Automated charting of physiological variables in anesthesia: A quantitative comparison of automated versus handwritten anesthesia records

Eight physiological variables—tidal volume, breathing rate, end-tidal carbon dioxide fraction, oxygen fraction in the anesthetic circuit, oxygen saturation by pulse oximetry, systolic and diastolic blood pressure, and heart rate—recorded on-line by a commercially available automated system were compared with the same variables recorded on handwritten anesthesia records. We quantified the differences between the automated and handwritten records generated from the same 30 patients (2,412 minutes of general anesthesia for elective eye surgical procedures). Considering the design of the study, we claim that the differences between both records were caused by the incompleteness or inaccuracy of the handwirtten records, except in two instances. The amounts of missing or erroneous data for these eight physiological variables were expressed as fraction (error fractions) of the time being recorded, designated EF_m and EF_e, respectively. For the first five variables the EF_m on the handwritten records ranged between 0.23 and 0.31, and the EF_c ranged between 0.01 and 0.06. For the last three variables the EF_m range was 0.08 to 0.13, and the EF_e range was 0.05 to 0.11. Most of these missing or erroneous data occurred during the period of induction (first 15 minutes) and at the end of the case (last 10 minutes). The EF_m and EF_e during induction had increased to 0.62 and 0.26, respectively, and to 0.76 and 0.06, respectively, at the end of the case. Erroneous data were observed on the automated records for the tidal volume during induction (EF_e=0.0044) and for the oxygen fraction during maintenance (EF_e=0.0024). The effect of averaging by the recordkeeper is discussed. The results of this study indicate the clinical relevance of automated record keeping.     

A reference system in clinical anesthesia

We have developed, as a pilot project, a hypertext-based reference system for anesthesia in a neurosurgical clinic. The intention is to give passive decision support prior to the operation. The information is presented according to the different needs and knowledge levels of the anesthesiologists. Therefore, with the aid of a simple and universal user interface (mouse and screen keys, no keyboard) the information is structured from basic to more detailed information. To complete the operation-related material the system also offers relevant data from the anesthesia-related background knowledge (pre-existing diseases, drugs, and medication, etc.). As an example, a short path through the system is shown to illustrate the presentation and linkage of information.This reference system was developed and implemented while Dr Zapf was an anesthesiologist at the Institute of Anesthesiology at the University of Erlangen-Nuremberg.     

Automating the recording and improving the presentation of the anesthesia record

Although anesthesia records have been kept for over a hundred years, there is still discussion of their value and content. Two uses of the record are widely accepted: (1) review after the anesthetic event (as in medicolegal disputes), and (2) support of patient care during the delivery of an anesthetic. Although the anesthetic record is mandatory in much of the world, there is not a single standard for its format. Automating the generation and presentation of the record will enhance its value and help develop a consensus as to content. Merely automating the steps used to produce the manually generated record does not realize the full benefit of automation. For maximum benefit, the primary goal of automation should be to support the uses of the record. Specific techniques that are discussed include increasing time resolution, optimizing the type and location of input and display equipment, and tailoring the human interface. Particular attention is paid to the issue of how much detail is acceptable in the record, how to use visual cues to present detail properly, how to exclude extraneous detail, and how to avoid misleading presentations (erroneous interpretation of the data). Specific elements discussed include line width, the use of color, presentation of gradients, statistical summaries, contexts for reporting data, graphical techniques for increasing data content, and pictorial presentations. Current records are more often confusing because presented information is inconsistently displayed or irrelevant than because too much information is offered, and automation can ameliorate this problem.Electrical EngineeringPhysiology     

The uses of the anesthesia record

The typical, handwritten anesthesia record of the 1980s does not satisfy its many users. The document is used for clinical care by the anesthetist, nurses, physicians, and technicians in postanesthesia, intensive, and postoperative surgical care units; for historical information by the billing officer, the statistician, and the anesthetist in preparation for a future anesthetic; and for the review of the quality of care by clinical peers and lawyers. For all of these users the typical record contains some to much unnecessary information and lacks some to much needed information. Electronic capture, storage, retrieval, and formatting of data can generate electronic displays or paper records tailored to answer the needs of specific users. The anesthetist in particular will benefit from a well-designed system that takes the place of the traditional handwritten anesthesia record.     

Tidal volume changes due to the interaction of anesthesia machine and anesthesia ventilator

Tidal volume (VT) delivered by mechanical ventilation during anesthesia may be influenced by factors related not only to the patient and the breathing circuit, but also to the interaction between the anesthesia machine and the anesthesia ventilator. To characterize this interaction, we studied in a test lung the effect of fresh-gas flow (FGF) (0.25, 2.5, 5, and 10 L/min), inspiratory-to-expiratory time ratio (I: E) (1 : 1, 1 : 2, and 1 : 3), and ventilatory frequency (8, 12, and 16 breaths/min) at fixed ventilator bellows excursions of 300, 600, and 900 ml. The influence of these variables was also estimated mathematically for a pediatric situation: a bellows excursion of 50 ml at 20 and 30 breaths/min. Each variable studied was associated with an increase, sometimes dramatic, in the delivered V_T compared with that which was set. The V_T augmentation was greatest at the highest FGF rate, largest I : E ratio, and slowest respiratory rate. Because the magnitude of the augmentation is independent of the V_T setting, the percent increase is much larger for pediatric settings. For example, with V_T set at 50 ml, delivered V_T ranged from 71 ml (FGF 2.5 L/min, I : E 1:3, and 30 breaths/min) to 300 ml (FGF 10 L/min, I : E 1:1, and 20 breaths/min). Thus it is possible in the pediatric situation to increase the delivered VT by sixfold without changing the ventilator bellows excursion. The magnitude of the changes was slightly larger for the VT settings for adult patients because of the slower respiratory rate. This VT augmentation can be predicted by the product of FGF (ml/s) and inspiratory time (seconds). Therefore it is possible to dramatically affect delivered V_T without changing the ventilator bellows excursion; this is an important consideration when FGF, I : E, or ventilator rate settings are selected or adjusted.     

Evaluation in animals of a system to estimate tracheal pressure from the endotracheal tube cuff

Objective. Flow through an endotracheal tube (ETT) causes a pressure loss across the tube. This loss results in a difference between pressure measured at the airway and pressure measured in the trachea. This difference can lead to errors when calculating pulmonary mechanics and when setting ventilators. We have tested a method of estimating tracheal pressure from the pressure in the ETT cuff.Methods. Pressure transducers were placed in the proximal ETT connector, in the trachea, and in the ETT cuff (through the inflation port). Instantaneous periods of zero flow, detected with a flow meter, were used to calculate the slope and offset of the line relating cuff pressure to tracheal pressure. The system was tested on the bench using a ventilator and lung simulator and in 2 dogs and 5 pigs. Tests were performed at various cuff pressures, trachea diameters, ETT sizes, respiratory rates, tidal volumes, and airway obstructions. Results. In bench tests, our estimate of tracheal pressure was within –4.0±2.6% of the actual tracheal pressure (mean = standard deviation [SD]). In animal tests, our estimation of tracheal pressure was within –0.6±5%. In all bench test measurements and in 40 of 42 animal measurements, the error was less than 1 cm H₂O.Conclusions. The cuff estimation technique gives real-time, continuous, noninvasive tracheal pressure measurements in intubated animals with cuffed ETTs.     

A program designed to facilitate the repetitive acquisition,display, and disk storage of sensory evoked responses

We present an algorithm that permits automated acquisition, display, and disk storage of single or dual channel sensory evoked potential waveforms using a Pathfinder II evoked response monitor. Once initiated, data acquisition and storage continue at user-specified intervals without additional user input. A cascade screen display provides trend monitoring. The file for data storage is established and searched using standard Pathfinder commands. The algorithm is written in MECOL, the Pathfinder specific programming language. This programming code imposed a number of limitations that had to be overcome to create a versatile and user friendly algorithm. During clinical use in our operating rooms, the program has been found effective and easy to use by both veteran and novice Pathfinder operators. The “handsoff” trend monitoring permitted by this algorithm has reduced the labor-intensive aspect of sensory evoked potential monitoring. The opinions expressed herein are those of the authors and are not to be construed as reflecting the views of the Navy Department, the Naval Service at large, or the Department of Defense.     

An automated, interactive analysis system for ultrasound sequences of the common carotid artery

Structural parameters of the common carotid artery (CCA) have shown to correlate with the risk of cardiovascular disease, but their precise measurement is challenging. We developed an automatic detection system with manual interaction capabilities that can reliably analyze B-mode ultrasound sequences of the CCA over several heart cycles. We evaluated 3824 frames from 40 sequences in two data qualities. Two readers measured the intima media thickness (IMT) and the lumen diameter at two evaluation times (T1/T2). A Bland-Altman analysis of the average IMT showed a bias ± SD of 0.002 ± 0.010 mm (T1), -0.004 ± 0.008 mm (T2) for completely automatic detections and -0.004 ± 0.010 mm (T1), -0.003 ± 0.010 mm (T2) for clips with manual corrections. The combination of automated analysis and manual intervention provides precise parameters as biomarkers for the atherosclerotic process and makes the system suitable for large scale epidemiological research, diagnostic and clinical practice.     

Viral load and inflammatory cytokine dynamics associated with the prognosis of severe fever with thrombocytopenia syndrome virus infection: An autopsy case

Severe fever with thrombocytopenia syndrome (SFTS) is an emerging tick-borne disease caused by a novel bunyavirus. The mechanism underlying disease progression remains unknown, and effective treatment strategy for SFTS is yet to be completely established, making its increasing incidence and subsequent mortality a great concern. Here, we present the autopsy case of a patient with rapidly progressed, fatal SFTS infection. Her viral titer and serum cytokines levels were measured daily and compared with the values of a survivor of the infection. Our findings elucidate the clinical features and pathophysiology of SFTS.