Background
Postoperative echocardiography after congenital heart disease surgery is of prognostic importance, but variable image quality is problematic. We implemented a quality improvement bundle comprising of focused imaging protocols, procedural sedation, and sonographer education to improve the rate of optimal imaging (OI).
Methods
Predischarge echocardiograms were evaluated in 116 children (median age, 0.51 years; range, 0.01–5.6 years) from two centers after tetralogy of Fallot repair, arterial switch operation, and bidirectional Glenn and Fontan procedures. OI rates were compared between the centers before and after the implementation of a quality improvement bundle at center 1, with center 2 serving as the comparator. Echocardiographic images were independently scored by a single reader from each center, blinded to center and time period. For each echocardiographic variable, quality score was assigned as 0 (not imaged or suboptimally imaged) or 1 (optimally imaged); structures were classified as intra- or extracardiac. The rate of OI was calculated for each variable as the percentage of patients assigned a score of 1.
Results
Intracardiac structures had higher OI than extracardiac structures (81% vs 57%; adjusted odds ratio [OR], 3.47; P < .01). Center 1 improved overall OI from 48% to 73% (OR, 4.44; P < .01), intracardiac OI from 69% to 85% (OR, 3.53; P = .01), and extracardiac OI from 35% to 67% (OR, 5.16; P < .01). There was no temporal difference for center 2.
Conclusions
After congenital heart disease surgery in children, intracardiac structures are imaged more optimally than extracardiac structures. Focused imaging protocols, patient sedation, and sonographer education can improve OI rates.
Highlights
- •
QI methods are useful to improve echocardiographic quality in CHD.
- •
Procedural sedation improved optimal visualization after CHD repairs.
- •
Imaging protocols specific to CHD repairs improved discharge echocardiographic quality.
- •
Extracardiac structures have a lower rate of OI despite intervention.
Optimal image quality in transthoracic echocardiography is affected by patient factors such as patient motion and acoustic windows, technical factors such as equipment and lighting, and operator-specific factors such as experience and imaging technique. Echocardiographic imaging of young children in the postoperative period after cardiac surgery is particularly challenging because of agitation, pain, interference from bandages and drains, and suboptimal acoustic windows due to disruption of tissue planes during surgery. Postoperative echocardiograms are not only important in guiding the immediate management of these children but also may be useful for predicting short- and midterm outcomes. Quality assurance initiatives that have been performed as part of multicenter studies, both pediatric and adult, have demonstrated that image quality can vary by participating center, and factors such as increasing familiarity with prespecified study protocols, sonographer education, and periodic feedback can improve individual centers’ performance.
In 2013, as part of retrospective data collection for an ongoing research study that was based at center 2, we noted significant variability in the rate of optimal imaging (OI) by echocardiography for postoperative patients at center 1, particularly for extracardiac structures such as the branch pulmonary arteries and cavopulmonary anastomoses. The key drivers of this quality problem appeared to be lack of patient cooperation and variable knowledge and experience among the sonographers in imaging the areas that were relevant to the postoperative cardiac anatomy. Measures undertaken to correct the problem included the implementation of a quality improvement (QI) bundle. The primary objective of this study was to determine if this strategy improved the rate of OI (the primary outcome measure) by transthoracic echocardiography in postoperative congenital heart disease (CHD). A secondary objective was to elucidate the rate of OI of specific cardiac structures.
Methods
Four commonly performed CHD repairs—complete repair of tetralogy of Fallot, arterial switch operation (ASO) for d-transposition of the great arteries, the bidirectional Glenn operation, and the Fontan operation—were selected for this study because of their varied complexity and the different echocardiographic imaging techniques required for visualization of the surgical repair. Children undergoing these repairs at center 1 (Children’s Mercy Hospital) during two time periods, July 1 to December 31, 2011 (baseline), and July 1 to December 31, 2013 (after the implementation of a QI bundle), were included. Because there existed no established benchmarks that defined optimal performance in echocardiography of postoperative CHD, a cohort of children matched for the surgical repairs were recruited from center 2 (Boston Children’s Hospital) during the same time periods. We selected center 2 as the comparator because it has long served as the core laboratory in numerous multicenter pediatric studies requiring echocardiographic expertise and had incorporated the components of the proposed QI bundle into clinical routine for a number of years. To account for center differences in surgical volume, children operated on in each month from center 2 were included chronologically (by date of surgery) in the study until study numbers were equal to numbers from center 1 for each month. Figure 1 shows the overall study design, including the number of subjects evaluated from each center after the four specified repairs before and after the implementation of the QI intervention. Study identification numbers were assigned by a random number generator. The institutional review boards of both centers approved the study, and data use agreements were in place.
QI Interventions
The QI bundle was implemented at center 1 between March and June 2013. This comprised of three components: (1) the implementation of prespecified, targeted imaging protocols for each of the surgical repairs; (2) the use of sedation for imaging children (at provider discretion); and (3) sonographer education. Disease-specific protocols were designed to focus on the structures intervened upon during the surgical repair. For example, the protocol for the ASO focused on imaging of the septa, semilunar valves, outflow tracts, supravalvar anastomoses, coronary artery flow, and branch pulmonary arteries from various imaging windows. The protocols specific to the four lesions are detailed in Supplemental Tables 1–4 . Sonographers from center 1 were educated regarding these protocols during biweekly echocardiography laboratory meetings. The protocols were laminated and placed on all ultrasound machines for easy access by the sonographer. An interactive case review conference was held for the sonographers to provide overall feedback regarding performance, including review of echocardiographic examinations in which imaging was optimal and reinforcement of techniques to be used to obtain optimal images. Sonographers were instructed to call the provider responsible for the care of the child to request sedation, and if deemed necessary and safe, the child was given a single dose of oral midazolam (0.5 mg/kg) for anxiolysis, along with oral narcotic pain medication as needed.
Echocardiographic Assessment of Image Quality
Postoperative, predischarge echocardiograms of the children included in the study were selected for review. A single reader from each center (A.P., J.C.L.) independently reviewed all deidentified echocardiograms from both time periods and both centers. The readers were aware of the diagnosis and the operative procedure performed but had no access to the echocardiographic reports that were generated for clinical decision making and were blinded to the center and the time period. For each study, reviewers assessed the quality of imaging for those structures that were part of the imaging protocol. Any structure that was not imaged or was suboptimally imaged was assigned a score of 0 and if imaged optimally received a score of 1. Variables that were scored for quality included two-dimensional, spectral Doppler, and color Doppler data. For two-dimensional imaging, OI was defined as imaging of sufficient quality to allow (1) clear and distinct delineation of the edges of a structure for the reader to perform measurements with confidence and (2) sufficient visualization of the lumen of a structure to rule out a thrombus. For color Doppler imaging, OI was defined as (1) appropriate scale and gain settings to allow visualization of color flow in the structure without excessive color bleed and (2) ability of the reader to discern laminar versus nonlaminar flow patterns. For spectral Doppler imaging, an image was judged to be of adequate quality if every effort was made to align the plane of Doppler interrogation parallel to the direction of flow, and appropriate adjustments to gain, baseline, and scale settings were made to decrease noise such that a distinct Doppler flow profile could be captured. Structures were classified as intracardiac or extracardiac ( Table 1 ). Figures 2 A–2C show examples of echocardiographic images that were assigned scores of 0 and 1. Twenty-five percent of the studies (six ASO, nine tetralogy of Fallot, four Glenn, and nine Fontan) were examined for intrareader reproducibility. The readers each reassigned quality scores to the same images after ≥4 weeks had elapsed since the initial read and while blinded to their initial reads.
| Intracardiac structures | Extracardiac structures |
|---|---|
| Atrial septum Ventricular septum Right ventricular outflow tract Left ventricular outflow tract Aortic valve Pulmonary valve | Pulmonary arteries (main and branches) Supra-aortic and supra-pulmonary anastomoses Coronary arteries Superior cavopulmonary anastomoses Inferior vena cava–Fontan conduit anastomosis Fontan conduit–pulmonary artery anastomosis |
Statistical Analysis
Quality scores assigned by both readers were averaged to determine rate of OI (percentage OI [%OI]) for individual cardiac structures as follows:
% OI = Number of optimal quality scores ( 1 ) Total number of scores ( 0 or 1 ) .
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
