SCCM Resource Library-Examining the Association Between ICU Operational Conditions and Clinical Decision-Making

Video Transcription

Hello, everyone. My name is Jae Young Park. I'm a Ph.D. candidate from the Department of Industrial and Systems Engineering at the University of Florida. Today, I'm going to  talk about my research examining the association between ICU operational conditions and clinical  decision making using a systems engineering approach. This is joint work with my advisor, Dr. Shang Zhong, and Dr. Brian Pickering as a co-PI from the Mayo Clinic in Rochester,  Minnesota. This study is supported by the Agency for Healthcare Research and Quality. Intensive care units provide care for patients with life-threatening illness. They require  timely and appropriate interventions. To achieve this, multidisciplinary teams make decisions  together and share their information with each other frequently. For the decision-making,  the team continually monitors patients' health condition and collects patient-generated data and information. These are inputs of decision-making. Based on them, the team plans  interventions such as labs, medications, and so on. These are outputs of decision-making.  When it comes to operational conditions, the patient census tends to determine care team's  workload. A high patient census represents highly demanding work environment. As the number of patients increases, the amount of data patients generate and decision-making regarding  interventions increase. Additionally, depending on patients' compositions in ICUs, operational conditions can be more demanding. For example, at the high number of severely ill  patients, the operational condition is more demanding since the patients need more interventions and the care team has to handle much more amount of and complicated data.  Based on systems engineering initiative for patient safety, we found the relationships among operational conditions, inputs, and outputs that I mentioned. The operational  conditions are associated with patient data. Corresponding to the increase in the amount of the data and physical workload, cognitive workload for individual team members increases.  And team distributed cognition worsens. Subsequently, this is associated with the risk of interventions. In other words, a demanding IC operational condition will  increase the risk of cognitive overload among the care team.  We assume that the drop of the number of medication orders, that is, unstable medication ordering behavior indicates the team cognitive overload. Here, according to literature,  the number of medication orders represents a unit of productivity of the care team and a surrogate for distributed cognitive function. Therefore, we hypothesize that the  number of medication orders per patient drops significantly in a demanding operational condition, such as a high number of patients in ICUs. We examine the relationship between the  patient census and the number of medication orders. Furthermore, this study includes extremely demanding operational conditions, such as a high number of severely ill patients.  Let's move on to the methods. This study is a retrospective study using observational data.  We collected data from the 32-bed medical ICU at Mayo Clinic in Rochester, Minnesota from 2016  to 2018. We identified 4,800 unique patients and 6,200 MEQ admissions. We collected numbers of medication orders, patients, and severely ill patients every hour.  We aimed to examine the association between operational conditions and decision-making  outcomes. Also, we wanted to find the patient census where ordering behaviors became different.  Thus, we separated patient census into two groups, low and high patient census, and we conducted  univariate piecewise Poisson regressions. Then, we investigated the rates of the multiple pieces  and whether they were identical or not using generalized F-test. We identified a breakpoint  if the rates were significantly different and the number of medication orders dropped in the  high patient census. In addition to the rate per piece, we compared the number of medication orders per patient per piece in hourly basis to easily interpret the results.  These two figures show the piecewise Poisson regression regressing the number of medication  orders on the patient census. In the left figure, we can see that the rate in the low patient census  between 10 to 18 and that in the high patient census between 19 to 25 are different. In the  high patient census, the rate is much lower. We also found that the extremely high patient census  between 26 to 27 recovering the rate, but the rate was not significantly different because of  lack of data points. So, we considered the patient census between 19 to 27 as the high patient census.  Using the generalized F-test, we confirmed that the medication order rate was significantly reduced  when there were more than 18 patients. This patient census is called a breakpoint.  In the right figure, we compared low and high presence of severely ill patients. When there were many severely ill patients colored in orange here, the rate dropped at the  lower breakpoint of 16 and the reduced rate was long lasting up to 26.  This suggests that the care team more suffered from cognitive overload in more demanding operational conditions.  Here is the number of medication orders per patient per hour obtained per patient census group.  We can see that the number of medication orders per patient significantly dropped in the high patient census from 0.74 to 0.65 and much significantly dropped at the high number  of severely ill patients from 0.81 to 0.63. This confirms that demanding operational conditions are associated with the distributed cognition  based on our assumption regarding the strained distributed cognitive function.  In conclusion, we identified the association between demanding operational conditions and reduction in decision making. This study highlights quantitative approach  to investigate cognitive overloads and decision making. Additionally, based on our findings, we can propose an adaptive clinical resource allocation strategy such as backup tele-ICU  consultants. Alright, if you have any questions, please email me. I would be happy to answer your  questions. Here is the QR code for my email address and also our full paper is available here. The paper contains details I might miss in this presentation due to the time limit  and additional operational conditions I did not mention in this presentation.  Thank you for listening to my presentation.

Video Summary

Jae Young Park, a Ph.D. candidate from the University of Florida, discusses his research on the association between ICU operational conditions and clinical decision making using a systems engineering approach. The study found that demanding operational conditions, such as a high patient census and a high number of severely ill patients, can increase the risk of cognitive overload among the care team, leading to a reduction in decision-making. The study used observational data from a 32-bed medical ICU at the Mayo Clinic in Minnesota and identified a breakpoint in patient census where medication ordering behavior changed. The findings suggest the need for adaptive clinical resource allocation strategies, such as backup tele-ICU consultants.

Asset Subtitle

Professional Development and Education, Administration, 2022

Asset Caption

INTRODUCTION: The ICU work environment requires care teams to continually make appropriate, accurate, and timely decisions. This study aimed to identify the association between the operational conditions and the decision-making in a quantitative way. METHODS: This study included temporal data at one medical ICU (MICU) of Mayo Clinic in Rochester, MN, between February 2016 to March 2018. During this period, 4,822 unique patients had been admitted to the facility (a total of 6,240 MICU admissions). We developed a univariate piecewise Poisson regression model where the number of hourly medication orders was regressed on the hourly patient census. Here, the medication order and the patient census are considered as a surrogate of the decision-making and the operational condition, respectively. Then, a generalized F-test was conducted to examine whether the increase rates of multiple pieces were identical. Also, we compared the numbers of medication orders per patient per hour obtained per piece. In addition, conditioned on a high presence of severely ill patients, the regression model with the same independent and dependent variables was constructed to compare different operational conditions. RESULTS: We identified a breakpoint that the increase rate started to be reduced; the number of medication orders per patient per hour was significantly reduced over 56% occupancy (18 patients) of the ICU (average=0.74; standard deviation (SD)=0.56 vs. average=0.65; SD=0.48; p < 0.001). This indicates the amount of decision-making decreased by 13% when at a high presence of patients. That is, it was associated with the delay of decision-making. Furthermore, the delay deteriorated when at a high presence of severely ill patients; the breakpoint shifted to a lower patient census (16 patients), and the quantity decreased by 22% over the breakpoint (average=0.81; SD=0.59 vs. average=0.63; SD=0.47; p < 0.001). CONCLUSIONS: Our statistical models revealed that demanding operational conditions were associated with the delay of decision-making. We propose adaptive clinical resource allocation (e.g., teleICU consultants as backup consultants) to mitigate the delay. Also, this study highlights quantitative comparisons of decision-making in different operational conditions. This allows for effective clinical resource allocation.

Meta Tag

Content Type Presentation

Knowledge Area Professional Development and Education

Knowledge Area Administration

Knowledge Level Advanced

Membership Level Select

Tag Informatics

Tag Intensivist

Tag Innovation

Tag Staffing

Tag Healthcare Delivery

Year 2022

Keywords

Jae Young Park

ICU operational conditions

clinical decision making

systems engineering approach

cognitive overload

	Society of Critical Care Medicine 500 Midway Drive Mount Prospect, IL 60056 USA Phone: +1 847 827-6888 Fax: +1 847 439-7226 Email: support@sccm.org	Contact Us About SCCM Newsroom Advertising & Sponsorship DONATE	MySCCM LearnICU Patients & Families Surviving Sepsis Campaign Critical Care Societies Collaborative	GET OUR NEWSLETTER
© Society of Critical Care Medicine. All rights reserved. \| Privacy Statement \| Terms & Conditions The Society of Critical Care Medicine, SCCM, and Critical Care Congress are registered trademarks of the Society of Critical Care Medicine.