (κ = 0.42-0.48), MR (κ = 0.56), and WMAs (κ = 0.44) with moderate agreement, and pericardial effusion (κ = 0.30) + AR (κ = 0.35) with fair agreement | |
See 2014 [55] | 318 ICU patients | PLAX, PSAX, A4, SC4, IVC (+ M-mode, color doppler) | LV function (EF by Simpson's method), RV dilation, MR, IVC size, pericardial effusion | Review of images by an intensivist experienced in critical care ultrasound | Acquisition: views obtained improved from 40% acceptable after 1-10 exams to 91% acceptable after at least 30 exams; learners took 21.3 ± 9.5 min for each exam and 18.9 ± 7 min for each after the first 30 studies; scanning duration decreased by 0.14 min (CI 0.10 - 0.18 min) after each successive study | Interpretation: pathologies were accurately assessed in > 80% after 11-20 exams, and > 90% after 30 exams for all pathologies except for estimation of LVEF fraction (85% after 30 exams) |
Smith 2018 [56] | 3 standardized patients | PLAX, PSAX, A4, SC4, IVC | none | Focused exam performed by a level 1 certified cardiology fellows and by experienced sonographers, interpreted by cardiologists (2 independent evaluators) | Acquisition: learners obtained comparable quality images as fellows for PLAX and SC4 views, but fellows obtained higher quality images for PSAX, A4, and IVC views; learners and fellows exam durations were similar (15.3 min and 13.8 min, respectively) | Efficiency: there was no difference in efficiency between learners and senior cardiology fellows, but experts consistently performed better than cardiology fellows and learners |
Vignon 2010 [57] | 201 ICU patients | PLAX, PSAX, A4, SC4, IVC | LV size + function (EF, WMAs), RV size + function, IVC size, pericardial effusion | Focused exam performed by an experienced, level 3 certified intensivist | Acquisition: leaners performed longer exams than experts (7±2.5 min vs 3±1 min) and obtained fewer views (82% vs 88%); image quality was significantly better when performed by experts for the PSAX view and similar (good to excellent) for all others | Interpretation: learners had perfect agreement with experts in identifying tamponade (κ=1), near-perfect agreement when interpreting LV function (κ=0.84) and LV dilation (κ=0.90), and substantial agreement for identifying RV dilation (κ=0.76), IVC dilation (κ=0.79) and collapse (κ=0.66), + pericardial effusion (κ=0.79) |
Yan 2010 [58] | 107 patients (mixed: floor, clinic) | PLAX, PSAX, A4, SC4 (+ color doppler) | Valvular lesions | Findings on focused exam performed by trained nurse or cardiologist | Interpretation: learners had moderate agreement (κ=0.45) with experts when identifying valvular findings and were most skilled at identifying MS, MR, and AS, and substantially less skilled at identifying AR (κ=0.23) | |