Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Apr 1.
Published in final edited form as: Paediatr Anaesth. 2023 Dec 6;34(4):318–323. doi: 10.1111/pan.14817

A Pragmatic Methodology to Extract Anesthetic and Physiological Data from the Electronic Health Record

Arshia Aalami Harandi 1,2,*, Katherine McPherson 1,2,*, Yungtai Lo 3, Rodrigo Gutiérrez 4,5, Jerry Y Chao 1,#
PMCID: PMC10922302  NIHMSID: NIHMS1952518  PMID: 38055618

Abstract/Summary

Background/Aims:

Traditional manual methods of extracting anesthetic and physiological data from the electronic health record (EHR) rely upon visual transcription by a human analyst that can be labor intensive and prone to error. Technical complexity, relative inexperience in computer coding and decreased access to data warehouses can deter investigators from obtaining valuable EHR data for research studies, especially in under-resourced settings. We therefore aimed to develop, pilot and demonstrate the effectiveness and utility of a pragmatic data extraction methodology.

Methods:

Expired sevoflurane concentration data from the EHR transcribed by eye was compared to an intermediate preprocessing method in which the entire anesthetic flowsheet narrative report was selected, copy-pasted and processed using only Microsoft Word and Excel software to generate a comma delimited (.csv) file. A step-by-step presentation of this method is presented. Concordance rates, Pearson correlation coefficients, and scatterplots with lines of best fit were used to compare the two methods of data extraction.

Results:

A total of 1,132 datapoints across 8 subjects were analyzed, accounting for 18.9 hours of anesthesia time. There was a high concordance rate of data extracted using the two methods (median concordance rate 100% range [96%, 100%]). The median time required to complete manual data extraction was significantly longer compared to the time required using the intermediate method (240 IQR [199, 482.5] seconds versus 92.5 IQR [69, 99] seconds, p=0.01) and was linearly associated with the number of datapoints (rmanual = 0.97, p<0.0001), whereas time required to complete data extraction using the intermediate approach was independent of the number of datapoints (rintermediate = −0.02, p=0.99).

Conclusions:

We describe a pragmatic data extraction methodology that does not require additional software or coding skills intended to enhance the ease, speed, and accuracy of data collection that could assist in clinician investigator-initiated research and quality/process improvement projects.

Keywords: Data science, Electronic health record (EHR), Anesthesia Information Systems (AIMS), Data extraction/scraping, Pediatric anesthesia, Concordance rate, Gold standard, Accuracy rate, Rigor and reproducibility in clinical research

Introduction

Clinical research studies involve gathering and analyzing different types of data. Two common examples of anesthetic data of interest include (1) pharmacological drug administration such as expired end-tidal sevoflurane concentration and (2) physiological hemodynamic data such as mean arterial blood pressure. In practice settings not utilizing paper anesthetic records, these data are now typically recorded continuously in the electronic health record (EHR) anesthesia information systems (AIMS)1-3 and can be accessed and gathered post hoc by using a variety of methods that range from a manual data extraction approach from patients’ EHR flowsheets – a process traditionally involving a human data analyst(s) visually reading and transcribing minute-by-minute anesthetic-physiological data by hand onto a separate spreadsheet software program, a time-consuming and error-prone process4-9 – to more technically complicated data extraction approaches involving computer coding in languages such as Structured Query Language (SQL) that require additional expertise and resources10-13. Data can be accessed by relying on the support of (1) helpful intradepartmental tech savvy clinicians, (2) EHR company staff with computer coding expertise to increase access in a more scalable fashion, (3) data scientists in larger academic practices and (4) through independent direct access to a variety of monitored variables in a perioperative data warehouse. At all times, potential limitations in data granularity and artifacts in the perioperative EHR and AIMS require careful examination, exploration and analytics to avoid the hazards of “garbage in, garbage out” (GIGO)3,14 in order to enhance data quality and mitigate inaccurate analyses and study interpretation.

Inexperience with coding or practice settings with less well-resourced research infrastructure are two examples of important barriers that can potentially deter investigators from obtaining the valuable clinical data necessary to perform analyses in research and quality/performance improvement projects. An intermediate approach beyond iterative, manual transcription by eye (but not quite approaching the level of computer coding) could be of potential benefit and interest to the community of “data interested” pediatric anesthesiologists. Such an approach might empower clinicians with a more efficient method to perform their own self-initiated projects who may otherwise lack the knowledge, time, access or resources to engage more sophisticated approaches of obtaining anesthetic and physiological data from the EHR.

In this study, we aimed to describe such an intermediate preprocessing approach intended to be pragmatic, useful and potentially easy to learn. This method involves (1) a simple cut-and-paste procedure of an entire EHR anesthetic flowsheet report using conventional word processing software which could reduce manual transcription errors8, (2) followed by a filtering and selection for specific data of interest using conventional spreadsheet software functions. While not fully automated or error-proof, this pragmatic method is nonetheless intended to help enhance the ease, speed, and potential accuracy of data collection without requiring any coding skills or additional software other than Microsoft Excel and Microsoft Word. We hypothesized that data collected via an intermediate preprocessing procedure would have a high concordance rate with data extracted using a traditional manual method. We further hypothesized that the time required to perform manual data extraction would increase proportionally with the number of observations in the dataset whereas for the intermediate method, the time required to perform data extraction would be independent of the number of datapoints.

Methods

A subset of de-identified subject data from an ongoing observational study approved by the Institutional Review Board (IRB) of Albert Einstein College of Medicine was examined in this study15,16. De-identified anesthetic-physiological data from the electronic health record (EHR), (EPIC Systems Corp., Verona, WI), were accessed and extracted using a manual data extraction procedure that was compared to an intermediate preprocessing data extraction procedure (Figure 1). Data extraction was performed primarily by two medical student members of the research team with no prior experience with any data extraction approaches in the EHR.

Figure 1.

Figure 1.

Open in a new tab

Flowchart of Study Procedures.

For the manual data extraction method procedure, anesthetic data were manually extracted by a human analyst from the anesthetic record via visual identification using “Control-F” to find instances of the desired variable phrase (i.e., “Sevoflurane Expired”). Each datapoint’s time and value was then manually inputted into a spreadsheet software program (Microsoft® Excel for Mac Version 16.58, Microsoft Corp., Redmond, WA).

An intermediate data extraction procedure was developed in which for a given subject, the entirety of a subject’s minute-by-minute anesthetic-physiological data in the EHR flowsheet report was accessed, manually selected and copied using “Control-C” and pasted into a word processing program (Microsoft® Word for Mac Version 16.58) using “Control-V.” This document was then saved as a “plain text file” and imported into Microsoft Excel using the legacy function, “import from text,” with no delineator screening. Once imported into Microsoft Excel, these data were filtered using the “Begins With” function to select for data points beginning with “Sevoflurane Expired.” Finally, the filtered data were separated using “Space” and “Other: )” as delimiters to isolate the time of each datum and corresponding expired sevoflurane datum of interest. The columns generated for time and corresponding anesthetic data of interest, now isolated, were copied to a separate spreadsheet and saved as a comma delimited (.csv) file for import to statistical software for further analysis (Stata 17.0 BE, StataCorp, College Station, TX). A complete, step-by-step presentation and walkthrough of the data extraction procedure with screenshots can be found in the Supporting Information. Recognizing that different EHR systems produce anesthesia flowsheet reports that may differ in format, syntax or the order in which data is outputted, we have added specific guidance to potentially handle these differences. Provided that the flowsheet contains repeated instances of both the desired physiological datum with corresponding time (e.g., “Expired Sevoflurane: 0.25 % (Device Time: 17:01:12)”), the process outlined should allow the analyst to successfully filter, isolate and obtain the desired data in a customizable manner, even for flowsheets produced from different EHRs.

In response to the editor and reviewers’ feedback, a post hoc analysis was performed in which anesthesia data obtained from EHR flowsheets using the intermediate data extraction procedure were compared to corresponding anesthesia data pulled from a Clarity database query to determine if they matched.

Statistical Analysis.

From a sample of subject charts identified and selected based on a priori hypothesis for an unrelated study that required analysis of anesthetic-physiological data, a subset of 8 subject charts had minute-by-minute expired sevoflurane concentration data extracted using both a manual procedure and an intermediate preprocessing procedure (Figure 1). Datapoints acquired using both methods were assessed for concordance, defined as both data extraction methods recording the same time-specific expired sevoflurane concentration in a patient chart. The concordance rate was calculated as the number of concordant (or identical) expired sevoflurane values divided by the total number of expired sevoflurane concentration values extracted in a patient chart.

Pearson correlation coefficients were computed to assess the association between the time required to complete data extraction and the number of data points for extraction from a single data analyst (K.M.) using each method. These were further shown graphically using scatterplots of the data with a line of best fit. For the manual procedure, time was measured beginning from first datum entry into the spreadsheet until the last datum entry. The duration of time required for the intermediate approach to data extraction was measured beginning with "control-C" key entry to copy all data from the anesthetic flowsheet for pasting into a text file until completion of data entry into the new spreadsheet.

Differences in median time required for data extraction between these two methods were compared using the Wilcoxon signed-rank test for paired data. Values obtained using the intermediate data extraction procedure were compared to values obtained from a Clarity database query by performing an arithmetic subtraction of the data at identical time stamps. An arithmetic difference of “zero” indicated matching values. All statistical analyses were two-tailed with a level of statistical significance of p<0.05 and performed using Stata 17.0 BE (StataCorp, College Station, TX).

Results

A total of 1,132 paired datapoints were assessed accounting for 18.9 hours of anesthesia monitoring time obtained from 8 infants (median age 8.18 IQR [7.15, 10.68] months) consisting of 6 males (75%) all undergoing general anesthesia. The values of expired sevoflurane concentration ranged from 0% to 7.85% and the anesthetic monitoring time ranged from 23 to 339 minutes. The median concordance rate was 100% range (96%, 100%). One chart had a concordance rate that was not 100% and consisted of a discordant datum value in the hundredths decimal place.

The time required for data extraction using the manual vs. intermediate method was compared, with times ranging from 66 seconds to 977 seconds (1.1 to 16.3 minutes). The median time required for manual data extraction was significantly higher than the median time required using the intermediate data extraction procedure (240 IQR [199, 482.5] seconds versus 92.5 IQR [69, 99] seconds, p=0.01). On scatterplots, the manual data extraction method demonstrated a positive linear association between the time required for completing data extraction and the number of datapoints (rmanual = 0.97, p<0.001). This was not observed for the intermediate methodology (rintermediate = −0.02, p=0.99, Figure 2).

Figure 2.

Figure 2.

Open in a new tab

Scatterplot of the time required for completion of data extraction and number of datapoints with lines of best fit by manual compared to intermediate preprocessing method procedures. Pearson correlation coefficient: rmanual = 0.97, p<0.0001 and rintermediate = −0.02, p<0.99.

Post hoc analyses demonstrated that anesthetic data obtained from the Clarity database matched values obtained using the intermediate data extraction method.

Discussion

In this study, we demonstrate a high concordance rate between data obtained from the electronic health record (EHR) using a traditional manual approach compared to an approach employing an intermediate preprocessing procedure. As there was no a priori “gold standard” method of data extraction assessed in this study, we avoid using the term “accuracy.” Assuming that data obtained using the intermediate method is a proxy for gold standard values of expired sevoflurane concentrations in the EHR, the high concordance rate reflects a high accuracy rate. Post hoc comparisons with values obtained from a direct EHR Clarity database query support this assumption, suggesting that the intermediate method has a high accuracy rate. There was one discordant end-tidal sevoflurane concentration value in our sample of data extraction that was later confirmed to be an error in the manual method. The discordance was observed in the hundredths decimal place and potentially attributable to a keystroke error or analyst fatigue1,4-9.

Errors in data transcription employing either manual8 or intermediate preprocessing methods such as the one described in this manuscript can arise for a number of different reasons. Manual errors can be classified as (1) conceptual errors such as those that arise from reading the wrong data (e.g., wrong line of data, wrong variable, or wrong variable unit) leading to systemic errors in the overall data or (2) simple, one-off transcription errors8 that may occur unexpectedly and at random. Although automated or “copy and paste” methods do not have transcription errors per se, they can be sensitive to errors in programming, human error or differences in the formatting of the flowsheet report, with slight variations arising in output files leading to potentially significant errors in the dataset analyses. As an example of this, time periods with no captured patient expired sevoflurane data led to a gap in the time stamps in the flowsheet report and, ultimately, in the .csv file generated using the intermediate method. These data were not “missing” per se, but potentially led to irregular time intervals that were not present in the EPIC Clarity database .csv file. These inconsistent time intervals could affect downstream analyses.

In our analysis of concordance rates, we extracted the same number of datapoint observations per subject chart in order to standardize across all subjects. The specific number of 23 was due to the shortest anesthetic duration lasting 23 minutes in our sample. We further avoided combining different subject data together into one large dataset that could lead to different propensities for human data analyst fatigue and human error, particularly with longer subject records, which could potentially affect the results and inferences from the analysis.

In analyses of time required to perform data extraction, the intermediate method was not dependent on the number of data points. A longer duration of anesthesia/surgery generating a greater number of datapoints did not require a longer time for data extraction. In contrast, for the manual data extraction method, there was a direct linear relationship observed between the time required to complete manual data extraction and the number of datapoints (rmanual=0.98, p<0.0001). As an example, the shortest anesthetic record in our sample was 23 minutes and required 88 seconds to complete manual data extraction, which was 22 seconds slower than the intermediate method, which required 66 seconds to complete. For the longest anesthetic record lasting 339 minutes (5.7 hours), the manual extraction procedure required a substantially longer time to complete compared to the intermediate preprocessing procedure (977 seconds [16.3 minutes] vs. 67 seconds) and was likely more prone to human error. The median time to complete the intermediate procedure was 92.5 IQR (69, 99) seconds. It is possible that the manual extraction procedure is quicker for patients with anesthetic duration less than 23 minutes, but such short procedures are uncommon in pediatric anesthesia17 and overall, the intermediate procedure may offer important advantages for both accuracy and time required for data extraction, especially for a larger number of datapoints.

This study demonstrates that an intermediate preprocessing data extraction procedure is feasible using only word processing and spreadsheet software programs widely available and familiar to most clinicians without requiring additional resources or specialized training in coding methods that may be important barriers to implementation, particularly for early-stage research studies. Although relatively rudimentary, this approach could be valuable to the community of “data-interested” pediatric anesthesiologists who can effectively and easily implement the process outlined in this report. More broadly, increased knowledge of a variety of approaches for EHR data extraction can enhance research in underrepresented settings such as non-academic anesthesia departments that may be less well-resourced. The intermediate method described in this manuscript has the advantage of being easily applied to any flowsheet in the EHR that is familiar and readily accessible in the course of clinical workflow, which may be more intuitive to some investigators. As research projects grow in complexity and sampling of datapoints increases, issues of scale arise that would decrease the effectiveness of this approach, requiring additional data science expertise10.

Another important limitation of the study is that the approach outlined was developed from flowsheets generated in only one EHR system and not tested in other systems; there may be important differences that could prevent the successful implementation of our method. Other EHR systems such as Cerner may also offer built-in functionality to provide anesthesia physiological values as a table that can be readily copied and pasted directly into a spreadsheet program, thereby bypassing the need for intermediate preprocessing steps like the ones described in this manuscript. Notwithstanding these limitations, provided that the flowsheet output contains repeated instances of both the desired physiological datum with corresponding time, the step-by-step procedure outlined in the Supporting Information should be adaptable and customizable to other EHR flowsheets. Finally, the calculations of the time required to perform the two methods were obtained from a single data analyst who had already begun to develop a familiarity with both manual and intermediate preprocessing data extraction methods (and not from naïve users). This likely introduced bias by shortening the time required to perform both methods. Future work could incorporate comparisons of the data obtained from naïve vs. experienced users.

The intention of this manuscript is to assist clinical investigators by providing an example of one data scraping approach that could set the stage for further career development training activities to potentially learn increasingly more complex and sophisticated approaches in the continuum of data science methods, or at a minimum a facilitation of conversations with members of the data team. Future studies could investigate other pragmatic data scraping approaches in EPIC as well as other EHR systems. At all times, investigators should heed the important adage “garbage in, garbage out” (GIGO) in any data extraction and analysis approach14,18, as limitations in data granularity arising from “noise” or artifacts in the EHR AIMS have been well reported19,20. One study found that up to over 30% of invasive blood pressure data could contain artifacts19. Therefore, data should be reviewed using visual analytics3 as well as other approaches to assist artifact processing21,22 in an effort to enhance data quality and mitigate the potential for inaccurate analyses as well as study interpretations. Manual data collection might also contain artifacts and the advantage of the proposed method (or any data extraction from the EHR) might be to facilitate increased sampling of more data to facilitate further examination of the data characteristics such as its distribution, number of gaussians or outliers. Overall, the intermediate preprocessing approach demonstrated in this report allows investigators to obtain data from the anesthesia EHR flowsheet with a high concordance rate with data obtained using traditional manual extraction procedures, suggesting it may have a high accuracy rate, be less prone to human error, and serve as a pragmatic approach that clinicians with less knowledge of data science methods can add to their skillset to assist in self-initiated research and quality/process improvement projects.

Supplementary Material

Supinfo2

Supporting Information Table 1. Demographics and characteristics of infant subjects in whom expired sevoflurane data were obtained.

Supinfo1

Clinical Implications

What is already known about the topic?

Traditional manual methods of extracting anesthetic and physiological data from the electronic health record (EHR) rely upon visual transcription by a human analyst that can be labor intensive and prone to error, with technical complexity and inexperience in data science methods potentially deterring investigators from obtaining valuable clinical data, especially in under-resourced settings. Does there exist a practical and relatively technically uncomplicated methodology to extract data from the EHR not requiring computer coding skills and using only software that is already familiar to most clinicians (i.e., Microsoft Word and Excel)?

What new information this study adds:

In an analysis of 1,132 minute-by-minute anesthetic-physiological datapoints in the EHR of 8 subjects accounting for 18.9 hours of anesthesia monitoring time, a high concordance rate was observed between data extracted using an intermediate preprocessing approach compared to traditional manual data extraction; the preprocessing approach also required less time to complete, especially for a large number of datapoints. The approach presented in this report (outlined step-by-step in the Supporting Information) could be useful to help enhance the ease, speed, and accuracy of data collection to assist clinicians in their research and quality/process improvement studies.

Acknowledgements:

The authors would like to acknowledge (1) Drs. Maíra Rudolph and Ling Zhang for their help with an EPIC Clarity database pull for post hoc assessments confirming that EHR flowsheet values are a proxy for values in the database, (2) Dr. Matthias Eikermann and (3) the Digital Health Research Team in the Department of Anesthesiology at Montefiore Einstein.

Funding:

This study was supported by an Albert Einstein College of Medicine Summer Research Fellowship (to K.M.), National Institutes of Health (NIH) National Center for Advancing Translational Science (NACTS) Einstein-Montefiore CTSA Grant Nos. UL1TR002556, KL2TR002558 (to J.Y.C.), and K23DA057499 (to J.Y.C.).

Footnotes

Conflicts of Interest: None

Supporting Information. Step-by-step walkthrough of the intermediate preprocessing methodology.

Data statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

  • 1.Callahan A, Shah NH, Chen JH. Research and Reporting Considerations for Observational Studies Using Electronic Health Record Data. Ann Intern Med. 2020;172(11 Suppl):S79–S84. doi: 10.7326/M19-0873 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Ehrenstein V, Kharrazi H, Lehmann H, Taylor CO. Obtaining Data From Electronic Health Records. In: Tools and Technologies for Registry Interoperability, Registries for Evaluating Patient Outcomes: A User’s Guide, 3rd Edition, Addendum 2 [Internet]. Agency for Healthcare Research and Quality (US); 2019. Accessed May 22, 2023. https://www.ncbi.nlm.nih.gov/books/NBK551878/ [Google Scholar]
  • 3.Simpao AF, Ahumada LM, Rehman MA. Big data and visual analytics in anaesthesia and health care. Br J Anaesth. 2015;115(3):350–356. doi: 10.1093/bja/aeu552 [DOI] [PubMed] [Google Scholar]
  • 4.Buczek L, Azar F, Bauzon J, Batra K, Murphy C, Wahi-Gururaj S. The Data Error Criteria (DEC) for retrospective studies: development and preliminary application. J Investig Med Off Publ Am Fed Clin Res. 2023;71(4):448–454. doi: 10.1177/10815589231151437 [DOI] [PubMed] [Google Scholar]
  • 5.Cheng AC, Banasiewicz MK, Johnson JD, et al. Evaluating automated electronic case report form data entry from electronic health records. J Clin Transl Sci. 2023;7(1):e29. doi: 10.1017/cts.2022.514 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Hong MKH, Yao HHI, Pedersen JS, et al. Error rates in a clinical data repository: lessons from the transition to electronic data transfer--a descriptive study. BMJ Open. 2013;3(5):e002406. doi: 10.1136/bmjopen-2012-002406 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Paulsen A, Harboe K, Dalen I. Data entry quality of double data entry vs automated form processing technologies: A cohort study validation of optical mark recognition and intelligent character recognition in a clinical setting. Health Sci Rep. 2020;3(4):e210. doi: 10.1002/hsr2.210 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Mays JA, Mathias PC. Measuring the rate of manual transcription error in outpatient point-of-care testing. J Am Med Inform Assoc JAMIA. 2019;26(3):269–272. doi: 10.1093/jamia/ocy170 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Paulsen A, Overgaard S, Lauritsen JM. Quality of data entry using single entry, double entry and automated forms processing--an example based on a study of patient-reported outcomes. PloS One. 2012;7(4):e35087. doi: 10.1371/journal.pone.0035087 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Gamal A, Barakat S, Rezk A. Standardized electronic health record data modeling and persistence: A comparative review. J Biomed Inform. 2021;114:103670. doi: 10.1016/j.jbi.2020.103670 [DOI] [PubMed] [Google Scholar]
  • 11.Cai T, Zhang L, Yang N, et al. EXTraction of EMR numerical data: an efficient and generalizable tool to EXTEND clinical research. BMC Med Inform Decis Mak. 2019;19(1):226. doi: 10.1186/s12911-019-0970-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Joseph N, Lindblad I, Zaker S, et al. Automated data extraction of electronic medical records: Validity of data mining to construct research databases for eligibility in gastroenterological clinical trials. Ups J Med Sci. 2022;127: 10.48101/ujms.v127.8260. doi: 10.48101/ujms.v127.8260 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Milinovich A, Kattan MW. Extracting and utilizing electronic health data from Epic for research. Ann Transl Med. 2018;6(3):42–42. doi: 10.21037/atm.2018.01.13 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Simpao AF, Gálvez JA. Reporting of Observational Research in ANESTHESIOLOGY: The Importance of Data Quality: Trust but Verify. Anesthesiology. 2018;128(2):250–251. doi: 10.1097/ALN.0000000000002007 [DOI] [PubMed] [Google Scholar]
  • 15.Chao JY, Legatt AD, Yozawitz EG, Adams DC, Delphin ES, Shinnar S. Electroencephalographic Findings and Clinical Behavior During Induction of Anesthesia With Sevoflurane in Human Infants: A Prospective Observational Study. Anesth Analg. 2020;130(6):e161–e164. doi: 10.1213/ANE.0000000000004380 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Chao JY, Gutiérrez R, Legatt AD, et al. Decreased Electroencephalographic Alpha Power During Anesthesia Induction Is Associated With EEG Discontinuity in Human Infants. Anesth Analg. 2022;135(6):1207–1216. doi: 10.1213/ANE.0000000000005864 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Bartels DD, McCann ME, Davidson AJ, Polaner DM, Whitlock EL, Bateman BT. Estimating Pediatric General Anesthesia Exposure: Quantifying Duration and Risk. Paediatr Anaesth. 2018;28(6):520–527. doi: 10.1111/pan.13391 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Simpao AF, Nelson O, Ahumada LM. Automated anesthesia artifact analysis: can machines be trained to take out the garbage? J Clin Monit Comput. 2021;35(2):225–227. doi: 10.1007/s10877-020-00589-6 [DOI] [PubMed] [Google Scholar]
  • 19.Kool NP, van Waes JAR, Bijker JB, et al. Artifacts in research data obtained from an anesthesia information and management system. Can J Anaesth J Can Anesth. 2012;59(9):833–841. doi: 10.1007/s12630-012-9754-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Hoorweg ALJ, Pasma W, van Wolfswinkel L, de Graaff JC. Incidence of Artifacts and Deviating Values in Research Data Obtained from an Anesthesia Information Management System in Children. Anesthesiology. 2018;128(2):293–304. doi: 10.1097/ALN.0000000000001895 [DOI] [PubMed] [Google Scholar]
  • 21.Pasma W, Peelen LM, van Buuren S, van Klei WA, de Graaff JC. Artifact Processing Methods Influence on Intraoperative Hypotension Quantification and Outcome Effect Estimates. Anesthesiology. 2020;132(4):723–737. doi: 10.1097/ALN.0000000000003131 [DOI] [PubMed] [Google Scholar]
  • 22.Pasma W, Wesselink EM, van Buuren S, de Graaff JC, van Klei WA. Artifacts annotations in anesthesia blood pressure data by man and machine. J Clin Monit Comput. 2021;35(2):259–267. doi: 10.1007/s10877-020-00574-z [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supinfo2

Supporting Information Table 1. Demographics and characteristics of infant subjects in whom expired sevoflurane data were obtained.

NIHMS1952518-supplement-Supinfo2.docx (13.9KB, docx)
Supinfo1
NIHMS1952518-supplement-Supinfo1.docx (2.5MB, docx)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

RESOURCES