Indian Journal of Anaesthesia  
About us | Editorial board | Search | Ahead of print | Current Issue | Past Issues | Instructions
Home | Login  | Users Online: 242  Print this pageEmail this pageSmall font sizeDefault font sizeIncrease font size    




 
 Table of Contents    
ORIGINAL ARTICLE
Year : 2019  |  Volume : 63  |  Issue : 10  |  Page : 797-804  

Completeness of manual data recording in the anaesthesia information management system: A retrospective audit of 1000 neurosurgical cases


1 Department of Neuroanaesthesia and Neurocritical Care, National Institute of Mental Health and Neurosciences, Bengaluru, Karnataka, India
2 Brains Neuro Spine Centre, Bengaluru, Karnataka, India

Date of Submission04-Jun-2019
Date of Decision29-Jul-2019
Date of Acceptance09-Aug-2019
Date of Web Publication10-Oct-2019

Correspondence Address:
Dr. Kamath Sriganesh
Department of Neuroanaesthesia and Neurocritical Care, Neurosciences Faculty Block, 3rd Floor, National Institute of Mental Health and Neurosciences, Bengaluru - 560 029, Karnataka
India
Login to access the Email id

Source of Support: None, Conflict of Interest: None


DOI: 10.4103/ija.IJA_450_19

Rights and Permissions
 


Background and Aims: Anaesthesia information management system (AIMS) is increasingly implemented in many hospitals. Considering the capital cost involved in its installation and maintenance, it is important to evaluate its performance and adoptability by end users. This study assessed the completeness of manual data recording in the AIMS one year after its implementation and also evaluated potential predictors for completeness. Methods: In this retrospective audit of AIMS, 1000 electronic anaesthesia records of patients undergoing neurosurgical procedures over one year were assessed for completeness of 41 preidentified items, one year after its implementation. Parameters evaluated were patient identifiers, personnel identifiers, demographics, airway management parameters, anaesthesia management items and end-of-anaesthesia parameters. We hypothesised that completeness of anaesthesia record can be predicted by nature of surgeries, case sequence, seniority of anaesthesiologist and phase ( first or second) of the study period. Results: We observed higher completeness of manual data recording during phase 2 of AIMS use compared to phase 1. Higher grade of anaesthesiologist, second case of the day and emergency surgery led to reduction in completeness of data entry. Anaesthesiologist grade significantly predicted complete entry of 18 (44%) variables, case number predicted 8 (20%) variables and phase- and procedure-type predicted 6 (15%) and 5 (12%) variables, respectively. Conclusion: Completeness of manual data recording in the electronic AIMS is poor after one year of implementation. First case of the day, second phase of study period, elective cases and trainee anaesthesiologist are associated with better completeness of manual data recording in the AIMS.

Keywords: Data analysis, electronic medical record, integrated information management system, medical audit, quality improvement


How to cite this article:
Palaniswamy SR, Jain V, Chakrabarti D, Bharadwaj S, Sriganesh K. Completeness of manual data recording in the anaesthesia information management system: A retrospective audit of 1000 neurosurgical cases. Indian J Anaesth 2019;63:797-804

How to cite this URL:
Palaniswamy SR, Jain V, Chakrabarti D, Bharadwaj S, Sriganesh K. Completeness of manual data recording in the anaesthesia information management system: A retrospective audit of 1000 neurosurgical cases. Indian J Anaesth [serial online] 2019 [cited 2019 Oct 18];63:797-804. Available from: http://www.ijaweb.org/text.asp?2019/63/10/797/268719




   Introduction Top


The electronic anaesthesia information management system (AIMS) provides a permanent paperless method for capturing and storing anaesthesia-related information during the perioperative period. The traditional method of manual record-keeping has various shortcomings.[1],[2],[3] To overcome this, the AIMS is being implemented in many hospitals including those in the developing countries. Considering the capital cost involved in its installation and maintenance in resource constraint settings, it is important to evaluate its performance and adoptability by the end users. Our hospital recently installed AIMS (Centricity™ Perioperative Anaesthesia, GE Healthcare) replacing the paper record to facilitate seamless and complete data acquisition and maintenance. This change of recording from manual to digital version was aimed to achieve several objectives such as legibility, complete and accurate capture of data, longer preservation of record, easy access to database for clinical, academic and research purpose and to establish a system compatible for any future medico-legal and quality assurance framework.

Our AIMS collects physiological data automatically from multiple sources (multiparameter monitor, anaesthesia workstation, target-controlled infusion pump and cardiac output monitor) and combines them into a consistent record of the perioperative period to support informed decisions and to improve quality of patient care. However, many other items such as demographics, drugs and infusions require manual entry. Our AIMS is independent from our picture archiving system and hospital information system. With completion of one year since the installation of this system, we planned to assess our AIMS for its efficiency, inadequacies and anaesthesiologists' adaptation to the change from manual to electronic record. Previous reports have observed improvements in the completeness of electronic anaesthetic data in comparison to manual records[4] but also noted inaccuracies in some of the items documented in the electronic AIMS.[5]

The primary objective of this audit was to assess completeness of the manual data recorded in the AIMS during neurosurgical procedures. Our secondary objective was to identify potential predictors of completeness of data recording.


   Methods Top


This retrospective audit of our AIMS was conducted at a tertiary care neurosciences academic centre, after approval from the institute's ethics committee (NIMHANS Ethics Committee, Approval no.—NIMHANS/IEC (BS & NS DIV/11th meeting 2018, Date—15/03/2018). Requirement for written informed consent was waived by our ethics committee. Electronic anaesthesia records from our AIMS beginning from 1st January 2018 to 31st December 2018 were extracted to assess completeness of preidentified parameters listed in [Table 1]. This period coincided with completion of one year of installation of AIMS at our institution. We planned to analyse 1000 anaesthesia records in this audit. We a-priori identified items that were part of our earlier manual anaesthesia record and also additional items from the AIMS that were deemed important for a good anaesthesia record. These 41 items were broadly classified as patient identifiers (4 variables), personnel identifiers (3 variables), demographics (8 variables), airway management parameters (5 variables), anaesthesia management items (13 variables) and end-of-anaesthesia parameters (8 variables). Each item was given a score of 0 for missing data entry, 1 for partial data entry and 2 for complete data entry or 0 = No and 1 = Yes, as applicable. Data regarding these preidentified parameters and predictors of their completeness obtained from 1000 electronic anaesthesia records were collected on a Microsoft Excel worksheet by two researchers for analyses.
Table 1: Percentages of completeness of variables in the AIMS

Click here to view


We a-priori identified certain factors that could predict completeness of manual data entry into the AIMS. These factors were elective or emergency nature of surgeries, first or subsequent case of the day, seniority of the anaesthesiologist (resident or faculty) and phase 1 (January to June) or phase 2 (July to December) of the study period. We hypothesised that completeness of electronic anaesthesia chart would be better for 1) elective surgeries than emergency surgeries, 2) first case than subsequent cases of the day, 3) junior anaesthesiologists than senior anaesthesiologists and 4) phase 2 of the study period than phase 1. We also planned to solicit feedback from the users to suggest improvements in enhancing the completeness of manual recordings and efficiency of the AIMS.

Data were collated offline into a Microsoft Excel spreadsheet (version 2007). Data analysis was conducted using R software (ver. 3.5.2).[6] Data were complete for all of the samples (n = 1000), and all were included in the analyses. Predictor variables were predefined as phase of study period (1/2), procedure type (emergency/elective), case number of the day (1/2) and anaesthesiologist grade (residents- first year, second year, third year and faculty). All residents were supervised by a faculty but all faculties functioned independently. For each of the predefined predictor variables, Chi-square test was used to find association with all outcome variables. The variables found significantly associated at P < 0.05 were selected for inclusion in the final models. The outcome variables were grouped into six groups—patient identifiers, personnel identifiers, demographics, airway management details, anaesthesia management details and end-of-anaesthesia details.

Due to multiple possible correlations within the predictors and the outcomes, and need for modelling multiple predictors for multiple outcomes, structural equation models (SEM) were built for each separate group of outcomes based on the variables found significant at univariate level. The predictors were entered into the model as ordinal variables such that phase 2 > 1, case number 2 > 1, procedure type emergency > elective and anaesthesiologist grade scored by seniority. The model was estimated using diagonally weighted least squares (DWLS) estimator and robust “sandwich type” standard error calculation, with no constraints placed on any of the model parameters. The SEM was conducted using “lavaan” package of R.[7]

For the uninitiated to our statistical analyses, the interpretation of our statistical analyses is as follows. The estimates produced from the SEM models using DWLS estimator are akin to those from a proportional odds model. The estimates are additive in the native format and exponent of the estimate is the odds ratio for prediction of the outcome variable. The benefit of the models is that the estimates of the predictors for any given outcome within that model can be added up and exponent of the sum of estimates denotes the odds of the presence of the outcome.


   Results Top


A total of 2155 electronic anaesthetic records for neurosurgical procedures from seven operating rooms were retrieved from the AIMS for the study period. We excluded records of patients from three emergency operating rooms in a different building where AIMS was not installed, and records from elective operating rooms where technical reasons precluded use of AIMS. Emergency surgeries performed in the elective operating rooms were included. We then randomly selected 1000 electronic anaesthesia records for analyses of our objectives considering logistical reasons. The first author (SRP) blinded to the content of the records, randomly picked 500 electronic anaesthesia records from our AIMS server for each of the two study phases. This method (about 83 records for each month for both the phases) was used to avoid any selection bias.

The degree of completeness of all variables is represented in [Table 1] as proportions. The detailed tables of univariate tests of association are presented as Supplementary Tables (S1-S4) [Additional file 1]. The outcome variables that were found to be significantly associated are represented in Table S5. These supplementary Tables (S1-5) are accessible in the online version of this article. These significant variables were then entered into SEM models for each respective set of outcomes. The results of the models are represented in [Table 2], [Table 3], [Table 4], [Table 5]. The models were found to have a good fit with model fit statistic Chi-square test (all P > 0.05), comparative fit index (all models CFI >0.95), Tucker Lewis index (all models TLI >0.95) and root-mean-square approximate error <0.05 for all models.
Table 2: Results of SEM models of patient identifiers and personnel details completeness

Click here to view
Table 3: Results of SEM model of completeness of patients' demographics

Click here to view
Table 4: Results of SEM model of completeness of airway and anaesthesia details

Click here to view
Table 5: Results of SEM model of end of anaesthesia details completeness

Click here to view


The completeness of intubation technique entry in the airway management model was significantly predicted by anaesthesiologist grade and procedure type with estimates of −0.136 and −0.552, respectively. Thus, if a second year resident conducts an emergency case, the estimate would add up to −0.688, which denotes an OR of 0.503. This means that in such a case the chance of complete recording of intubation technique reduces by approximately 50%.

Overall, it was found that for most outcomes, phase 2 of AIMS use led to improvement in completeness compared to phase 1. Increase in anaesthesiologist grade (higher seniority) and case number (second case of the day) led to reduction in completeness of data entry. Emergency cases had poorer data completeness compared to elective cases. Also, anaesthesiologist grade significantly predicted complete entry of 18 (44%) variables, case number predicted 8 (20%) variables and phase- and procedure-type predicted 6 (15%) and 5 (12%) variables, respectively.


   Discussion Top


The use of AIMS has benefits such as legibility, faster data entry, reduction in human errors, enhanced data completeness, cost savings and easy access to previous records. However, the presence of AIMS by itself does not guarantee completeness and accuracy of anaesthesia information during surgery. This is predominantly observed during initial phase of introduction of AIMS from lack of familiarity and training. This study demonstrated that manual data entries in AIMS remains incomplete for many items and degree of incompleteness is observed more for the end-of-anaesthesia parameters. Our predefined factors were predictive of completeness of manual data entry in the AIMS.

An African audit of manual anaesthesia records noted only 30% (85/284) completion and for 71/284 (25%) anaesthetics, records were not used at all.[3] Driscoll et al. noted completeness in documentation of electronic anaesthesia record of 59% to 92% for six variables studied. They observed that dependence on free-text remarks and inability to automatically present entries in logical sequences by AIMS was associated with incomplete data entry.[8] In contrast, study examining introduction of context-sensitive mandatory fields in AIMS documented high (>98%) completeness rate and data concordance, and high rating for usability by anaesthesiologists.[9]

The wide variation in completeness of manual aspects of AIMS reported in the literature reflects deficiencies such as inadequate training prior to introduction of AIMS, lack of user friendliness, absence of mandatory field application and haphazard workflow of components requiring completion. Earlier studies have reported increased completion rate with education, workflow integration and individual feedback,[10] automated text prompts,[11] and including context-sensitive mandatory data entry fields.[9] Sandberg et al. observed significant improvement in completion of nonmandatory allergy information in the AIMS by implementing an automated text message to the user if no allergy information was documented within 15 min of the start of case.[11]

The suggested measures to improve data completeness of manual components of AIMS based on the literature and feedback from users in our study are 1) mandatory data entry fields for essential items 2) periodic training of anaesthesiologists in the use of AIMS to increase familiarity, 3) increase user-friendliness of commercially available system by customising record as per local needs, 4) restricting manual entry fields to minimum required 5) identify pathways to capture core items to avoid duplication 6) time-sensitive on-screen prompts to complete missing items 7) sign-out of chart by attending consultant (to ensure double checking) and 8) regular screening by medical records department for incompleteness and providing timeline for its completion. Our users provided feedback to mandatorily add certain missing elements into AIMS such as documentation of bilateral air entry in lungs on auscultation after intubation, number of attempts at intubation and mark of tracheal tube fixation. A need for developing filter to automatically remove artefacts was deemed necessary in place of the current option to manual recording of artefacts in our AIMS as several instances of artefacts being recorded as events are reported in the literature.[12]

We observed poor completeness with emergency cases. An Australian study analysing 850 anaesthesia records also reported poor completeness for emergency surgeries.[13] Similar findings were noted by Ige et al. for obstetric manual anaesthesia records.[14] These findings are understandable as the focus of clinicians is more on resuscitation and maintenance of patient's clinical condition with data completeness taking a back seat. Greater education and emphasis is required to improve completion rates in this population.

Our residents attended training more frequently than faculty on AIMS and, therefore, demonstrated more compliance with data completeness. The records of trainees are randomly verified by faculty but similar verification is absent for faculty. These factors might have contributed to better completeness among residents. A previous study noted no difference in completeness based on anaesthesiologist's age, level of training or number of years in practice.[15] However, close observation of anaesthesiologists during data entry increases completeness suggesting role of human behaviour during supervision of tasks.[16]

We noted that first case of the day had better completeness than the subsequent cases. Fresh start to the day, adequate time and clarity of workflow might have contributed to this finding. Improper hand-over between shifts, anaesthesiologist fatigue, extension of elective surgeries beyond routine work hours and increased events during latter part of the day could have contributed to poor completeness during the subsequent cases of the day.

We observed that phase-1 was associated with poor completeness of manual entries in the AIMS. It is likely that increased familiarity with time and more training sessions (two additional sessions during this period) contributed to increased completeness during phase-2 of study. An earlier study also documented improved completeness of manual record with passage of time. The percentage of adequately documented intraoperative records increased to 35.1% in 2014 in comparison to 25.5% in 2009.[17] Likewise, significant improvement in adequacy of documentation of anaesthetic record for obstetric spinal anaesthesia was noted after a teaching intervention.[18] Similar improvement for in-patient medical record completeness (from 73% to 84%) was seen after modifying the record format and training.[19]

The strength of our study is that this study evaluated more than 40 variables important to anaesthesia. Anaesthesia record is a faithful compilation of all aspects of peri-anaesthetic care and, therefore, should document all information that can help improve anaesthetic processes and reduce untoward outcomes. Therefore, we assessed all variables that we considered as important from this perspective, unlike previous studies that selectively examined few items such as drug entry[20] and six predetermined clinical documentation elements.[8] Second, we evaluated potential predictors that can affect completeness of anaesthesia record. As some of these factors are modifiable, addressing these issues is likely to improve completeness.

This study is not without limitations. We performed assessment of only one type of AIMS customised to a neuroanaesthesia setup after one year of installation, and hence, our findings may not be generalisable to other hospitals or systems or time frame. Secondly, we did not assess time-sensitiveness of the manual entries, which reflects accuracy of data entry in the AIMS. This requires a prospective audit and, hence, could not be performed in our current study. Thirdly, we did not compare manual record with our AIMS. This would have provided better insight into factors contributing to different completion rates.


   Conclusion Top


The completeness of manual data entry into the electronic AIMS is poor after one year of its implementation at a tertiary neurosciences centre. First case of the day, second phase of the study period, elective cases and trainee anaesthesiologist are associated with better completeness of the manual data recording into the AIMS. Frequent audit of the system and implementation of the above-suggested corrective measures is likely to increase completeness of manual data recording in the AIMS.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.



 
   References Top

1.
Yunuswangsa Q, Nimmaanrat S, Wasinwong W. Completion and accuracy in charting of anesthetic records in Songklanagarind hospital. J Med Assoc Thai 2008;91:1002-10.  Back to cited text no. 1
    
2.
Wrightson WAG. A comparison of electronic and handwritten anaesthetic records for completeness of information. Anaesth Intensive Care 2010;38:1052-8.  Back to cited text no. 2
    
3.
Raff M, James MFM. An audit of anaesthetic record keeping. Southern Afr J Anaesth Analg 2003;9:7-9.  Back to cited text no. 3
    
4.
Jang J, Yu SH, Kim CB, Moon Y, Kim S. The effects of an electronic medical record on the completeness of documentation in the anesthesia record. Int J Med Inform 2013;82:702-7.  Back to cited text no. 4
    
5.
Wilbanks BA, Moss JA, Berner ES. An observational study of the accuracy and completeness of an anesthesia information management system: Recommendations for documentation system changes. Comput Inform Nurs 2013;31:359-67.  Back to cited text no. 5
    
6.
R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. 2018. Available from: https://www.R-project.org. [Last cited on 2019 May 31].  Back to cited text no. 6
    
7.
Lavaan RY. An R package for structural equation modeling. J Stat Softw 2012;48:1-36.  Back to cited text no. 7
    
8.
Driscoll WD, Columbia MA, Peterfreund RA. An observational study of anesthesia record completeness using an anesthesia information management system. Anesth Analg 2007;104:1454-61.  Back to cited text no. 8
    
9.
Avidan A, Weissman C. Record completeness and data concordance in an anesthesia information management system using context-sensitive mandatory data-entry fields. Int J Med Inform 2012;81:173-81.  Back to cited text no. 9
    
10.
Vigoda MM, Gencorelli F, Lubarsky DA. Changing medical group behaviors: Increasing the rate of documentation of quality assurance events using an anesthesia information system. Anesth Analg 2006;103:390-5.  Back to cited text no. 10
    
11.
Sandberg WS, Sandberg EH, Seim AR, Anupama S, Ehrenfeld JM, Spring SF, et al. Real-time checking of electronic anesthesia records for documentation errors and automatically text messaging clinicians improves quality of documentation. Anesth Analg 2008;106:192-201.  Back to cited text no. 11
    
12.
Hoorweg AJ, 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;128:293-304.  Back to cited text no. 12
    
13.
Elhalawani I, Jenkins S, Newman N. Perioperative anesthetic documentation: Adherence to current Australian guidelines. J Anaesthesiol Clin Pharmacol 2013;29:211-5.  Back to cited text no. 13
[PUBMED]  [Full text]  
14.
Ige OA, Adesina KT, Fatoba MA. Completeness of manual anaesthesia records in a tertiary facility in Nigeria. J Med Trop 2017;19:86-9.  Back to cited text no. 14
  [Full text]  
15.
Devitt JH, Rapanos T, Kurrek M, Cohen MM, Shaw M. The anesthetic record: Accuracy and completeness. Can J Anesth 1999;46:122-8.  Back to cited text no. 15
    
16.
Edwards K, Hagen SM, Hannam J, Kruger C, Yu R, Merry AF. A randomized comparison between records made with an anesthesia information management system and by hand, and evaluation of the Hawthorne effect. Can J Anesth 2013;60:990-7.  Back to cited text no. 16
    
17.
Curtis WB, Sethi R, Visvanathan T, Sethi S. Peri-operative anaesthetic documentation: A report of three sequential audits on the quality of outcomes, with an insight into surrounding legal issues. Turk J Anaesthesiol Reanim 2018;46:354-61.  Back to cited text no. 17
    
18.
Olateju SO, Adenekan AT, Owojuyigbe AM. The effect of teaching on the completeness of the anesthesia record charts for obstetric subarachnoid blocks in a low resource area hospital. J Obstet Anaesth Crit Care 2015;5:16-21.  Back to cited text no. 18
  [Full text]  
19.
Tola K, Abebe H, Gebremariam Y, Jikamo B. Improving completeness of inpatient medical records in Menelik II referral hospital, Addis Ababa, Ethiopia. Adv Public Health 2017. doi: 10.1155/2017/8389414.  Back to cited text no. 19
    
20.
Avidan A, Dotan K, Weissman C, Cohen MJ, Levin PD. Accuracy of manual entry of drug administration data into an anesthesia information management system. Can J Anesth 2014;61:979-85.  Back to cited text no. 20
    



 
 
    Tables

  [Table 1], [Table 2], [Table 3], [Table 4], [Table 5]



 

Top
 
  Search
 
    Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
    Access Statistics
    Email Alert *
    Add to My List *
* Registration required (free)  

 
  In this article
    Abstract
   Introduction
   Methods
   Results
   Discussion
   Conclusion
    References
    Article Tables

 Article Access Statistics
    Viewed171    
    Printed2    
    Emailed0    
    PDF Downloaded44    
    Comments [Add]    

Recommend this journal