|
|
October 16-20, 2010 San Diego, CA Home Abstract Archive Search Abstracts 2010 Session Grid 2010 Meeting Website Policy Statements ASA Website Feedback 
Visit Anesthesiology.org |
| Previous Abstract | Next Abstract |
| Printable Version |
|
A1241
October 25, 2005 2:00:00 PM - 4:00:00 PM Hall C4 |
| A Continuous Video Buffering System for Recording Unscheduled Medical Procedures |
|
Richard P. Dutton, M.D., Peter F. Hu, M.S., C.N.E., Colin F. Mackenzie, M.D., Steven R. Seebode, Yan Xiao, Ph.D. Anesthesiology, University of Maryland, School of Medicine, Baltimore, Maryland, United States |
|
Introduction Emergency invasive procedures have high risk to patient safety but are difficult to consistently observe and document. As part of a human factors research effort examining emergency care, we designed an automated video buffering system that fulfilled Institutional Review Board (IRB) and medico- legal requirements for recording emergency medical procedures. We report the experiences of using the system in a study of central venous catheterization (CVC). Methods In a major trauma center, three cameras and one audio feed in each of nine-bays in a trauma resuscitation unit (TRU) were connected to a central video hub. A Digital Video Buffer (DVB) was designed with commercially available digital video recording devices. DVB could buffer up to 32 video and 12 audio inputs simultaneously 24x7. The buffered video quality is selectable from low (88x72) to highest (1408x1152) and from 1 to 30 frames/sec for each video input. The duration of buffer retention was controllable from hours to weeks with continuous looping and overwrites after retention time was exceeded. With password and hardware key protection, buffered video in the DVB system could be rapidly screened for occurrences of the procedures targeted in a research protocol on patient safety in CVC. Sixteen video images on one screen could be simultaneously screened and the buffered video could be searched by date, time and input source (e.g. TRU Bay #). The video buffer could be screened at the rate of 1:16 for all the video inputs simultaneously. Video segments with emergency medical procedures were extracted from the video buffer and recorded as AVI or MPEG digital video files. The IRB and hospital legal counsel approved the protocol to record all CVC insertions in the TRU. During the study period, all 27 video and 9 audio feeds were pre-scheduled and automatically buffered for 12 hrs Monday through Saturday. A technician extracted CVC procedures from the buffered video within the 12 hr buffering window and recorded them to AVI video files (one video file per camera input) for later expert analysis. Results The DVB system was used for 101 days, during which a total of 32,724 hours (101 Days x 12 Hours per day x 27 cameras) of video were buffered. A total of 101 CVC cases were recorded with 126 hours (0.39% of total buffered length) of video files (26.2GB). The video data screening for CVC insertions was rapid (ratio of 1:54, one hour screening fifty-four hours of buffered video). Average case length was 25 min, between shortest case 4 min (66MB) and longest case 61 min (638 MB). No medico-legal or privacy complaints were raised. Conclusion We successfully recorded all CVC insertions using the DVB system. An unmanned automatic video buffering system used for identifying and extracting CVC insertion reduced medico-legal concerns and minimized the staff time needed to record these procedures. This technique is generalizable to other medical, surgical and anesthesia procedures and processes (e.g. induction of anesthesia, transition of care, emergence, etc.) whose collection would otherwise be extremely difficult and time consuming. Previous efforts [1] have shown the benefits of video collection and analysis in identifying training needs and improving patient safety. Acknowledgement The work was supported by Agency for Healthcare Research and Quality (P20 HS11562) [1] Cogn Techn & Work 6:139-147, 2004 Anesthesiology 2005; 103: A1241 |