SELECT caregivers.cgid,
COUNT(*) AS chart_count
FROM caregivers
INNER JOIN chartevents
ON caregivers.cgid = chartevents.cgid
WHERE caregivers.label = "RN"
GROUP BY caregivers.cgid
HAVING chart_count > 100
ORDER BY chart_count DESC
LIMIT 10Analysis Report Three - Information Systems and Healthcare Provider Interactions
Executive Summary
Every minute that passes, healthcare entities produce vast amounts of new data surrounding patient encounters from electronic health records, use of clinical documentation, patient monitoring logs, laboratory test results, and medication administration logs. When leveraged properly, those pieces of information hold incredible capability in contributing to the betterment of patient care as well as improving operational efficiencies of organizations through data-driven decision-making. However, many challenges remain for executives, managers, and their organizations regarding the documentation burden on clinicians; clinician burnout; data quality; workflow design; and improving patient outcomes via technology. Therefore, using evidence-based articles that outline this technology being used as less of a result to deliver clinical care, and finding ways to support the clinical judgment of healthcare professionals. National Academy of Medicine, American Medical Association, AHRQ, and many peer-reviewed articles demonstrate that organizations need to manage EHR systems to lower administrative burdens placed on their staff and thus improve clinician well-being. Through literature analysis and hands-on analysis utilizing a family-created SQL Database Querying program (MIMIC-III), I provide insights from the literature combined with statistical data to show how caregiver activity occurs as both a function of the caregiver’s employment role and level of caregiver output. Both sets of data provide evidence that will allow healthcare organizations using the MIMIC-III Database to develop meaningful metrics for staffing strategies and workflow improvements and establish an evidence-based management framework.
Introduction
The introduction of electronic health records has brought about changes to the healthcare system through improved communication, greater access to patient information, and the facilitation of evidence-based clinical decision-making. EHR systems function as a means by which providers can retrieve relevant patient information more effectively; this, in turn, facilitates improved coordination among departments and enhances the continuity of care. However, according to their study, clinicians often encounter an increased number of documentation requirements, interruptions to workflow, inconsistent design of systems, and lack of usability in the system, and the researchers noted that rather than reducing workloads through the use of EHR systems, frontline clinicians are often taking on even more administrative duties as a result of EHRs. This situation was corroborated by a recent study conducted by the National Academy of Medicine (2020), wherein both the lack of usability within EHR systems as well as the need for excessive documentation were cited as key factors contributing to clinician burnout(Rathert et al. 2019; National Academy of Medicine 2024). A systematic review published in JMIR Medical Informatics (2024) similarly suggested that healthcare providers who spend a greater amount of their time on completing electronic documentation are likely to experience a greater level of job dissatisfaction than those who are not subjected to unnecessary administrative levels of documentation. As healthcare organizations develop more sophisticated algorithms utilizing artificial intelligence, thereby advancing predictive analytics and creating more advanced clinical decision support systems, it is imperative that leaders utilize technology to support the delivery of clinical care, as opposed to creating new burdens on the administrative side of clinical care.(Bannon 2023)
The Healthcare Context
Today, health care organizations are collecting more information than ever before. Electronic health records, laboratory information systems, pharmacy databases, wearable devices/equipment, patient monitoring equipment, and other clinical data sources are available to help improve patient outcomes and operational performance. However, the mere collection of these data does not guarantee improvement in health care delivery.(McAlearney et al. 2015) Rathert et al. showed that while clinicians may appreciate the ability to instantly access their patients’ records, they report that they spend an inordinate amount of time documenting care, rather than spending more time interacting with patients. Many clinicians also report that, despite the benefits of an electronically based health care record system, their clinical environment consists of inconsistent workflows, a lack of proper training, and “gray areas” due to EHR systems not supporting actual clinical practice. Although workarounds may temporarily create greater efficiency, they also tend to create decreased standardization, thus introducing new patient safety hazards.
The excessive amount of documentation also continues to be one of the largest contributors to physician burnout, as stated by the American Medical Association. Because of the administrative responsibilities related to EHR systems, physicians are spending a significant portion of their workday on administrative tasks and will therefore have less time to devote to direct patient care, which, in turn, will cause them to be less satisfied with their profession. The Agency for Healthcare Research and Quality has also stated that employer-imposed burnout among physicians is a serious issue threatening the provision of quality health care, continuation of health care providers, and patient safety.(Agency for Healthcare Research and Quality 2024) The piece assigned to you demonstrates how tech is changing the way EHRs make decisions in a healthcare setting through AI-assisted processes. AI has the ability to help healthcare organizations quickly assess the large amounts of clinical data on individual patients and find trends and patterns from that data that might not be obvious. However, AI should only assist clinicians in decision-making as an additional source of information; therefore, clinician engagement as an active participant in the decision-making process must be maintained through the use of AI in conjunction with clinician expertise. The National Academy of Medicine recently indicated that optimizing EHR systems requires improving usability, streamlining documentation processes, and providing input from frontline clinicians during the implementation phase. In other words, instead of measuring the success of EHR technology by the amount of data they collect, healthcare leaders need to assess how the data collected from EHR technology can be used to improve patient care and decrease the number of administrative tasks placed on clinicians.
Data Visualizations
Visualization 1:
ggplot(data = myquery1,
aes(x = reorder(cgid, chart_count),
y = chart_count)) +
geom_col() +
coord_flip() +
theme_minimal() +
labs(
title = "Top 10 RN Caregivers by Charted Events",
x = "Caregiver ID",
y = "Number of Charted Events"
)
To create the first visualization of the analysis, a join was performed between the CAREGIVER table and the CHARTEVENT table to show the number of charters recorded by each Registered Nurse caregiver. The methods associated with SQL that are applicable for this week’s practice include INNER JOIN, COUNT, GROUP BY, HAVING, and ORDER BY. The aggregate data will be presented in geom_col format. As shown on the graph, the documentation activity is not the same between caregivers. There are some RN caregivers who documented far more charted events than other caregivers. This disparity may suggest differences in workload, patient assignment, shifts, or how caregivers document. Although it cannot determine precisely why there are differences between RNs, it shows how healthcare administrators can utilize operational data to identify potential workflow mismatches that would require additional review.
This finding is supported by the literature. Rathert et al. describe the documentation burden as one of the greatest challenges for healthcare providers who utilize electronic health records(Rathert et al. 2019). The American Medical Association also supports this by encouraging organizations to evaluate EHR activity data on a regular basis to identify opportunities to cut unnecessary documentation and enhance clinician effectiveness. By looking at documentation counts beyond merely being a measure of performance, healthcare organizations can leverage the data to determine staffing requirements, change workflows, and lessen the administrative burden.
Visualization 2:
SELECT caregivers.label,
AVG(CAST(outputevents.value AS INTEGER)) AS avg_output
FROM caregivers
INNER JOIN outputevents
ON caregivers.cgid = outputevents.cgid
GROUP BY caregivers.label
HAVING avg_output > 0
ORDER BY avg_output DESC#Put your ggplot visualization in this block
ggplot(data = myquery2,
aes(x = reorder(label, avg_output),
y = avg_output)) +
geom_col() +
coord_flip() +
theme_minimal() +
labs(
title = "Average Output Value by Caregiver Role",
x = "Caregiver Role",
y = "Average Output Value"
)
The second visualization is a combination of the CAREGIVERS and OUTPUTEVENTS tables; average recorded output values for each of the different caregiver roles are displayed. In contrast to the first visualization, which primarily counted documentation events, this analysis demonstrates the use of AVG(), CAST(), GROUP BY, and HAVING in order to summarize quantitative data. The secondary representation of average recorded output values demonstrates variation among average recorded output values for different caregiver roles, however this variation should not be interpreted as indicative of caregiver performance because it does provide evidence of how operational data can be aggregated to evaluate differences in clinical activity across provider groups(McAlearney et al. 2015). A notable example of this variation is the lack of standardization in the labels for caregiver roles that appear to have different case sensitivities including “RN”, “Rn”, and “rn”. While inconsistencies in data entry practices do not directly impact the quality of patient care, they do pose a significant challenge for healthcare organisations regarding the ability to report, display, and analyse the data. HIMSS and the National Academy of Medicine both note that the standardization of documentation practices will improve the quality of clinical data and enhance the effectiveness of healthcare analytics. Therefore, this visualization illustrates not only how quantitative data can be summarised but also illustrates why healthcare organisations should make the standardisation of clinical data their highest priority when establishing their enterprise analytics program.(Pham et al. 2024)
Visualization One
SELECT caregivers.cgid,
COUNT(*) AS chart_count
FROM caregivers
INNER JOIN chartevents
ON caregivers.cgid = chartevents.cgid
WHERE caregivers.label = "RN"
GROUP BY caregivers.cgid
HAVING chart_count > 100
ORDER BY chart_count DESC
LIMIT 10ggplot(data = myquery1,
aes(x = reorder(cgid, chart_count),
y = chart_count)) +
geom_col() +
coord_flip() +
theme_minimal() +
labs(
title = "Top 10 RN Caregivers by Charted Events",
x = "Caregiver ID",
y = "Number of Charted Events"
)
Visualization Two
SELECT caregivers.label,
AVG(CAST(outputevents.value AS INTEGER)) AS avg_output
FROM caregivers
INNER JOIN outputevents
ON caregivers.cgid = outputevents.cgid
GROUP BY caregivers.label
HAVING avg_output > 0
ORDER BY avg_output DESC#Put your ggplot visualization in this block
ggplot(data = myquery2,
aes(x = reorder(label, avg_output),
y = avg_output)) +
geom_col() +
coord_flip() +
theme_minimal() +
labs(
title = "Average Output Value by Caregiver Role",
x = "Caregiver Role",
y = "Average Output Value"
)
Recommendations for Industry
Healthcare providers must keep putting money to work with technologies that assist clinicians with terms of clinical decision-making and decreasing the work needed for unnecessary paperwork. The focus for leaders with successes should be on improving the quality and usability of healthcare data rather than measuring success by the amount of data collected for use. To measure the extent and frequency of the work being done through caregiver documentation, organizations should use operational dashboards to monitor caregiver documentation activities, as shown in the first visualization. Identifying an unusual quantity of workload in terms of documentation may address staffing imbalances or workflow areas that contribute to clinician burnout. In order to be able to produce data that is useful for reporting purposes and produce reliable data for analytics, organizations need to establish documentation standards for all departments as an initiative for data governance. When there are changes within technology, healthcare leaders must engage their frontline clinicians when designing, developing, and implementing any type of new systems. Many studies demonstrate that clinician involvement in system design will improve usability and reduce workarounds and documentation burden. Organizations will continue to expand their clinical decision-making, support tools, and artificial intelligence systems while ensuring the clinician remains in charge. Technology can be a tool to support clinical judgment, not replace it. Therefore, any healthcare administrator who can successfully balance technology, usability, and staff wellness can more efficiently enhance the delivery of care through improving clinical outcomes.(American Medical Association 2024)
##Conclusion Healthcare organizations today have access to more types of clinical data than ever before, and while the challenge of collecting this data has changed, the challenge of transforming the available information into useful knowledge that will improve patient care and help an organization perform at its best remains. The analysis contained within this report illustrates how SQL query language and visualization techniques utilize the MIMIC-III database to summarize both how clinicians record their activities and what kind of data is being created. There is no doubt that healthcare technology improves how care is delivered or that it creates new ways of rendering clinics as administrators. The assigned readings and numerous published research articles emphasize the critical role that technology plays in improving the delivery of care while presenting significant challenges to the administration of those clinics. Major administrative issues currently faced by healthcare administrators as a result of the integration of technology into the healthcare delivery process include clinician workload burden, clinician fatigue, data quality, and workflow design. Leveraging effective data governance, engaging clinicians in their clinical decision-making, and applying evidence-based analytics enable healthcare organizations to utilize EHR data to achieve both operational excellence and quality patient care.
References
Agency for Healthcare Research and Quality. 2024. “Physician Burnout.” https://www.ahrq.gov/prevention/clinician/ahrq-works/burnout/index.html.
American Medical Association. 2024. “Digging into the Data to Cut EHR Burdens That Drive Burnout.” https://www.ama-assn.org/practice-management/digital-health/digging-data-cut-ehr-burdens-drive-burnout.
Bannon, Lisa. 2023. “Who Makes Decisions in a Crisis: The Human or the Machine?” The Wall Street Journal.
McAlearney, Ann Scheck, Jennifer L. Hefner, Cynthia J. Sieck, and Timothy R. Huerta. 2015. “The Journey Through Grief: Insights from a Qualitative Study of Electronic Health Record Implementation.” Health Services Research 50 (2): 462–88. https://doi.org/10.1111/1475-6773.12227.
National Academy of Medicine. 2024. “Electronic Health Record Optimization and Clinician Well-Being: A Potential Roadmap Toward Action.” https://nam.edu/perspectives/electronic-health-record-optimization-and-clinician-well-being-a-potential-roadmap-toward-action/.
Pham, Minh et al. 2024. “Evaluating the Prevalence of Burnout Among Health Care Professionals Related to Electronic Health Record Use: A Systematic Review.” JMIR Medical Informatics 12: e54811. https://doi.org/10.2196/54811.
Rathert, Cheryl, Tracy H. Porter, Jessica N. Mittler, and Michelle Fleig-Palmer. 2019. “Seven Years After Meaningful Use: Physicians’ and Nurses’ Experiences with Electronic Health Records.” Health Care Management Review 44 (1): 30–40. https://doi.org/10.1097/HMR.0000000000000168.