SELECT caregivers.label, COUNT(*)/1000 AS chart_count
FROM chartevents
INNER JOIN caregivers
ON chartevents.cgid = caregivers.cgid
WHERE caregivers.label = "RN"
OR caregivers.label = "IMD"
OR caregivers.label = "RRT"
GROUP BY caregivers.label
HAVING chart_count > 1
ORDER BY chart_countAnalysis Report Three - Information Systems and Healthcare Provider Interactions
Executive Summary
Electronic health record systems and AI driven clinical tools have been integrated into how healthcare is delivered. These tools have shown many benefits but also challenges that need to be addressed. Many healthcare organizations still do not understand the full functionality of the EHR system that has been adopted and do not use it to it’s full capacity, even though it has been years since the integration (Rathert et al. 2017).When reviewing and analyzing the articles/references listed in this report, they all correlate to one central idea, technology advancements alone do not improve healthcare. In order to have success in technological advances, organizations have to invest in their people first. Providers have been and continue to be asked to adopt complex EHR systems without in depth training and support for their day to day workflow in patient care. Some of the common issues I found in the articles lack of training and ongoing training for EHR systems, poor communication between systems, for example, EPIC EHR and CERNER EHR systems, and overall lack of support for the clinical team when learning about new systems and system upgrades (Rathert et al. 2017; McAlearney et al. 2015). Providers emphasized how they felt rushed into learning a EHR system, new documentation, etc while still being responsible for maintaining exceptional patient care (McAlearney et al. 2015). Although it has been almost 20 years since the HITECH acts was enforced and created a ‘pivotal moment’ in healthcare as far as EHR systems go (Morris 2025), many clinicians have said they still have to create their own shortcuts and agenda just to get through their average day because the EHR systems do not match the way care is actually delivered (Rathert et al. 2017). When taking this into consideration, healthcare executives have to consider the people using these systems everyday and realize they need and deserve better support for integration, upgrades, and their input in what needs to be adjusted and improved within their workflows. Without a heavy support system in place, even the most advanced technology, or those using that technology, will fall short of all it is meant to do.
Introduction
Rathert and her coauthors conducted a qualitative study to obtain a better understanding on how physicians and clinical staff were experiencing EHR systems in healthcare organizations that had been using them for many years (Rathert et al. 2017). In the study, they chose to focus on providers who were no longer in the initial learning phase of the EHR systems but rather in the ‘Stage 3: Meaningful Use’ phase where providers were actively working in the EHR systems. In this study, they focused on 30 clinicians all together, half providers the other half clinical staff/nurses and used a ‘thematic framework’ that helped them identify strengths and challenges in the EHR system (Rathert et al. 2017). When looking over this study, it was shown that the core issues are that even though it has been nearly 20 years since the EHR was implemented, there is still continued challenges today. One of the most frequent concern that was mentioned in this article/study was the use of workarounds. Clinicians would find ways to ‘work around’ the EHR system, for example, printing the last note from the prior visit, writing the new visit data down on a piece of paper and then entering the data in the system at the end of clinic hours , although many clinicians are not open about this workaround phrase (Rathert et al. 2017). This article/study also discussed the Conservation of Resources (COR) theory as a way to better explain why clinicians respond to EHR challenges the way they do. When the EHR system demands outweigh the support system providers need, this is when providers tend to shut down and workaround the system, rather than ask for help (Rathert et al. 2017). When looking at this from a high-level overview, some providers adapt and work through the challenges to better understand the EHR system and new technologies to embrace how it can improve their work and impact on patient care, some are not there yet and still see the new technologies as a hindrance to quality patient care. Some see EHR systems and new technologies as an increased workload, privacy and security concerns and workflow disruptions where as others see it as improved patient care and an increase in efficiency with data collection (Meigs 2016). Overall, this article and the research that is in it makes a clear point that successful EHR use is not only about having the right technology and newest EHR system, but ensuring those who will be using the technology understand how to use it effectively and efficiently.
The Healthcare Context
Healthcare organizations across the country continue to struggle with the same EHR challenges that have persisted for years, and in many cases, continue to get more complicated with more technologies being introduced (Janett and Yeracaris 2020). One of the most common issues is the ‘lack of interoperability’ between different systems, where providers at different organizations on different systems cannot easily access patient information which impacts the quality of direct patient care (Rathert et al. 2017). Providers complain about information being in the system but not easily accessible when and where they need it. Although these EHR technology issues are persistent, AI driven clinical tools are being rolled out in healthcare organizations, such as documentation tools and flag tools that notify of sepsis risk and medication alerts (Bannon 2023). Although these tools, if used properly, have a positive impact for clinicians and how they make medical decisions, if they rely to heavily on AI and the healthcare organization does not have clear, specific policies in place, it could lead to a severe patient endangerment, sometimes even death (Bannon 2023). This relates back to what Rathert and McAlearney emphasized, providers felt pressure to utilize the system and new technologies, even when they felt they should make their own decision and not rely solely on AI (Rathert et al. 2017; McAlearney et al. 2015). Although these systems provide more efficient documentation and have potential to continue to improve along with implementing new technologies, research shows that many physicians still struggle with the implementation of EHR systems and actually exhibit grief like symptoms like in the five stages of grief in the Kübler-Ross’s model; denial, anger, and frustration before finally reaching acceptance (McAlearney et al. 2015). Healthcare organizations have not realized how much of an impact this has had on providers and the way they see healthcare. They struggle to understand how physicians struggle with this high level of change and technology, and yet still are not provided with adequate training and follow ups after EHR/technology upgrades. Although EHR systems have provided many positives, the way they are rolled out and shown to end users has caused burnout, along with staffing shortages and pressure on providers to maintain efficiency continues to increase (Bannon 2023). Although the United States has been on EHR systems for nearly two decades, the top two issues remain the same, systems need to communicate with each other and healthcare organizations need to implement proper training and pay close attention to the most used workflow and improve it for the clinicians and clinical staff utilizing it daily. EHR systems are a great tool but the overall issues with upgrades and workflows fall on those who are required to utilize it heavily on a daily basis (Janett and Yeracaris 2020). At the end of the day, the EHR and technology is only as good as the support system behind it and at this point in time, that support system continues to let good providers down.
Data Visualizations
Visualization One
ggplot(data = myquery1.1,
mapping = aes(x=label, y=chart_count, fill = label))+
geom_col(show.legend = FALSE) +
scale_fill_manual(values = c("RN"="blue", "IMD" = "light blue", "RRT" = "light blue"))+
theme_minimal()+
labs(
title= "Chart Events Recorded by Caregiver Type",
subtitle = "Data taken from caregivers and chartevents table within MIMIC-III:
The graph shows caregiver events in thousands ",
x= "Caregiver Type",
y= "Number of Chart Events",
caption = "Source: MIMIC-III Clinical Database v1.4"
)
In this visualization, I have joined together the ‘caregivers’ table with the ‘chartevents’ table from the MIMIC-III dataset to count the number of chart events recorded by each type of caregiver using the ‘count’ and ‘group by’ function. The ‘having’ clause filtered out the results to show only the caregiver types that had more than 10,000 chart events. I made the dominate caregiver type a dark blue and the less dominate caregiver types a light blue/gray color. According to the data that was pulled and what is displayed on the visualization, RN’s record the majority of chart events when comparing it to the other caregiver types. When looking at this type of information, it is relevant to healthcare organizations because it shows the organization which type of caregiver is more documentation heavy than the other, which in this dataset is RN’s. When organizations consider EHR suggestions, they can retain information and suggestions from RN’s, seeing as they use the EHR system and have the most chart events. Knowing which caregiver type documents the most, can help healthcare organizations see which workflow needs improvements first, or which workflow could take priority over the other, especially with EHR upgrades.
Visualization Two
SELECT caregivers.label,
COUNT(*) AS output_count,
SUM(CAST(value AS INTEGER)) AS total_output,
MAX(CAST(value AS INTEGER)) AS biggest_output,
MIN(CAST(value AS INTEGER)) AS smallest_output,
AVG(CAST(value AS INTEGER)) AS avg_output
FROM outputevents
INNER JOIN caregivers
ON outputevents.cgid = caregivers.cgid
WHERE caregivers.label = "RN"
OR caregivers.label = "NS"
OR caregivers.label = "PCT"
OR caregivers.label = "MD"
OR caregivers.label = "RRT"
GROUP BY caregivers.label
HAVING output_count > 10
ORDER BY output_count ggplot(data = myquery2,
aes(y = label, x = output_count, fill = label))+
geom_col()+
theme_minimal()+
labs(
title = "Output Events Recorded by Caregiver Type",
subtitle = "Data taken from Caregivers and Outputevents tables within MIMIC-III",
x= "Number of Output Events",
y= "Caregiver Type",
caption= "Source: MIMIC-III Clinical Database v1.4"
)
In this visualization, I have joined together the ‘caregivers’ table with the ‘outputevents’ table from the MIMIC-III dataset to count the number of output events recorded by each type of caregiver. I then used the COUNT, CAST, SUM, MAX, MIN, and AVG function with the GROUP BY function so the data that is being pulled is grouped together by the caregiver label. I added the WHERE function to filter the results to include the clinical caregivers in this chart, such as, RN, NS, PCT, MD, and RRT. I then included the HAVING function that ensures that the graph displays caregiver types with more than 10 output events, which are being displayed. When looking at the data, we can see that RN’s record more output events than any other caregiver type that is displayed. According to the data, we can also see that MD’s recorded very few output events compared to RN’s, which reflect the day to day data and reality, that output monitoring is typically primarily recorded by registered nurses, rather than medical doctors. This data is relevant to healthcare organizations because understanding which caregiver type is primarily responsible for output events can help healthcare administrators understand who the workload typically falls on and who to rely on to understand the output events information. By asking someone who is primarily documenting the output events can help administrators when trying to improve EHR workflows for documentation.
Recommendations for Industry
Based on the research, articles, data and findings reviewed in this report, I believe there are several actionable steps that healthcare administrators should prioritize when it comes to EHR and AI implementation and upgrades. Both visualization one and two show that RN’s tend to carry a large portion of documentation and output events, or chart events, that happen within the EHR system. This data that was given from the MIMIC-III data set aligns with what research has been saying for years, the documentation portion of an EHR system mainly falls on the nurses portion of their job, this should be part of the conversation when making upgrades to the EHR system and workflows (Rathert et al. 2017). I believe that the industry needs to focus on four main things, repeat training, technology policies, interoperability, and an open floor/survey for staff that uses EHR workflows daily. First, repeat training, healthcare organizations cannot teach a system at the start of employment or EHR system roll outs and expect their staff to have a full understanding on all the ins and outs of a EHR system. Healthcare organizations need to invest in proper training and learning tools/resources for clinicians and clinical staff, or overall end users, who use the EHR system on a daily basis. This sound be a hands on training system that continues to update as the system does (McAlearney et al. 2015). Next, healthcare organizations should have clearly stated policies when it comes to how to utilize AI tools to make sure clinicians know when to act based on their own judgement rather than strictly rely on AI for their clinical information. Implementing these policies will not only legally protect the clinical staff and clinician but will also ensure providers understand they will not receive backlash from the organization for using their own judgement (Bannon 2023). Next, interoperability needs to increase to ensure providers have all the data, medical records and information they need about a patient to properly treat that patient rather than the patient experience a gap in care. (Janett and Yeracaris 2020; Rathert et al. 2017). EHR systems are almost two decades old with continued technological advancements including AI, before continuing to progress with AI advancements, making sure EHR systems integrate and talk to each other should be a top priority for the healthcare industry as a whole to ensure patients are receiving the top quality care organizations and providers know is out there. Lastly, healthcare organizations need to ensure their team members, especially those who utilize these technologies, feel comfortable speaking up when something is not working or something can be improved. These clinicians and employees who utilize the EHR systems on a daily basis know where the issues lie and where the improvements could be impactful to the organization (Rathert et al. 2017). My overall recommendation for the healthcare organization industry would be to truly look at the data and research, dive into the feedback the clinicians are giving us, because better outcomes start with an overall better support for those in the front line delivering the care.
References
Bannon, Lisa. 2023. “Wehn AI Overrules the Nurses Caring for You.” The Wall Street Journal. https://www.wsl.com.
Janett, Robert S., and Peter Pano Yeracaris. 2020. “Electronic Medical Records in the American Health System: Challenges and Lessons Learned.” Ciencia & Saude Coletiva 25 (4): 1293–1304. https://doi.org/https://doi.org/10.1590/1413-81232020254.28922019.
McAlearney, Ann Scheck, Jennifer L. Hefner, Cynthia J. Sieck, and Timothy R. Huerta. 2015. “The Journal Through Grief: Insight from a Qualitative Study of Electronic Health Record Implementation.” Health Services Research 50 (2): 462–88.
Meigs, Soloman, Stephen L. 2016. “Electronic Health Record Use a Bitter Pill for Many Physicians.” Advances in Health Information Science and Practice 13: 1–17. https://ahisp.kglmeridian.com/view/journals/phim/13/Winter/article-1d.xml?%3Cext_search%3E=pubMedWidget-542563&body=FullText.
Morris, Oliver. 2025. “When Did Electronic Health Records Become Mandatory?” AEANET. https://www.aeanet.org/when-did-electronic-health-records-become-mandatory/#The_HITECH_Act_A_Pivotal_Moment.
Rathert, Cheryl, Tracy Porter, Jessica Mittler, and Michelle Fleig-Palmer. 2017. “Seven Years After Meaningful Use: Physicians’ and Nurses’ Experiences with Electronic Health Records.” Health Care Management Review 44 (June). https://doi.org/10.1097/HMR.0000000000000168.