Exploring the Relationship Between Severity of Patient Illness and In Room Primary Care Nursing Time in a Medical Intensive Care Unit

Abstract

This study explores a connection between sequential organ failure assessment (SOFA) of patients in a medical intensive care unit (ICU) and primary care nursing time. Although ICU nurses have many responsibilities that impact patient outcomes, little is known about how the severity of a patient’s illness influences the time and activities that nurses dedicate to primary care within a medical ICU. Data collection was set for February to April 2020, as suspended on March 11 due to COVID-19. It resumed on July 6, 2020, after clinical research activities were reinstated. We collected data for a time-motion study. We combined the manual observation data with the real-time location data of a Near Field Electromagnetic Ranging (NFER) system to analyze how time is structured in primary care. Our goal was to investigate the impact of a patient’s illness severity on how time is allocated in intensive care nursing. Overall, there was no significant difference between February and July in terms of time spent on primary care and peer support activities. However, for specific primary-care tasks, ICU nurses spent less time on patient assessment, patient care, and medication tasks compared to February, significantly more time on patient transport in July. In the February data, we found a significant correlation between the SOFA score and the time devoted to primary care, indicating a meaningful association between patient illness severity and the allocation of nursing time. However, in the July data, no correlation was found between the SOFA score and the time spent on primary care. These findings improve our understanding of the challenges that nurses experience in medical ICU units, emphasizing the necessary adjustments to enhance nursing care.

INTRODUCTION

Patients in the intensive care unit (ICU) are especially vulnerable to a decline in their clinical condition. This heightened risk for deterioration is primarily due to the severe nature of their illnesses, which often involve complex health issues, critical organ dysfunction, and the need for continuous monitoring and advanced medical interventions (Gooch & Kahn, 2014). To prevent failure to detect clinical deterioration cues, measuring accurate health processes and understanding the impacts of the multiple factors related to the staff and ICU environments are crucial for improving the quality of healthcare service. Over the past decade, there have been notable advancements in nursing care within medical ICUs, aiming for a more patient-centered approach. Previous studies on time allocation have contributed to identifying efficient and cost-effective methods for delivering patient care (Blay et al., 2014). However, in recent times, a variety of new factors have emerged that could greatly affect the amount of time nurses spend on clinical activities and their overall workload (Kotfis et al., 2020). Several studies have explored the ICU nurses’ activity patterns and time allocation within ICUs (Fernández-Castillo et al., 2021; Pourteimour et al., 2021). Also, Previous studies have examined how the severity of illness affects nursing time allocation. Research conducted by Tafti et al. (2025) found that for every additional 10 minutes a nurse dedicates to patient care, there is a noteworthy 0.05% reduction in mortality rates. Additionally, Pan and Wu (2020) demonstrated that implementing time management strategies in nursing can significantly enhance the efficiency of emergency treatments provided to ICU patients suffering from severe heart failure. However, there is limited research regarding the scope to which the severity of a patient’s illness impacts the allocation of time and the specific activities that nurses engage in while providing primary care within a medical ICU.

To investigate the connection between patient severity of illness and nurse primary care time, we conducted a time-motion study two times (February and July 2020). The study was designed prior to COVID-19 in our region, with the original plan for continuous data collection over a two-month period. Its objectives do not specifically address questions related to the pandemic. Throughout the study period, the 18-bed medical ICU studied was formally designated a non-COVID unit. However, we recognize that the suspension of on-site research from mid-March to June 2020 might have introduced an environmental confounder. We explicitly consider this factor in our analysis and acknowledge it as a limitation of our study.

The present study addresses two research questions: “What are the differences in time distributions for ICU nurses between two different periods?” and “Suppose there is no significant difference in the amount of time spent on primary care between two distinct periods. Are there any notable variations in the correlations between the severity of patient illnesses and the time allocated to in-room primary care during these two time frames?” The first question is explored by quantifying the allocation of nurses’ time, specifically about primary care, peer support, out-of-room, and non-nursing activities. Further details regarding these tasks will be provided in a later section. In addition, a regression analysis was conducted to address the second research question, focusing on how the Sequential Organ Failure Assessment (SOFA) score relates to the total time spent on primary care within patient rooms.

By uncovering a significant correlation between time data generated during nursing primary care and the SOFA score, the findings of this study will offer valuable insights that can lead to meaningful process improvements within healthcare systems. Ultimately, these will enhance the quality of patient care, ensuring that resources are used effectively to meet patient needs.

METHOD

Sample and Recruitment

We completed a total of 75 observations (888.453 hours), with total 9 nurses in February and 11 nurses in July (Vargas et al., 2022). Data collection was originally scheduled to take place from February to April 2020 in order to capture routine practices. However, this process was suspended on March 11, 2020, due to the wave of COVID-19, which led to the cessation of all non-essential on-site research. Data collection resumed on July 6, 2020, after clinical research activities were reinstated. July 2020 was the earliest date that the institution allowed on-site data collection to restart under the same clinical conditions in non-COVID units. In February 2020, eight female nurses and one male nurse took part in the first time-motion study, and all the female nurses returned for a follow-up study (July 2020), and three additional female nurses joined for the second time-motion study. 38 observations (448.37 hours) were done in February, and 37 observations (440.08 hours) were done in July. Institutional review board (IRB) approval was obtained from the University of Missouri. All ICU nurses were registered nurses with experience working in multiple ICUs for at least one year. The mean years of experience for February were 9.31 years (SD = 3.15 years), while for July, it was 7.75 years (SD = 2.35 years). All patients were adults (≥ 18 years). No significant difference in experience was observed between the two time periods.

Time-Motion Data and SOFA Score Collection Procedures

The time-motion study has long been a common method to gain insights into nurses’ time allocation. This approach allowed us to gain deeper insights into the patterns and dynamics of time management within the healthcare environment (Abbey et al., 2012). It has been widely employed in healthcare systems for many years, enabling accurate assessment of task durations and providing a precise understanding of workflow sequences (Ballermann et al., 2011). Direct time-motion study observations from qualified workers have allowed us to create accurate stand times for their tasks and generate a precise picture of the sequence of steps in a workflow (Bratt et al., 1999; Larson et al., 2004). Hence, in this study, we thoroughly compared how ICU nurses allocated their time during two distinct periods: February 2020 and July 2020.

To investigate the relationship between the severity of patient illness and in room primary care nursing time in a medical ICU, a regression analysis was conducted to explore the impact of the association between the SOFA score and the total time spent on primary care within patient rooms. The SOFA score is a widely used tool for predicting ICU mortality and assessing dysfunction across six organ systems (respiration, coagulation, liver, cardiovascular, central nervous system, renal) based on clinical data (Arts et al., 2005). Previous literature (Altafin et al., 2014; Despins et al., 2019; Hefter et al., 2016) suggests that ICU nurses tend to spend more time on direct patient care when the patient’s SOFA score is higher, indicating a positive relationship between primary care time and the patient’s SOFA score. However, establishing a clear correlation between these two factors is quite complex. Each ICU nurse is responsible for multiple patients, each with varying SOFA scores, making it challenging to determine how much time a nurse spends with each patient during their shifts. To address this issue, we implemented an advanced indoor location tracking system. This technology allowed us to accurately record the time each nurse spent near every patient’s room, enabling us to directly and systematically compare care time across different patients. Every medical ICU room is designed for single occupancy. This means that the time nurses spend directly with each patient—referred to as NFER in-room dwell time—reflects the individual experience of just one person. Each nurse is typically responsible for attending two different rooms, maintaining a nurse-to-patient ratio of 1:2.

The time-motion study data was collected from the University Hospital, University of Missouri-Columbia. The medical ICU at University Hospital, University of Missouri Health Care, has eighteen beds composed of two pods with nine rooms each. The area equips a station in each pod with three cabinets for medications and multiple desktop computers. Traditionally, time-motion study data has been collected using mobile apps or a stopwatch and paper-and-pencil records. However, these conventional tools possess certain limitations. Firstly, observation accuracy may be compromised when the observation period is lengthy. Secondly, the time-motion study observer must be well-trained and have sufficient experience conducting such studies. Thirdly, emotional and sociological factors, such as the Hawthorne Effect, can potentially impact the performance of healthcare workers and the overall quality of the time-motion study. To overcome these limitations, the current study incorporates a Near Field Electromagnetic Ranging (NFER) sensor as an innovative work technology, aiming to supplement the shortcomings of traditional tools like mobile apps or the stopwatch and paper-and-pencil method.

Compared to Radio Frequency Identification (RFID) sensors, the NFER system offers superior accuracy in real-time location tracking, providing precise x and y coordinates of a nurse’s position with up to 40 cm accuracy, measured in feet —even capable of penetrating walls with precision (Jones et al.; Schantz et al., 2011). In contrast, traditional Wi-Fi methods often fall short, delivering an error margin of 8 to 15 meters (Cantón Paterna et al., 2017), while passive RFID systems can only provide basic portal-based detection, typically within a mere one meter of a reader (Ni et al., 2011; Wurms, 2021). The original study protocol further highlights the limitations of Wi-Fi tracking in hospital settings, which typically struggles within a range of just 10 to 20 feet (approximately 3 to 6 meters), proving inadequate for the stringent demands of efficient workflow (Kim, 2024; Okoniewska et al., 2012).This means that the NFER system can determine the exact location of a nurse within a monitored area. Near-field signals are perturbed only gradually and monotonically by the electrical characteristics of the environment (i.e., the structure of the building, attenuation through dielectric media, and conduction via wiring and plumbing). These characteristics are defined by the gross, large-scale structure of the building. However, small-scale variations like the opening and closing of a patient door or the presence of medical equipment or personnel do not affect NFER signals. By employing this innovative tracking method for healthcare staff, the study analyzed the workflow of a service process line and the physical movements of nurses, including factors such as distance traveled and speed, during 12-hour day shifts. The data obtained from the NFER system was combined with manual observation data, allowing for a comprehensive examination of nurses’ time allocation in primary patient care and comparing data collected in February 2020 and July 2020.

Data collection occurred over 15 days in February, from 2/17/2020 to 3/11/2020. Additionally, data was collected over 18 days in July from 7/6/2020 to 7/31/2020. During the data collection period, one or two ICU nurses were observed daily from 7:00 a.m. to 7:30 p.m. It is important to note that the data was collected in a non-COVID intensive care unit.

Materials

The observation data collection form was carefully developed from a previous study to prevent inaccuracy issues related to monitoring nurse activity for a long period (Song et al., 2017). A group of volunteer medical ICU nurses guided revisions regarding the primary care tasks and descriptions. They reviewed and approved the final version of the form. A senior medical ICU nurse worked with the research team as they piloted the final form, identifying care activities and explaining workflow processes. There are four main categories in this time-motion study observation form (i.e., primary care, peer support, out-of-room, and non-nursing). Table 1 shows the four primary categories, detailing the specific tasks encompassed within each category. In this study, peer support is described as: “A voluntary, non-evaluative and mutually beneficial collaboration between two health care professionals of similar experience” (Waddell & Dunn, 2005).

Table 1:

Activity Categories Recorded in the ICU Nurse Time-Motion Study Observation Form

Category	Included Activities
Primary care	All in-room, nurse-performed healthcare tasks: patient assessment; procedures; routine patient care (e.g., turning, hygiene); EMR charting; medication administration; patient transport; equipment setup or troubleshooting; collecting laboratory specimens; stocking supplies; room cleaning; participation in bedside or multidisciplinary rounds.
Peer support	Assisting others (nurse- or physician-led) with patient care and procedures, including preparing supplies, handing instruments, monitoring vitals, or documenting during a colleague’s intervention.
Out-of-room	Direct or indirect care carried out outside the patient’s room: printing ECG strips; teaching residents/students; hand hygiene; consulting or communicating with other healthcare personnel; retrieving supplies or medications; unit-wide tasks; transporting laboratory specimens, and similar activities.
Non-nursing	Managerial, social, or personal tasks not directly related to patient care (e.g., administrative paperwork, personal phone calls, meal or comfort breaks, informal conversations).

This study elaborates on the nurses’ time allocation related to primary care tasks. Table 2 shows the definition of primary care activities in a patient room.

Table 2:

Primary Care Task Definition in Time-Motion Study

Primary Care Task (In-Room)	Definition
Patient Assessment	Examining/evaluating a patient during the day shift
Performing Procedure	Performing/ assisting medical procedures
Patient Care	Assisting patients, including turning, bathing, comforting, and talking to a patient.
EMR Charting	Documenting electronic health record data about a patient
Medication	Preparing/administrating medications
Transport Patient	Preparing and transporting a patient, general cleaning and sanitation for a new patient
Working on Equipment	Watching monitors (inside the patient room) and checking/changing ventilator settings
Attending/Participating in Clinical Rounds	Clinical morning rounds and reports

Based on the results of our previous studies, there was no critical accuracy problem with the ability of the observers to monitor nurse activity for a long period of time. According to the earlier studies (Despins et al., 2019; Song et al., 2017), there was approximately 80% power to detect an assumed time average difference of 1.20 with 12 ICU nurses and a base correlation of 0.1 with an exchangeable correlation structure.

In addition, the research team implemented the NFER system to improve the accuracy of the time-motion study. The NFER system consists of IEEE 802 (O’hara & Petrick, 2005) and FCC 15.219 wireless devices: a position server, wireless router, transmitters, and receivers (NFER tags and locators) (Schantz et al., 2016). The NFER tag transmits a low-frequency signal. After the locator receives the NFER tag signal, the data is delivered to the wireless router and the position server. The algorithms built into the position server calculate a tracking solution and display position information on the server screen in real time. We used 12 locators (see Figure 1 (b)) and sixteen NFER tags to monitor the whole ICU area for this study. The weight of the NFER tag is 153g (see Figure 1 (c)). The tag battery can run up to 14 hours. Each locator-receiver localizes an NFER tag to coordinate tracking data about 165 – 350m². Before initiating data collection, the NFER system underwent a thorough calibration and validation process, which was conducted on-site. For this study, approximately 200 calibration points were selected, which included a total of 92 to 95 locations within hallways and nursing stations, along with a total of 105 to 118 designated single-patient rooms. An investigator equipped with an NFER tag remained still at each calibration reference point for 3 seconds. This data was crucial, as the raw X and Y coordinates obtained from this were subsequently compared against established ground truth metrics to ensure accuracy. At the beginning of each 12-hour shift, we took the time to synchronize and calibrate all tags with the NFER system and the time-motion study tool used by our human observers. The NFER system logs a location label roughly every minute. Observers then recorded the start and end times of tasks using a minute-based time unit. For spatial segmentation of NFER coordinates (see Figure 1(a)), the x-y coordinate was categorized into one of six predefined zones: Patient-room—bed, Patient-room—EMR computer, Patient-room—other areas, Hallway (the common corridor outside the rooms), Nursing station (the work island and chart-storage space), and Storage/clean-utility (which includes supply closets and central equipment areas). While location alone does not pinpoint the exact task being performed, it does help narrow down the likely activities (for instance, if the nurse is at the EMR computer, they are probably charting or entering orders; if they are at the bed, they could be engaged in direct patient care or conducting a physical assessment). For each nurse, we identified the zone into which each minute fell by using the NFER data. Additionally, we had access to the observers’ lists of task intervals, complete with start and end timestamps. For any specific task, we counted the number of minutes that corresponded to each NFER zone. After matching every minute to their respective “(zone, task)” pairs, we computed the totals: NFER zone total represented the sum of all minutes labeled by NFER as “bed,” “EMR computer,” etc., while the observer zone total encompassed the sum of all minutes of observer-coded tasks that aligned with the same zone. By following these steps, each minute during every shift is accompanied by both a validated location (through NFER) and a validated task label (via manual observation), thereby ensuring the integrity and accuracy of our time-motion study. We selected observers from our team of research assistants, each of whom had prior experience in healthcare environments or human-factors observation. Every observer underwent 20 hours of standardized training. Following this training, two observers independently recorded the time-motion data. As the NFER data was already categorized into six zones (bed; EMR computer; other areas within the patient room; hallway; nursing station; storage/clean-utility) on a per-minute basis, we compared the observers’ zone assignments for each minute of data. In total, we found an overall percent agreement of 92%.

Figure 1:

NFER Real-Time location system in a medical ICU

At the end of each daily shift for data collection, a senior nurse with more than 10 years of experience in critical care independently extracted the necessary physiological and laboratory parameters from the EMR. She then calculated the SOFA score using the original six-organ algorithm. All values were verified before being entered into the study database. The variable ‘minutes in a patient room per patient’ represents the total direct-care contact time that each individual patient received from their assigned nurse during that shift. When we examined associations with SOFA score, we linked each patient’s room-specific time to that patient’s own SOFA score. Each patient-specific total was then paired with that patient’s SOFA score, producing one data row per patient.

Statistical Analysis

All analyses were conducted using JMP^® 17 software. The nurses’ times for primary care, peer support, out-of-room activities, and non-nursing tasks were presented as mean ± SD for normally distributed data and as median for skewed distributions. The normality of each outcome was assessed using the Shapiro–Wilk test. For the primary care, peer support, and out-of-room categories, comparisons of means between February and July were performed using independent-samples t-tests, as these datasets satisfied the normality assumption. Conversely, differences in non-nursing activities were evaluated using the Wilcoxon rank-sum test due to violations of normality. Additionally, the relationship between the time spent in the patient room (in minutes per patient) and the SOFA score was examined using ordinary least-squares linear regression. All statistical tests were conducted two-sided, with a significance level set at alpha = 0.05.

RESULTS

Table 3 summarizes nurses’ time allocation on primary care, peer support, out-of-room, and non-nursing activities. On average, in a 12-hour observation, ICU nurses spent 35% of their time on primary care (Feb: 4.42 hours, July: 4.02 hours) and out-of-room activities (Feb: 4.37 hours, July: 3.87 hours). They also spent about 6 – 7% of their time supporting other nurses daily. For the non-nursing activities, ICU nurses spent about 2.16 hours (19.7%) in February and 3.28 hours (28.03%) in July.

Table 3:

Comparing time spent (mins) on the Main Categories

Activity	Level	Mean	SD	Median	F	P
Primary Care	Feb	265.0	88.8	262.0	1.55	0.217
	Jul	240.9	78.6	235.0
Peer Support	Feb	50.73	26.37	53.50	0.62	0.434
	Jul	45.44	31.28	43.50
Out-of-Room	Feb	262.3	51.59	263.0	7.09	0.01 *
	Jul	231.9	47.17	222.0
Non-Nursing	Feb	129.9	65.7	115.5	22.64	<.001 *
	Jul	196.7	55.21	186.0

^*p < 0.05 indicates significance

No notable time disparity was observed between February and July in primary care and peer support activities. However, in July, ICU nurses allocated considerably less time to out-of-room activities and dedicated more time to non-nursing tasks. Among the in-room primary care activities (see Table 4), ICU nurses spent much less time on patient assessment, patient care, and medication-related tasks compared to the data gathered in February. In contrast, they allocated significantly more time to transporting patients in July. No significant time difference was observed between February and July for other in-room primary care activities such as performing procedures, EMR charting, working on equipment, and clinical rounds.

Table 4:

Nurses’ time allocation on in-room primary care activities (mins per day shift)

Work Task	Level	Mean	StDev	F	P
Patient Assessment	Feb	23.03	19.24	7.35	0.008 *
	Jul	13.42	9.06
Performing Procedure	Feb	87.82	35.10	1.46	0.231
	Jul	100.95	56.80
Patient Care	Feb	54.26	31.94	16.16	<.001 *
	Jul	29.94	17.69
EMR Charting	Feb	22.87	19.63	0.34	0.560
	Jul	19.94	19.98
Medication	Feb	21.83	14.85	15.34	<.001 *
	Jul	10.33	8.34
Transport Patient	Feb	17.95	18.09	8.95	0.005 *
	Jul	52.2	47.1
Working on Equipment	Feb	4.923	3.498	0.09	0.762
	Jul	5.217	3.219
Attending/Participating in Clinical Rounds	Feb	12.50	9.24	3.47	0.068
	Jul	17.63	12.13

^*p < 0.05 indicates significance

For the out-of-room activities, Table 5 illustrates the top five activities among a total of fourteen tasks that collectively accounted for over 85% of the total time allocated outside the patient’s room. ICU nurses spent much less time on reviewing paper documents, getting/preparing medications, and EMR charting compared to the data gathered in February. In contrast, they allocated significantly more time to talking with physicians and other healthcare personnel in July. No significant difference in time usage for telephone calls was observed between February and July.

Table 5:

Nurses’ time allocation on Out-of-Room activities (mins per day shift)

Work Task	Level	Mean	StDev	F	P
Reviewing Paper Documents	Feb	13.18	3.11	5.63	0.021 *
	Jul	2.67	3.15
Using Telephones	Feb	15.68	10.89	1.23	0.271
	Jul	13.19	8.38
Talking with Physicians and Other Healthcare Personnel	Feb	67.65	6.13	9.21	0.003 *
	Jul	94.13	6.21
Getting/Preparing Medications	Feb	16.10	0.91	17.05	<.001 *
	Jul	10.73	0.92
EMR Charting	Feb	113.47	36.15	7.13	0.009 *
	Jul	90.27	39.07

^*p < 0.05 indicates significance

According to nurse’s real-time location data from the NFER system, each ICU nurse spent about 84 – 90 mins in a patient room per patient (Feb: Mean = 84.14 mins, SD = 44.56 mins; July: Mean = 89.85 mins, SD = 48.52 mins). With a nurse-to-patient ratio of 1:2 in a medical ICU, it can be inferred that nurses typically spend approximately 3 hours in patient rooms. Two sample t-test result shows that there was no significant in-room spending time difference between Feb and July.

Fitted line plots were generated to examine the potential significance of the relationship between the average time spent in a patient room in a medical ICU and the SOFA score (see Figure 2). The results indicate a statistically significant association between the time spent in the patient room and the SOFA score in February 2020. The coefficient for the SOFA score suggests that, on average, the time spent in the ICU patient room increases by 8.97 minutes for each one-unit increase in the SOFA score (95% Confidence Interval (CI): 5.92 – 12.03). The model accounted for approximately 39% of the variation in the time spent in a medical ICU room. According to the R² value, it reveals that the model provides an acceptable fit to the data. Also, the Pearson correlation coefficient for this model shows a moderate positive relationship between them (r = 0.625, CI = (0.434,0.763)). We may further improve the model fit by introducing additional terms. There may be a nonlinear relationship between those two variables from the scatter plot, especially when SOFA is high. Based on the fit result by exponential growth curve fit, we can achieve an even better R² value, which is 0.46.

Figure 2:

Correlations between SOFA on Time spent in patient-room (mins per patient)

In July 2020, no significant relationship was observed between the average time spent in the patient room and the SOFA score (95% CI: −2.84 to 6.83). The mean response values were relatively similar when comparing the total time spent between February and July, with 84.14 minutes in February and 89.85 minutes in July. However, there is a notable difference in the root mean square error values, which estimate the standard deviation of the error term. In February, the value is 35.09 minutes, whereas in July, it is 48.65 minutes. This suggests that the total times in July exhibit more variability than those in February. Regarding the comparison of SOFA scores between February and July (see Table 6), no significant difference was observed between the two periods.

Table 6:

SOFA Comparison

Level	Mean	StDev	F	P
Feb	5.597	3.411	0.30	0.584
Jul	5.921	3.951

^*p < 0.05 indicates significance

DISCUSSION

In this study, we conducted time studies to observe nursing activities during their working day shifts from 7:00 a.m. to 7:30 p.m. To enhance the accuracy of the collected data, manual observations were supplemented with data obtained from the advanced real-time indoor tracking sensor, known as the NFER data. This study identified variations in time allocation and regression models among nurses in February 2020 and July 2020.

Comparing the Work of Intensive Care Nurses between February and July

This study aimed to document and compare the primary care, peer support, out-of-room, and non-nursing tasks performed by ICU nurses between February and July. As presented in Table 3, the findings indicated no significant difference in primary care and peer support activities. Previous studies have reported that direct care activities constituted approximately 41% of the overall tasks performed by ICU nurses (Norrie, 1997; Pelletier & Duffield, 2003). Our time-motion study data consistently revealed that direct nursing care activities accounted for over 41.8% of the ICU nurses’ time, with 43.8% in February and 39.8% in July. Although there was a slight reduction in the time allocation for these activities in July 2020, the change was not significant enough to be considered a major adjustment.

On the other hand, significant differences in time allocation were observed in out-of-room and non-nursing activities between February and July. The percentage of time spent on out-of-room activities decreased by 4.2%, whereas the percentage dedicated to non-nursing tasks increased by 9.3%. The disparities observed may be attributed to the pandemic. During July 2020, medical ICU nurses were required to adhere to newly implemented protocols resulting from the pandemic. These modifications are likely to have significantly redirected nurses’ focus towards following new infection-control protocols and addressing evolving patient needs. These changes in workload dynamics could significantly impact staff well-being and patient care delivery in critical settings. Existing literature (Harrison & Nixon, 2002; Wong et al., 2003) emphasizes the importance of understanding workload factors in the ICU, including indirect nursing care, even though it is not directly linked to the quality of patient care. The significant time increment in non-nursing activities serves as supporting evidence for these observations. The non-nursing activities are mainly associated with resting, lunch & personal breaks, and private conversations with staff. In the literature (Harrison & Nixon, 2002; Marasovic et al., 1997; Norrie, 1997), the percentage of the non-nursing is from 10% to 20%, similar to the current study non-nursing activities in February (observed around 18%). However, in July, there was a significant increase in the proportion of time spent on non-nursing activities (27.3%), surpassing the findings reported in other studies. If ICU nurses are allocating more time to personal breaks and resting, it could suggest greater physical and cognitive strain associated with personal protective equipment (PPE) use and adherence to new infection-control protocols. This interpretation is further supported by the significant increase in the time required for patient transport (see Table 4), which reflects more complicated procedures for moving patients under pandemic conditions. Additionally, Table 5 shows that in July, there was a significant decrease in the time ICU nurses spent reviewing paper documents. Conversely, there was an increase in their interactions with physicians and other healthcare personnel. This trend may indicate the growing demands placed on nurses, particularly concerning safety-oriented logistics. The medical ICU unit we observed did not have any patients who were positive for COVID-19 at any time; therefore, the changes in workflow that we observed are likely attributable to the hospital-wide measures implemented to ensure safety during the pandemic. The change in practices was part of a larger effort to protect the health of patients and healthcare workers in the hospital. Hence, it becomes essential to investigate how public health emergencies, like the pandemic, impact nurses’ workload in ICUs. Understanding the specific challenges nurses face during such crises is vital for improving patient care and supporting healthcare professionals in managing their responsibilities effectively. While several validated tools exist for monitoring workload, such as the Nine Equivalents of Nursing Manpower Score (Reis Miranda & Moreno, 1997) and Nursing Activities Score (Miranda et al., 2003), they do not comprehensively address the levels of nursing complexity management in ICU patients (Rengeiné et al., 2020). None of them consider the impacts of physiological or psychological nursing workload associated with public health emergencies.

Nurses’ time allocation in the ICU patient rooms

According to the study done by Schenk et al. (2017), ICU nurses allocate approximately 34% of their time within patients’ rooms, which aligns with the findings of our study. In February, nurses spent around 36.8% of their time engaging in in-room activities. In July, this percentage decreased to approximately 33.5% of their time spent in patient rooms. Table 4 revealed that the July time-motion study data significantly influenced the distribution of nurses’ time for various in-room primary care tasks. In February, ICU nurses dedicated most of their time to performing procedures, providing patient care, conducting patient assessments, EMR charting, and medication administration. These activities accounted for 33.1%, 20.5%, 8.7%, 8.6%, and 8.2% of their total work time within patient rooms, respectively. However, our study indicated a notable decrease in the time allocated to patient care, patient assessment, and medication activities in the July time-motion study data, with reductions of 12.4%, 5.6%, and 4.3%, respectively. Consequently, the time spent on these activities was nearly halved compared to the February time-motion study data.

In contrast, there was a significant increase in the time dedicated to transporting patients (Feb: 6.8%, July: 21.7%), nearly three times higher than before. This change can be attributed to the implementation of the hospital-wide pandemic safety measures. In July 2020, ensuring the safety of both patients and healthcare professionals became the utmost priority in medical ICUs. As a result, nurses were required to wear gowns, gloves, and masks while transporting patients to different locations. Furthermore, thorough cleaning and sanitization of high-touch areas within patient rooms became mandatory before admitting new patients. This involved the use of cleaning supplies and equipment to disinfect the area effectively. These rigorous physical tasks significantly increased the workload of ICU nurses. The findings of our study align with the existing literature (Burch, 2020; Qian et al., 2020), indicating that nursing professionals experienced heightened physical stress in their efforts to reduce the risk of infection during the pandemic.

The findings of this study shed light on the effects of the pandemic on nurse workload and changes in time allocation within a non-COVID ICU. These observations highlight the need for further investigation to comprehend the nature of the work burden faced by nursing professionals and develop strategies to effectively manage these increased workloads to meet the demands imposed by future public health emergencies successfully.

Relations between SOFA and Primary Care Time in a Medical ICU

Previous research in nursing time studies has predominantly focused on analyzing the time structure and nurses’ workflow in primary care. These studies aimed to determine the most and least time-consuming nursing activities in various settings. However, challenges arose when attempting to establish a meaningful relationship between the severity of illness in patients and nursing workflow within a medical ICU. This was primarily due to the limited human capacity to accurately record time-motion data in high-stress environments (Vankipuram et al., 2011). To overcome this limitation and establish a quantitative relationship, a regression analysis was conducted between the SOFA score and the total time spent on primary care in a patient room. Figure 2 presents the results of the linear regression models, indicating that a one-point increase in the SOFA score was associated with an 8.97-minute increase in primary care time in February 2020. However, in the July time-motion study data, the association between the SOFA score and primary care time was no longer statistically significant despite no significant difference in the SOFA score between February and July (refer to Table 6). This suggests that the pandemic may have altered the workflow patterns of nurses, resulting in more time being allocated to sanitization tasks and less time being dedicated to assisting patients in the ICU. Based on the findings presented in Table 4, it was observed that the average time spent performing or assisting medical procedures increased from 33.1% to 41.9% in July 2020. However, no statistical time difference was observed between February and July due to a significant increase in the response variance (SD Feb: 35.1 mins, SD July: 56.8 mins). This finding could potentially serve as important evidence supporting the limitation of the mandated staffing ratio policy. When there is substantial variability in direct care activities, it becomes necessary to reconsider the effectiveness of the current staffing ratio and include other variables beyond direct care activities in the staffing equation (Upenieks et al., 2007). The scatter plots in Figure 2 further illustrated that the variabilities in nurses’ primary care time with patients having a middle SOFA score (ranging from 3 to 7) were significantly increased when comparing the time in July to that of February.

The care requirements of ICU patients can vary significantly, leading to fluctuations in nurses’ primary care time. An increase in the acuity of illness among ICU patients is expected to correspond to more time spent on direct patient care (Hefter et al., 2016). However, our study results revealed no correlation between the SOFA score and primary care time in the July time-motion study data, primarily due to the impact of the hospital-wide pandemic safety measures. Instead, nurses experienced a fast-paced working environment characterized by multitasking, including sanitization and assisting other healthcare professionals. These identifiable risks highlight the potential negative impact on healthcare quality and the increased likelihood of medication errors (Cornell et al., 2010).

Recommendations and Limitations

Based on our findings, we recommend integrating real-time location systems to monitor how nurses allocate their time across various tasks continuously. Administrators should review this data weekly to identify any unintended shifts away from essential primary care activities. First, it’s important to develop or revise ICU protocols so that whenever patient acuity—measured by the SOFA score—rises, staffing assignments and task checklists clearly emphasize the need for direct patient assessment and medication management. Our findings from July indicate that nurses often found themselves diverted from direct care activities toward ancillary tasks such as transport. This shift stemmed from staff redeployments or the need for infection-control measures. To address this issue, administrators should consider implementing dynamic staffing models that allow for temporary reinforcements when overall patient acuity increases, instead of making blanket reassignments to non-critical tasks. By transforming our observations into these practical recommendations, we believe healthcare providers and administrators can more effectively align nursing resources with patient severity.

Further efforts are required to analyze the variabilities in direct and indirect care activities and evaluate the appropriate nurse-to-patient ratio in medical ICUs. Future studies should focus on reallocating nursing activities based on the workload experienced by nursing professionals. Additionally, it is important to note that the current study was limited to a single medical ICU within one hospital. Therefore, collecting data from multiple ICUs is necessary to determine if the findings of this study are consistent across different settings. While the NFER real-time location system has advanced the field of time-motion study in nursing, there is room for improvement. Specifically, the NFER sensor in the system should undergo further refinement to reduce its size and weight, ultimately enhancing nurses’ satisfaction with the technology. NFER-based location-tracking systems demonstrated an impressive mean positional error, indicating their strong potential for enhancing workflow analytics within intricate clinical environments. However, researchers and practitioners should consider several important practical limitations of this technology. Firstly, the tracking tags utilized in this system necessitate regular battery recharging, typically required daily, to ensure optimal functionality and continuous operation. This regular maintenance can be a logistical challenge in high-paced clinical settings. Secondly, the presence of metal-rich environments can significantly compromise the accuracy of location tracking. In such cases, it becomes essential to employ multipath-mitigation filters to counteract the negative effects of reflected signals and thereby improve the precision of the location data. Understanding these constraints is crucial for effective implementation and maximizing the benefits of this advanced tracking solution in clinical workflows. Another limitation of the study is that we did not include certain variables, such as the introduction of new house staff, variations in the staffing mix, and other relevant factors that may have influenced the outcomes. Lastly, nursing students during their clinical rotation, in July, were not documented. The arrival of nursing students in July may have altered the distribution of tasks, shifting the focus toward supervisory and teaching activities. This change has the potential to increase the number of indirect care minutes, which could weaken the connection between the time nurses spend providing direct care in patients’ rooms and the severity of those patients’ conditions. For future research, it would be beneficial to consider these limitations and adopt more comprehensive research designs to gain a deeper understanding of how these factors influence the observed outcomes.

CONCLUSION

ICU nurses operate under time constraints while carrying out various clinical activities, and their efforts and workload directly impact the quality of patient care. However, high-quality nursing studies are scarce that comprehensively capture the intricate nature of nursing activities. This study examines the variations in time allocation for ICU nurses between February 2020 and July 2020. It also investigates the relationship between the amount of time nurses spend in patient rooms of a medical ICU and the severity of the patients’ conditions. The results highlight the time adjustments made in response to the evolving demands of ICU patients between two different periods. Several primary care activities experienced a substantial reduction, by approximately 50% in July compared with February. Conversely, the time required for patient transport increased considerably. Also, we found that the association between the SOFA scores and the average time spent in a patient room in a medical ICU was no longer observed under pandemic safety measures. These findings contribute to a deeper understanding of the complexities and challenges nurses face caused by public health emergencies, offering insights into the necessary adjustments and interventions required to optimize nursing care in such circumstances.

HIGHLIGHTS.

• Under pandemic safety measures, SOFA scores no longer correlated with time in patient-room.

• Several primary care activity durations dropped ≈ 50% in July vs. February (p < 0.01).

• The pandemic protocols increased the time spent on transport patient activities.