Manual editing and artefacts marking of vital signs in an anesthesia information management system: A retrospective study on 193,414 anesthesia records
Alexander Avidan, MD, Charles Weissman, MD, Phillip D. Levin, MD
Department of Anesthesiology and Critical Care Medicine, Hadassah – Hebrew University Medical Center, Jerusalem Israel. General Intensive Care Unit, Shaarei Zedek Medical Center

Background
Vital signs recorded automatically into anesthesia information management systems (AIMS) are more accurate than those in the classic paper based anesthesia records. Alternately, artefacts are frequent and also recorded by the AIMS installed in our institution (Metavision, iMDSoft, Israel). Each vital sign value recorded automatically can be changed or marked manually as an error. Alternatively, a nonspecific "Artefact Monitor" flag can be entered. The goal of this study was to assess the rate of vital signs changed or marked as error and the use of the nonspecific "Artefact Monitor" flag.
Material and Methods
This study was approved by the Hadassah Institutional Review Board (IRB). Data (absolute numbers, number of changes, number marked as error) for vital signs (heart rate, SpO2, non-invasive and invasive blood pressure) and number of artefact flags were retrieved from the AIMS for the period February 2007 to December 2016.
Results
Data from 193,414 anesthesia records were analysed (Figure). In 20,689 (10.7%) cases anesthesiologists made changes to vital signs or marked them as error. This was mostly done for heart rates and SpO2 values. In 24,002 (12.4%) cases, 46,838 (mean 2.0±1.8) "Artefact Monitor" flags were used. In 151,901 (78.5%) cases no vital signs were changed or marked as error and no "Artefact Monitor" flags were used. Overall, very few vital signs were changed manually (0.15%).
Conclusions
Anesthesiologists rarely changed automatically collected vital signs data or marked them as errors. This may be related to the fact that changing the numeric value of a vital sign requires which requires several mouse clicks and may be too cumbersome. When changes are made, they were most frequent for heart rate and saturation. 
Table 1
Vital Sign	Automatically Recorded	Changed	Marked Errors	Total Manual Changes
		n	%	n	%	n	%
Heart Rate	18,998,098	19,026	0.1	24,473	0.1	43,499	0.23%
SpO2	18,407,827	14,800	0.1	20,571	0.1	35,371	0.19%
NIBP-Systolic	3,344,018	1,827	0.1	465	0.01	2,292	0.07%
NIBP-Mean	3,357,238	1,498	0.1	447	0.01	1,945	0.06%
NIBP-Diastolic	3,342,755	1,804	0.1	226	0.01	2,030	0.06%
IBP-Systolic	5,306,598	1,238	0.02	1,026	0.02	2,264	0.04%
IBP-Mean	5,425,001	1,193	0.02	1,914	0.04	3,107	0.06%
IBP-Diastolic	5,304,334	1,193	0.02	958	0.02	2,151	0.04%
Total	63,485,869	42,579	0.07%	50,080	0.08%	92,659	0.15%

Table 2
Total number of case:
193,414	Cases with recordings	Cases with Changes	Cases with Marked Errors	Cases with manual changes
	n	%	n	%	n	%	n	%
Heart Rate	191,288	98.9	5,112	2.7	8,337	4.4	13,449	7.0%
SpO2	190,840	98.7	4,715	2.5	8,949	4.7	13,664	7.2%
NIBP-Systolic	186,231	96.3	807	0.4	356	0.2	1,163	0.6%
NIBP-Mean	186,261	96.3	723	0.4	333	0.2	1,056	0.6%
NIBP-Diastolic	186,261	96.3	804	0.4	159	0.1	963	0.5%
IBP-Systolic	24,526	12.7	357	1.5	368	1.5	725	3.0%
IBP-Mean	23,869	12.3	335	1.4	538	2.3	873	3.7%
IBP-Diastolic	24,476	12.7	340	1.4	348	1.4	688	2.8%
Legend: NIBP: Non-invasive blood pressure, IBP: Invasive blood pressure

Manual editing and artefacts marking of vital signs in an anesthesia information management system: a retrospective study on 193,414 anesthesia records