SCCM Resource Library-58-Video-Based Capillary Refill Time Assessment in Children

Video Transcription

Hi, everyone. My name is Brandon Hunter, a third-year Pediatric Critical Care Fellow at the Children's Hospital  of Philadelphia, and I'll be presenting our research titled Video-Based Cafeteria Refill Time Assessment in Children on behalf of the authors displayed on this slide.  There are several disclosures to report from funding, co-author employment, and consulting relationships, which are listed here.  So capillary refill time is a commonly used bedside marker in shock assessment, and it's endorsed by multiple international guidelines.  Observational data and meta-analyses have repeatedly shown that abnormal capillary refill  time correlates with the development of critical illness and tracks with morbidity and mortality.  Recently, there has been further interest in studying capillary refill time as a clinical  indicator with the multi-center international randomized control trial in adults, the Andromeda shock trial.  The study compared an early eight-hour stepwise resuscitation strategy among patients with septic shock to either normalized lactate or normalized capillary refill time.  You can see from the Kaplan-Meier curves that although the study reported equivalence between the two strategies, there was, in fact, a trend towards improved mortality at 28 days  in the group that sought to normalize capillary refill time. With 43% mortality reported in the lactate group compared to 35% mortality in the capillary  refill time group, the p-value was 0.06. Of note, there was also a significant reduction in mean SOFA score at 72 hours in the cap refill time group compared to the lactate group.  Despite the apparent clinical utility of cap refill time, there's a tension in the literature  because cap refill time has limited and variable reporting of inter-rater and intra-rater reliability. Recorded kappa values range from less than 0.15 to greater than 0.7.  Of note, studies generally have different clinicians performing separate cap refill time collection events and compare the times that they report.  Furthermore, they often lack a strict protocol design, and importantly, those with more regimented protocols tend to have better clinician agreement.  Video-based cap refill time, which is the recording of a capillary refill event with controlled ambient lighting and collection protocol, allows multiple clinicians to assess  a single episode of capillary refill time and may be useful with remote care, though it's been poorly studied.  Overall, cap refill time seems to be an important clinical indicator, though it's unclear how reliable it is provider to provider or even with one provider assessing multiple times.  As such, our study has several aims. First, to evaluate the correlation between standardized bedside cap refill time and video-based cap refill time.  Second, to evaluate intra-rater and intra-rater reliability of video-based cap refill time evaluation.  And third, to apply a generalizability study to establish sources of variance in cap refill time detection.  Our hypothesis is that the correlation of bedside cap refill time and video-based cap refill time would be high, and that video-based cap refill time intra-rater and intra-rater  reliability would be moderately high across raters because of our strict collection protocol and the fact that multiple clinicians are reviewing a shared set of cap refill time  events recorded by video. Our study was a secondary analysis of an existing data set. Cap refill time videos were recorded by a Canon PowerShot video camera with optimal  hand position on a black background and controlled ambient lighting. Clinicians applied five seconds of firm compression followed by release and observation of a distal  second digit. The ventral or dorsal aspect was observed depending on clinician customary practice.  Bedside providers then verbalized when full cap refill had occurred, which was recorded by a stopwatch.  Following video collection, 30 clips were chosen based on time, 10 flash or less than one second, 10 normal or 1 to 1.99 seconds, and 10 prolonged or greater than or equal  to two seconds. The clips were played in random order twice for 10 critical care physicians.  Physicians clicked a button to indicate when they believe cap refill time began and ended. This was a fully cross-study design where each of the 30 video clips was reviewed by  each of the 10 critical care physicians two times. Intra-rater and inter-rater reliability were evaluated using intra-class correlation coefficient analysis.  We then ran a generalizability study, which in essence acknowledges that multiple factors may affect measurement variance.  Our generalizability study aimed to estimate variance components, also called facets, in their interaction underlying the measurement process.  The facets we examined were the video capillary refill time clinician raters, the specific video clips themselves, and the order of the video clip presentation.  This graph shows the correlation between bedside capillary refill time on the x-axis and video-based  capillary refill time, again, reflecting 10 clinicians reviewing 30 video clips on the y-axis. It revealed a moderate to good correlation between the two with an R-value of 0.65.  This shows box plots of the video-based capillary refill time assessment for the 30 video clips with diamonds representing the bedside clinician cap refill time estimate.  They're listed in order of bedside CRT, or cap refill time length. The measure was moderate inter-rater reliability with intra-class correlation coefficient ICC  of 0.61, good intra-rater reliability with an ICC of 0.71. On average, the bedside clinical capillary refill time was 0.17 seconds longer than the  video-based capillary refill time estimates. So looking at the results of the generalizability study with facets and facet interactions listed  in the left column, and the percentage of total variation in measurements attributed to each source in the right column, you can see that 57% of the variation in measurements  came from the specific video themselves, 11% from the interaction of an individual rater and the video, and less than 10% of variation came from the raters themselves.  In plain language, it shows what we can see from the prior box plot, that there were some videos which had very little variation in estimates, while others had quite wide variation.  The g-coefficients displayed at the right are consistent with the ICCs calculated previously,  showing moderate inter-rater reliability of 0.66 and good intra-rater reliability of 0.72.  So prior small studies have shown significant variance in video-based capillary refill time prediction among clinicians, but have not reported inter-rater reliability.  Our study shows moderate agreement between bedside and video-based capillary refill time review.  Inter-rater and intra-rater reliability of video-based capillary refill time estimation were moderate and good, respectively.  Generalizability analysis indicated that variance in capillary refill time estimation is primarily  caused by the CRT video, or capillary refill time video itself, rather than the reviewer.  Further analysis is being performed to ascertain which features intrinsic to these videos in capillary refill time events cause them to be the source of variation.  Things such as the length of bedside capillary refill time, patient clinical features, patient skin color, and ambient light are being considered. Our study has several limitations.  It was a single-center study at a tertiary care children's hospital. We utilized a convenient sample of pediatric patients in the operative and critical care settings.  And additionally, we had a strict capillary refill measurement protocol, but did permit bedside clinicians to assess either the ventral or dorsal aspect of the digit, depending on  their customary practice, to enhance the accuracy of their evaluation. Our results add to the literature regarding capillary refill time clinical reliability.  Cap refill time is a bedside marker with practical clinical utility and acceptable, but imperfect, reliability.  Further investigation into standardizing and automating collection and estimation of cap refill time is needed. And so that is the end of the presentation.  I just want to thank you all for listening and for your time, and hope you all take care.

Video Summary

The researcher presented a study on the correlation and reliability of capillary refill time (CRT) measurements. Cap refill time is an important clinical indicator used in assessing shock and critical illness. The study used video-based CRT assessments to evaluate reliability and correlation with bedside measurements. The results showed a moderate correlation between bedside and video-based CRT measurements and moderate to good reliability among clinicians. However, the study also found that the variation in measurements was primarily due to the specific videos rather than the clinicians. Further analysis is needed to determine the factors causing variation in video-based CRT measurements.

Asset Subtitle

Pediatrics, Procedures, 2021

Asset Caption

The Society of Critical Care Medicine's Critical Care Congress features internationally renowned faculty and content sessions highlighting the most up-to-date, evidence-based developments in critical care medicine. This is a presentation from the 2021 Critical Care Congress held virtually from January 31-February 12, 2021.

Ryan B. Hunter

Introduction/Hypothesis: Capillary refill time (CRT) is used to assess peripheral perfusion in children with suspected shock. Interclinician CRT assessment is reported as inconsistent. During pandemics remote video-based CRT (VBCRT) assessment may be warranted but is untested. We hypothesized the correlation of standardized bedside CRT and VBCRT would be high and VBCRT across raters would be consistent.

Methods: 99 children (1-12yr) had 10 consecutive bedside CRT measurements by an experienced critical care clinician. CRT measurement was standardized with 5s of firm compression and video recorded in a replicable fashion with optimal hand position in the camera window on a black background. 30 patient clips were prospectively selected: 10 flash (<1s), 10 normal (1-2s), and 10 prolonged (>2s). 10 pediatric critical care attending physicians received identical scripted instructions and reviewed 30 clips twice in randomized order on a 43in screen under identical lighting conditions. To ensure accurate recording of perceived CRT, raters were instructed to push a button when finger compression was released and again when finger color fully returned to baseline (CRT). The correlation and absolute difference between bedside and VBCRT were assessed by Pearson's correlation and paired t-test. Consistency across raters and repeated measures within each rater were analyzed using the intraclass correlation coefficient (ICC) and for absolute agreement using the Generalizability (G) study.

Results: 10 raters (practice years including fellowship: Mdn 11.5yr, IQR 9-14) reviewed 30 clips twice. Moderate agreement was observed between the bedside CRT and VBCRT (r=0.65, p<0.001). VBCRT was assessed as shorter by 0.17s (95% CI: 0.09-0.25, p<0.001) compared to bedside CRT. Among the video rating of each clip, the largest source of variance was patient (57%), followed by patient-rater interaction (11%). The consistency across raters over 30 clips x 2 measurements was ICC=0.58. Within each rater, the two measurements on the same video clip yielded good consistency: ICC=0.73. Overall G coefficient was 0.57.

Conclusions: Bedside and video assessment of CRT correlated with a statistically significant offset of 0.17s. Across raters, CRT measurement showed moderate consistency while repeated measurements within each rater were highly consistent.

Meta Tag

Content Type Presentation

Knowledge Area Pediatrics

Knowledge Area Procedures

Knowledge Level Intermediate

Knowledge Level Advanced

Membership Level Select

Tag Monitoring

Year 2021

Keywords

correlation

reliability

capillary refill time

video-based assessments

variation in measurements

	Society of Critical Care Medicine 500 Midway Drive Mount Prospect, IL 60056 USA Phone: +1 847 827-6888 Fax: +1 847 439-7226 Email: support@sccm.org	Contact Us About SCCM Newsroom Advertising & Sponsorship DONATE	MySCCM LearnICU Patients & Families Surviving Sepsis Campaign Critical Care Societies Collaborative	GET OUR NEWSLETTER
© Society of Critical Care Medicine. All rights reserved. \| Privacy Statement \| Terms & Conditions The Society of Critical Care Medicine, SCCM, and Critical Care Congress are registered trademarks of the Society of Critical Care Medicine.