Electronic health record (EHR) nursing documentation offers a method to record the patient's health status, individual needs, and responses to care, and to support clinical reasoning regarding the patient's future care.1 Additionally, the EHR has the potential to improve patient care, care team communication, and patient experiences through well-designed workflows and clinical content relevant to the role, venue, and patient's clinical status. The objective of this quality improvement study was to evaluate whether EHR system timers and event logs can measure the efficiency and quality of a clinical process in an EHR.
BACKGROUND AND SIGNIFICANCE
Over the last few decades, nurses have increasingly been burdened with documentation to meet regulatory and quality reporting requirements. This reporting requires nurses to collect data not intrinsically linked with knowledge and context of the patient's story and potential risk factors.2,3 Nurses play a crucial role in patient care and performance improvement in the healthcare system. This role requires documenting and managing patient information through coordinating care and communicating with other healthcare professionals.4
The nursing profession is striving to standardize care processes with clinically relevant content while considering the complexity of care, conditions in the environment, market and regulatory changes, and clinician preferences.5,6 These factors increase the need for health information technology (IT) professionals to have a mechanism to measure the effectiveness, efficiency, and quality of clinical processes supported within the EHR.
Today's health information systems capture significant data in event and timer logs. Combined with EHR data, these logs allow examination of temporal patterns of care, including mining, modeling, and measuring the clinician's experiences within the EHR. Event logs are extracts from EHRs that comprised lists of time-stamped process-steps created as a by-product of operations.7
Researchers and informatics professionals can use EHR logs to evaluate historical content, workflow design, and utilization patterns, including time and clicks required for an EHR workflow. Wu et al8 completed a study that used EHR audit logs, and found that these logs are a valid source for workflow analysis and can provide an objective view of clinician behaviors. Mining process information directly extracted from these logs can provide insight into how healthcare processes are executed.9 For the event logs to be useful, investigators must understand which events and timers are relevant to measure processes related to a clinician's role, venue, quality, and sometimes the patient's clinical condition.
Chen et al10 found that existing methods were insufficient, resource-intensive, and time-consuming in measuring the influences of EHRs regarding end-user adoption, efficiency, and utilization of workflows. Chen and colleagues introduced a framework to infer clinical workflows through the utilization of an EHR. This framework provides a method to generate workflows at multiple levels of granularity through data mining of EHR event logs, allowing healthcare systems and health informatics organizations to evaluate and refine inefficient and unstable workflows. The researchers categorized workflows into four classes based on their duration and variance:
- stable efficient blocks: a short average duration with small variance;
- unstable efficient blocks: a short average duration with large variance;
- stable inefficient blocks: a long average duration with small variance; and
- unstable inefficient blocks: a long average duration with large variance.10
The researchers selected the nursing admission patient history (APH) for this study because it meets the classification of a stable and efficient process. Prior to this quantitative quality improvement study, a qualitative research team consulted an expert panel of 20 chief nursing officers and chief nursing informatics officers, which included a doctor of nursing practice and PhD to establish an adult APH essential clinical data set (APH ECD). The panel represented a cross section of 12 healthcare organizations, which accounted for approximately 46 000 acute care beds and 192 hospitals, including a mix of academic medical centers, integrated delivery networks, and community hospitals.
The research team used a modified Delphi approach to analyze the historical form design and 3 months of utilization patterns for adult APHs. The research team pulled timer and discrete data for APHs of 12 organizations and analyzed each organization's admission history form and the average utilization patterns for the nursing documentation of patient histories. Based on Weiskopf and Weng's11 dimensions of quality as well as federal regulatory requirements, the research team defined an essential standardized data set of 40 elements for the adult APH that met the intended use and were of sufficient quality for a patient history assessment on admission.
This study was intended to measure the quantitative impact of implementing the reduced number of essential data elements utilizing EHR timers and event logs. The authors had two objectives for this study: (1) to determine whether system timers and event logs can measure the preintervention and postintervention status of a defined clinical workflow, and (2) to determine whether introducing a standardized APH ECD improves nursing efficiency and quality of documentation.
We hypothesize that reducing this data set to essential data elements will (1) increase the proportion of essential items completed, (2) decrease the documentation time required to complete the APH, and (3) reduce the number of mouse clicks to complete the APH.
MATERIALS AND METHODS
The research setting was a 600-bed academic medical center in the United States. The study included medical-surgical, ICU, emergency department, step-down, and telemetry clinical documentation for adult APHs performed by RNs, with a focus on US-based practices at the federal level. The research site's nursing leadership and informatics team compared their baseline APH 215 data elements to the APH ECD 40 data elements and adjusted for local and state requirements. This review resulted in the study site reducing its APH from 215 data elements to 58 data elements, reflecting a 73% reduction in its APH ECD.
The quality improvement research site integrated the APH ECD into their EHR for acute adult inpatient units. We collected 1 month of data preintervention and postintervention to measure nursing efficiency and quality measures to evaluate the impact of the APH ECD. We utilized timers and event logs embedded in the software to capture duration and the number of data elements documented by each RN.
Data sets were extracted from the EHR event files that included timer data. The timers systematically captured the beginning and end of the APH documentation. The system timer data were validated by an electronic video recording of a sample of nurses completing the forms and comparing the system timers to the video recording timers. Event files were also utilized to measure the data elements captured.
For the preintervention data set, the investigators analyzed baseline measures for patient admission histories completed over 30 days prior to implementation of the ECD in September 2017. The timer data were captured for each episode of care, which was the unit of analysis. The data set included all data elements in the APH form, the number of data elements documented, and time and number of mouse clicks required to complete the APH for each encounter.
The APH ECD was placed in production (ie, viewable to clinicians) on November 1, 2017. The postimplementation period included 30 days of APHs completed from November 20, 2017, to December 17, 2017. Data from November 1 to November 19 were excluded to allow nurses to adjust to the APH ECD.
To assess quality, the difference in the proportion of each of the 58 essential data elements captured before and after introducing the APH ECD was evaluated. This measure was based on the premise that if there are fewer total elements in the patient history, the most important elements were likely documented. The researchers included all assessments completed in preintervention (n = 904) and postintervention periods (n = 805).
To assess efficiency, we measured (1) the time it took, in minutes; and (2) the number of mouse clicks for nurses to complete an APH before and after implementation of the APH ECD. For the efficiency measures, the researchers included all assessments completed in one-sequence preintervention (n = 536) and postintervention periods (n = 640), due to limitations of the timer to measure time and clicks for assessments only partially completed initially and then completed later.
Descriptive statistics were used to evaluate the differences preintervention and postintervention on baseline characteristics, the percentage of essential data elements captured, and the number of clicks and time to complete the APH. The significance of the difference between means was determined using an independent-samples t test. Descriptive and statistical calculations were performed using SAS version 9.4 (SAS Institute, Cary, NC). This quality improvement study was deemed nonhuman research by the research site's institutional review board (IRB); no IRB approval was required.
Table 1 reflects the descriptive statistics for preintervention and postintervention quality and efficiency measures. Quality is measured by the number and percentage of essential variables captured as well as the proportion of APH completed in one sequence. These quality measures included all APHs during the study period. Efficiency measures included the mean number of clicks and the time spent completing an APH, which was performed on the subset of APHs that were completed in one sequence.
To assess quality, the difference in the number of the 58 essential data elements captured for the APH after introducing the APH ECD was evaluated. Figure 1 shows a statistically significant increase in the number of essential data elements captured from the preintervention measure (mean = 48% [SD = 0.10]) to the postintervention measure (mean = 54% [SD = 0.11]; t = −10.83, P = <.0001, two-tailed). The percentage of capture of data elements considered essential increased from 48% to 54%. In addition, the number of APHs completed in one sequence, without saving and signing or modifying later, increased by 24% from the preintervention measure (mean = 62% [SD = 0.48]) to the postintervention measure (mean = 86% [SD = 0.35]).
We analyzed the nurses' total time and clicks from start to completion of an APH for all patients admitted within the preintervention and postintervention periods. Only patient histories completed in a single sequence were included in this analysis. If a patient history was partially completed initially and then completed later, the history was excluded from the analysis due to timer data limitations. The time in minutes and the number of clicks to complete the APH were assessed using descriptive statistics.
The average active time spent documenting the APH decreased by 72%. Figure 2 shows the decrease in active time from the preintervention measure (mean = 9.30 [SD = 4.66] minutes) to the postintervention measure (mean = 2.55 [SD = 1.65] minutes), resulting in a difference of 6.76 minutes (t = 31.95, P = <.0001, two-tailed).
The average number of clicks to document the APH decreased by 76%. We found a statistically significant decrease in the number of clicks from the preintervention number of clicks (mean = 151.51 [SD = 37.35]) to the postintervention number of clicks (mean = 35.93 [SD = 15.01]; t = 67.40, P = <.0001, two-tailed). The mean decrease in the number of clicks was 115.6 with a 95% confidence interval and with a range of 112.2 to 118.9, P < .0001. The mean differences of both the time and clicks represent statistically significant changes that may have a noticeable impact on clinical workflows and efficiency.
The automation of clinical workflows provides new opportunities to measure efficiency and quality. Historically, researchers have used time and motion studies, observations, surveys, retrospective chart reviews, and structured and semistructured interviews to measure the effect of technology on the nursing process, documentation quality, and workflow.12–15 This study supports Wu and colleagues'8 and Chen and colleagues'10 findings that today's health information systems have significant data captured in event and timer logs that could research the influences of EHRs regarding end-user adoption, efficiency, and utilization of workflows.
This study demonstrates that, by analyzing EHR event files, health systems could discover whether automated clinical processes are meeting the intended objectives from a quality and efficiency perspective.10 This study also demonstrates that EHR event files are an additional source to support an objective analysis of workflow, clinicians' behaviors related to EHR use, and efficiency measures. This information can support quality and efficiency initiatives and increase knowledge and understanding of the clinician's experience.
This approach has not been used to evaluate the quality and efficiency of nursing documentation, so we cannot compare it to previous studies. However, we found it very effective to measure the impact of quality and efficiency of changing content within an EHR nursing process. Health systems and researchers are striving to standardize processes while considering the complexity of care, conditions of the environment, and clinician preferences.
This study demonstrates that an analytical framework maximizing the use of EHR event files can provide practice-based insights into clinical workflows and interventions, with added insight into the context of data collected during clinical care and the impact on clinical decision making. This framework can support a health system's ability to use EHR data for discoveries, creating data-driven contextual intelligence to develop strategies to optimize and advance knowledge related to clinical practice in an EHR environment.
Future studies could advance knowledge for healthcare informatics professionals and researchers about the evolution of EHR event logs and timers to measure end-user adoption, process, and outcome measures. Additional factors that could be measured include clinician experience as it relates to communication between nursing and other care providers, measurable secondary use of the data, and the impact on patient outcomes.
Further research should include the downstream use of clinical documentation by other members of the care team to support team-wide clinical reasoning and decision making and to improve patient outcomes. In future studies, we will include physical assessment documentation and how the care team utilizes the information. Documentation is a communication method with care team members; a mixed-methods study would add additional information related to the value the care team assigns to documentation as an effective method to support communication, collaboration, and coordination of care.
The results of this study may not be generalizable, especially if a health system has no EHR with timers and event logs to measure quality and efficiency. However, this work will inform future research to determine the impact and feasibility of an optimized approach to documentation, using system timers and event logs for preintervention and postintervention measurements.
This study occurred in an organization that used an EHR with embedded timers and event logs capable of discrete data capture. All EHRs may not have the same features embedded into the clinical workflows. Additionally, if the event logs exist, they may have limitations. For example, for this research project, the event logs did not allow the researchers to measure time and clicks if the provider documented, saved the form, then returned later to complete/sign the form. The internal timers could not collect the actual duration of intermittent documentation episodes. Due to this technical limitation, only uninterrupted documentation instances were included in the analysis of efficiency. Future research should examine whether timers can be adapted to capture breaks in time and ensure storage of that data.
Conditional logic needs to be considered when measuring quality. The essential data elements in an EHR event file might be relevant for a particular patient and not others. This functionality is called conditional logic, and it streamlines documentation by not displaying or requiring documentation of data elements unless relevant. For example, when nurses document patient belongings, a patient may have glasses and hearing aid, but not a cane. The nurse would select only the belongings relevant to the patient. Therefore, 100% capture of essential data elements could not be expected, because some data elements are not relevant for all patients.
This study did not assess how the content captured in the APH was utilized by other clinicians to support clinical reasoning or to meet regulatory requirements. A method that could provide insight into nursing documentation used by other team members would be to study the downstream use of the information captured in workflows, such as patient summaries, in other care team workflows and reporting or during clinical decision support. Those key components support care team communications and collaboration and warrant further study. Future studies should address the multiple factors that affect nursing efficiency so hospital systems can maximize the benefits of EHRs.
Both the literature and anecdotal evidence from nurses indicate that EHR systems must be revised to reduce the time spent on documentation so nurses can spend more time on patient care. The first step is to assess the characteristics of EHR systems and redesign them to optimize nursing processes by including only essential documentation to support clinical practice.11 Automation of clinical workflows in EHRs has created new opportunities for clinical informaticists to measure and evaluate the impact of health IT on clinical workflow effectiveness, efficiency, utilization, and quality of workflows within an EHR.
Changing an EHR clinical workflow is a multifactorial process. The outcome could be influenced by staff training, years of experience of staff, the acuity of the patients, the physical environment of the study, and the organizational culture of the study site. While patterns and trends can be measured, proving cause and effect may be influenced by more than the changes in content. Clinical informatics professionals should consider the use of EHR event files and timers to gain insight into process and workflow changes. The use of system data can substantiate the transformational value of informatics practice and inform future optimization efforts. The results may be the first step toward achieving the early promise regarding the value of EHRs from more than 20 years ago.
The authors thank Amy Peters and Cathy McFarland from the study site for their total support and belief that this study will make a difference to clinicians and patients. Also, they thank Darinda Sutton, Cameron Johnson, Carly Evans, and Lindsey Jarrett for their partnership on this study.
1. Akhu-Zaheya L, Al-Maaitah R, Bany Hani S. Quality of nursing documentation
: paper-based health records versus electronic-based health records. Journal of Clinical Nursing
. 2018;27(3–4): e578–e589.
2. Okaisu EM, Kalikwani F, Wanyana G, Coetzee M. Improving the quality of nursing documentation
: an action research project: original research. Curationis
. 2014;37(2): 1–11.
3. Kruse CS, Kothman K, Anerobi K, Abanaka L. Adoption factors of the electronic health record
: a systematic review. JMIR Medical Informatics
. 2016;4(2): e19.
4. Kahn MG, Callahan TJ, Barnard J, et al. A harmonized data quality assessment terminology and framework for the secondary use of electronic health record
data. EGEMS (Washington, DC)
5. Gocsik T, Barton AJ. Workflow optimization: essential in the electronic health record
journey. Clinical Nurse Specialist
. 2014;28(3): 132–134.
6. Ghattas J, Soffer P, Peleg M. A formal model for process context learning. In: International Conference on Business Process Management. Berlin, Heidelberg, Germany: Springer; 2009:140–157.
7. Fox F, Aggarwal VR, Whelton H, Johnson O. A Data Quality Framework for Process Mining of Electronic Health Record
Data. Paper presented at Healthcare Informatics (ICHI), 2018 I.E. International Conference; 2018.
8. Wu DTY, Smart N, Ciemins EL, Lanham HJ, Lindberg C, Zheng K. Using EHR audit trail logs to analyze clinical workflow: a case study from community-based ambulatory clinics. American Medical Informatics Association Annual Symposium Proceedings
. 2018;2017: 1820–1827.
9. Mans RS, van der Aalst WM, Vanwersch RJ, Moleman AJ. Process mining in healthcare: data challenges when answering frequently posed questions. In: Process Support and Knowledge Representation in Health Care
. Berlin, Heidelberg, Germany: Springer; 2013: 140–153.
10. Chen Y, Xie W, Gunter CA, et al. Inferring clinical workflow efficiency via electronic medical record utilization. AMIA Annual Symposium Proceedings/AMIA Symposium
. 2015;2015: 416–425.
11. Weiskopf NG, Weng C. Methods and dimensions of electronic health record
data quality assessment: enabling reuse for clinical research. Journal of the American Medical Informatics Association
. 2013;20(1): 144–151.
12. Hendrich A, Chow MP, Skierczynski BA, Lu Z. A 36-hospital time and motion study: how do medical-surgical nurses spend their time? The Permanente Journal
. 2008;12(3): 25–34.
13. Collins SA, Fred M, Wilcox L, Vawdrey DK. Workarounds used by nurses to overcome design constraints of electronic health records. NI 2012 (2012)
2012;2012: 93. eCollection 2012.
14. Banner L, Olney CM. Automated clinical documentation: does it allow nurses more time for patient care? CIN: Computers, Informatics, Nursing
. 2009;27(2): 75–81.
15. Larrabee JH, Boldreghini S, Elder-Sorrells K, et al. Evaluation of documentation before and after implementation of a nursing information system in an acute care hospital - improving patient record data validity. Computers in Nursing
. 2001;19(2): 56.