Sage Journals: Discover world-class research

Abstract

Background: Effective nurse scheduling promotes patient safety, nurse well-being, and compliance with regulations. Objective: To develop and evaluate a robotic process automation (RPA)-based scheduling system to improve accuracy, reflect nurse preferences, and align with national guidelines. Methods: A total of 102 nurses from a 500-bed hospital were assigned to the experimental or control group. The RPA system was integrated into nurse managers’ workflows to automate monthly shift planning. Nurses submitted shift preferences and the system generated schedules accordingly. Pre- and post-intervention data on work characteristics, health status, and work-life balance were analyzed using chi-square and paired t-tests with SPSS. Results: The RPA-based scheduling significantly improved nurses’ work-life balance. No significant differences were found in health status or work characteristics. Conclusion: Integrating RPA into nurse scheduling can enhance work-life balance and support fairer scheduling practices. Broader organizational adoption and supportive policies are recommended to ensure sustainable impact and improved care quality.

Keywords

health status nurse robotics scheduling work-life balance

Introduction

Nurses play a critical role in hospitals by continuously monitoring patients and responding promptly to emergencies, with most working within a three-shift rotation system,¹ often facing understaffing, excessive workloads, and irregular schedules.² Irregular schedules are linked to fatigue, obesity, sleep disturbances, infertility, and miscarriage,^3,4 and may disrupt work-life balance and nurse-patient interactions, potentially compromising patient safety.^2,5,6 Well-structured schedules, in contrast, improve nurses’ health, job satisfaction, and work-life balance,^7–9 enhancing both care quality and nurses’ quality of life.¹⁰ The American Nurses Association emphasizes the importance of equitable workloads and healthy scheduling to ensure safe, high-quality patient care.¹¹

Nursing scheduling involves planning and allocating work hours to maintain safe staffing and optimal care delivery.¹² Nurse managers must balance staffing levels, competencies, fairness, and organizational needs.^13,14 However, perception gaps often exist between schedulers and staff, as personality, family status, and individual preferences make it challenging to satisfy all nurses. Misuse of scheduling authority can lead to perceptions of unfairness, dissatisfaction, and lower morale.¹⁵ Furthermore, inappropriate scheduling has been associated with anxiety, stress, burnout, and sleep disorders, while fatigue and drowsiness reduce concentration, worsen work-life balance, and negatively affect patient safety.^2,6,16 Thus, hospitals should regularly monitor scheduling practices and intervene proactively when problems arise.

Efforts to develop equitable and efficient scheduling systems have been reported domestically and internationally. In South Korea, Jeon et al. proposed a scheduling model incorporating various constraints,¹⁷ while Song developed an automated scheduling system using a web-based interface to reflect nurse preferences.¹⁸ Internationally, preference-based scheduling systems in Egypt and India have demonstrated economic and operational benefits by reducing overtime and shortening scheduling time.^19,20 Nonetheless, most automated systems still rely on rule-based or preference-matching approaches, requiring substantial manual adjustments and limiting scalability and workflow integration.²¹ Many commercial systems operate as standalone applications, making real-time linkage with hospital information systems challenging and thus limiting dynamic schedule updates.²²

Artificial intelligence (AI) and robotic process automation (RPA) remain underutilized in nurse scheduling. RPA automates repetitive tasks, enhancing fairness and accuracy,²³ reducing manager workload, and improving satisfaction.²⁴ Moreover, integrating RPA directly into hospital information systems can substantially reduce administrative task time and errors.²² Compared to traditional automated or preference-matching approaches, RPA provides clear advantages, including direct data extraction, automated staffing-guideline checks, schedule generation without extensive system redevelopment, reduced administrative burden, and easier integration with existing hospital workflows.²²

In South Korea, the Ministry of Health and Welfare has implemented policies to improve shift work conditions, including the “Night Nurse Pay” program and the “Guidelines for Nurses’ Night Work.”²⁵ The Nurse Shift System Improvement Pilot Project, initiated in 2022, requires hospitals to implement predictable shift schedules and meet specific staffing requirements for night-duty nurses, with financial support provided based on compliance.²⁶ These policies highlight the importance of adhering to scheduling criteria and the need for research on effective implementation strategies.

Given the complexity of nurse scheduling and ongoing efforts to improve working conditions, this study aims to develop and implement an RPA-based nurse staffing program directly integrated into routine workflows. Building on previous systems’ limitations, this study evaluates the program’s impact on nurses’ work characteristics, health status, and work-life balance, providing foundational evidence to inform future workforce strategies.

Research hypotheses

• Hypothesis 1: There is no significant difference in work characteristics between the experimental group, where the nurse staffing program using RPA is implemented, and the control group.

• Hypothesis 2: The experimental group working in a unit where the nurse staffing program using RPA is implemented shows better health status compared to the control group.

• Hypothesis 3: The experimental group working in a unit where the nurse staffing program using RPA is implemented shows better work-life balance compared to the control group.

Methods

Design

This interventional study employed a non-equivalent control group pretest-posttest design to develop a nurse staffing program using RPA, implement it among nurses, and evaluate its effects on work characteristics, health status, and work-life balance.

Study sample

The study was conducted at a general hospital, participating in the Integrated Nursing Care Service Wards and the Nurse Shift System Improvement Pilot Project. Two general wards, two Integrated Nursing Care Service Wards, and two Intensive Care Units (ICUs) were randomly selected from six general wards, six Integrated Nursing Care Service Wards, and four ICUs. Nurses from these wards were recruited via convenience sampling and assigned to experimental or control groups. Data were collected from November 15, 2023, to June 30, 2024. The sample size was calculated using G*Power 3.1, based on Cohen’s formula for paired t-tests with a medium effect size of 0.5, significance level α = 0.05, and power of 0.95. The minimum required sample size was 54 per group, totaling 108. To meet this, 134 participants were recruited.

Inclusion criteria were nurses with at least 6 months of work experience in general wards, Integrated Nursing Care Service Wards, or ICUs, who were directly involved in patient care. Exclusion criteria were less than 6 months of work experience, assignment to non-clinical roles, extended leave, departmental transfer or resignation during the study period, or incomplete/invalid responses. Consequently, 102 nurses remained and were evenly allocated to the experimental (n = 51) and control (n = 51) groups. Cooperation with the study was facilitated by co-researchers affiliated with the hospital.

Ethical considerations

This study was approved by the Institutional Review Board of the affiliated hospital on August 9, 2023 (IRB No.: MJH 2023-06-048). Participants provided written informed consent after being briefed about the study purpose, confidentiality, voluntary participation, and right to withdraw.

Instruments

A structured questionnaire comprising 72 items was used in this study: 7 on general characteristics, 6 on work characteristics, 30 on health status, and 29 on work-life balance. The tools were used with original authors’ permission.

Work characteristics

Survey questions were developed based on monitoring indicators from the Nurse Shift System Improvement Guidelines (April 2022) and the Guidelines for Nurses’ Night Work (May 2022).^25,26 The questionnaire included items on the maximum number of consecutive workdays, total monthly shifts, leave provision after excessive night shifts, number of night shifts per month, rest after night shifts, and consecutive night shifts per cycle. Specifically, it assessed whether consecutive workdays were limited to five or fewer, whether 1 day of leave was granted when designated night shifts exceeded seven, whether at least 48 h of rest was ensured after two or more consecutive night shifts, and whether the number of consecutive night shifts was limited to three or fewer.

Health status

Health status was measured using the 30-item Health Status Scale, developed by Kim,²⁷ based on the health questionnaire from the University of Tokyo’s Department of Public Health, and later adapted into the Korean version of the Todai Health Index by Lim.⁹ The scale includes 10 for the physical domain, 7 for the psychological domain, 3 for the spiritual domain, and 10 for the social domain. Each item was rated on a 5-point Likert scale ranging from 1 (“strongly agree”) to 5 (“strongly disagree”). All items were negatively worded and reverse-coded, with higher scores indicating a better health status. Cronbach’s alpha was .87 in Lim’s study⁹ and .91 in the present study.

Work-life balance

Work-life balance was measured using the Work-Life Balance Scale developed by Kim and Park.²⁸ This scale comprises 29 items: 8 on harmony between work and family, 8 on harmony between work and leisure, 9 on harmony between work and personal growth, and 4 on general evaluation of life. Each item was rated on a 5-point Likert scale ranging from 1 for (“strongly agree”) to 5 (“strongly disagree”). As all items were negatively worded, reverse coding was applied during statistical analysis, with higher scores indicating higher levels of work-life balance. Cronbach’s alpha was .93 in Kim and Park’s study²⁸ and .95 in the present study.

Research procedure

Development of the nurse work scheduling program using RPA

Assessment of the current nurse scheduling process

Although the hospital used an information system, each nursing unit had distinct, non-standardized scheduling rules. Nursing department staff, co-researchers, and the system developer held meetings to identify challenges. Nurse managers entered into the system manually created rosters based on preferences and constraints. This delayed distribution and created undesirable sequences (e.g., E→D), conflicting with national guidelines. Privacy concerns arose as requests were visible to colleagues. Junior nurses felt pressured when preferences conflicted with senior staff. The developer conducted a 2-week observational study in three units to identify workflow patterns and constraints for automation.

Development of the nurse staffing program utilizing RPA

An RPA bot was developed to automate monthly schedules based on nurse manager preferences and ward rules. The bot generated schedules, submitted them for review, and registered final versions in the system. Nurses’ perspectives were incorporated during development through unit meetings and informal consultations, during which nurses provided feedback on fairness concerns, workflow issues, and preferred scheduling rules. Their input was used to refine constraint settings and request procedures, ensuring that end-user needs were reflected in both system design and the interpretation of study outcomes.

(1) Setting Document – Users selected functions (automatic generation or preference-based), periods (1–3 months), and registration options (upload or Excel file). Documents were Excel-based and linked to system.

(2) Google Sheet for Nurses – Nurses submitted requests via Google Sheets shared through KakaoTalk. After submission, editing privileges were restricted to view-only.

(3) Google Sheet for Nurse Managers

① Nurse Information – HR data (e.g., groups, employment dates, shift types) were synced with the nurses’ sheet. Staff were grouped by tenure and shift type.

② Duty Setting – Shifts included Day (07:00–15:30), Evening (14:00–22:30), Night (22:00–07:30) and fixed D2 (08:30–17:30). Monthly schedules reflected a 40-h workweek. Public holidays were recorded on a yearly sheet.

③ Constraints – Based on national guidelines:

Staffing Numbers - Minimum staffing requirements were set for each group and shift. Day shifts included D and D2.

Consecutive Shifts - Maximum 5 consecutive workdays, 3 consecutive night shifts, and rest days after 2 night shifts. Shift transitions like E→D or N→E were prohibited.

Shift Limits - Monthly caps for D/E/N matched individual preferences. Blank fields triggered default logic.

④ Requests – Nurses entered preferences and days off. Nurse managers added vacations and events, then reset sheets monthly. When nurse preferences exceeded staffing limits or violated constraints, the RPA bot used a simple conflict-resolution process. Preferences were prioritized based on staff group and submission order; if conflicts remained, the bot reassigned shifts using default rules that complied with national guidelines. Nurse managers could make targeted manual adjustments through an override function before finalizing the schedule.

Pre-test

Before implementation, both groups completed surveys—collected within three days—on demographics, work, health, and work–life balance.

Application of the RPA program

From January 1 to May 30, 2024, the RPA system was applied to the intervention units on a monthly basis.

• Nurse managers updated staffing data by the 30th.

• Nurses submitted input by the 5th.

• Reviews were completed by the 9th.

• The Bot generated and registered schedules by the 10th.

• Managers finalized and published schedules by the 11th (Figure 1).

Figure 1.

Application of nurse work scheduling program using RPA. Note. NRS, nurse rostering system.

Post-test

From June 10 to 26, 2024, post-tests were conducted using the same procedure. Data were collected within 5 days.

Data analysis

Data were analyzed using SPSS WIN 22.0 program.

Descriptive statistics (frequency, percentage, mean, and standard deviation) analyzed general characteristics and research variables. Group homogeneity was assessed using independent t-test and Chi-squared tests.

Program effectiveness was evaluated using Chi-squared or paired t-tests following normality tests.

Results

General characteristics

Most participants were female—48 (88.2%) in the experimental group and 47 (95.2%) in the control group—with no statistically significant difference between groups (p >. 05). Participants’ ages ranged from 23 to 52 years, with a mean age of 28.37 ± 5.17 years in the experimental group and 27.37 ± 4.09 years in the control group, showing no significant difference (t = 1.08, p = .281). No significant differences were observed in marital status, religion, education level, or work unit. Total clinical experience ranged from 6 to 327 months, with a mean of 5.17 ± 5.05 years in the experimental group and 4.57 ± 3.79 years in the control group (t = 0.68, p = .281), indicating homogeneity between the groups (Table 1).

Table 1.

Homogeneity test on general characteristics of participants (N = 102).

Variables	Categories	Exp. (n = 51)	Cont. (n = 51)	x² or t	p
Variables	Categories	n (%) or M ± SD	n (%) or M ± SD	x² or t	p
Sex	Female	45 (88.2)	47 (95.2)	0.44	.505
	Male	6 (11.8)	4 (7.8)
Age (years)	<30	36 (70.6)	36 (70.6)	0.00	1.000
	≥30	15 (29.4)	15 (29.4)
	M ± SD	28.37 ± 5.17	27.37 ± 4.09	1.08	.281
Marital status	Single	45 (88.2)	43 (84.3)	0.33	.565
	Married	6 (11.8)	8 (15.7)
Religion	Yes	17 (33.3)	12 (23.5)	1.21	.272
	No	34 (66.7)	39 (76.5)
Educational level	Diploma	7 (13.7)	10 (19.6)	0.64	.727
	Bachelor or higher	44 (86.3)	41 (83.4)
Work unit	General ward	11 (21.6)	13 (25.5)	0.43	.805
	Integrated nursing care service ward	20 (39.2)	17 (33.3)
	Intensive care unit	20 (39.2)	21 (41.2)
Total clinical experience (years)	<1	3 (5.9)	1 (2.0)	0.04	.841
	≥1 to <3	17 (33.3)	22 (43.1)
	≥3 to <5	10 (19.6)	6 (11.8)
	≥5	21 (41.2)	22 (43.1)
	M ± SD	5.17 ± 5.05	4.57 ± 3.79	0.68	.281

Con. = control group; Exp. = experimental group.

Homogeneity test for dependent variables between experimental and control groups before intervention

In the homogeneity test of work characteristics between the experimental and control groups prior to implementing the nurse staffing program using RPA, most nurses worked a maximum of five or fewer consecutive days (experimental group: 48 nurses, 94.1%; control group: 49 nurses, 96.1%). The mean number of maximum consecutive workdays was 4.63 ± 1.00 days in the experimental group and 4.61 ± 0.60 days in the control group (t = 0.12, p = .905). The mean number of monthly workdays was 19.06 ± 2.19 for the experimental group and 19.43 ± 2.44 for the control group (t = −0.81, p = .419). There were no statistically significant differences between groups in the number of days worked per shift type (Day, Evening, Night shift), the provision of 1 day off after more than 7 days, or ensuring 48 h rest after two or more consecutive night shifts. The mean number of night shifts per month was 5.67 ± 1.97 in the experimental group and 5.53 ± 1.78 in the control group, with no statistically significant difference (t = 0.37, p = .713). Most nurses had three or fewer consecutive night shifts (experimental group: 49 nurses, 96.1%; control group: 51 nurses, 100%), with mean values of 2.67 ± 0.68 days in the experimental group and 2.82 ± 0.48 days in the control group, showing no statistically significant difference (t = −1.34, p = .182). No statistically significant differences were observed between the groups in nurses’ health status (t = −1.23, p = .220) or work–life balance (t = −1.05, p = .297), confirming homogeneity between the two groups (Table 2).

Table 2.

Homogeneity test on main variables between groups (N = 102).

Variables	Categories	Exp. (n = 51)	Cont. (n = 51)	x² or t	p
Variables	Categories	n (%) or M ± SD	n (%) or M ± SD	x² or t	p
Maximum consecutive work days	≤5	48 (94.1)	49 (96.1)	0.21	.647
	6 to more	3 (5.9)	2 (3.9)
	M ± SD	4.63 ± 1.00	4.61 ± 0.60	0.12	.905
Numbers of shifts per month	Day shift	6.14 ± 3.71	5.92 ± 3.16	0.32	.753
	Evening shift	6.02 ± 3.02	6.82 ± 3.39	−1.26	.209
	Night shift	5.67 ± 1.97	5.47 ± 1.71	0.54	.592
	Working days	19.06 ± 2.19	19.43 ± 2.44	−0.81	.419
1 day off after 7 night shifts	Yes	41 (80.4)	41 (80.4)	1.01	.603
1 day off after 7 night shifts	No	10 (19.6)	10 (19.6)
Between-shift breaks shorter than 48 h after 2 consecutive night shifts	Yes	44 (86.3)	40 (78.4)	1.72	.423
	No	7 (13.7)	10 (19.6)
Numbers of night shifts per month	≤2	2 (3.9)	4 (7.8)	1.47	.690
	3∼4	9 (17.6)	6 (11.8)
	5∼6	30 (58.8)	29 (56.9)
	7 or more	10 (19.6)	12 (23.5)
	M ± SD	5.67 ± 1.97	5.53 ± 1.78	0.37	.713
Number of consecutive night shifts per cycle	≤3	49 (96.1)	51 (100)	2.04	.153
	4 or more	2 (3.9)	0 (0)
	M ± SD	2.67 ± 0.68	2.82 ± 0.48	−1.34	.182
Health status	Total	3.49 ± 0.57	3.62 ± 0.49	−1.23	.220
Work-life balance	Total	2.98 ± 0.73	3.12 ± 0.66	−1.05	.297

Con. = control group; Exp. = experimental group.

Effect of the nurse work program applying RPA for participants

Hypothesis 1 test

Hypothesis 1, which predicted no significant difference in work characteristics between the experimental group (where the nurse staffing program using RPA was implemented) and the control group, was supported. The experimental group’s maximum consecutive workdays decreased from 4.63 ± 1.00 to 4.31 ± 0.73 (t = −1.81, p = .077), while the control group showed no change (4.61 ± 0.60 to 4.61 ± 0.57; t = 0.00, p = 1.000). Between-group change was also not significant (−0.31 ± 1.24 vs. 0.00 ± 0.80; t = −1.52, p = .132). Monthly workdays slightly decreased in the experimental group (19.06 ± 2.19 to 18.47 ± 1.84; t = −1.51, p = .138) and remained similar in the control group (19.43 ± 2.44 to 19.10 ± 2.30; t = −0.73, p = .467). Between-group change was not significant (−0.59 ± 2.79 vs −0.33 ± 3.25; t = −0.43, p = .671), nor were differences in shift-specific workdays. Provision of 1 day off after more than 7 night shifts showed no significant differences at pre-test (χ² = 1.01, p = .603), post-test (χ² = 1.00, p = .590), or in change scores (t = 0.51, p = .636). Similarly, ensuring 48-h rest after 2 or more consecutive night shifts was not significantly different at pre-test (χ² = 1.72, p = .423) or post-test (χ² = 0.27, p = .796), with no significant difference in response change (“Yes”: χ² = 0.78, p = 1.000; “No”: χ² = 0.00, p = 1.000). Monthly night shifts slightly declined in the experimental group (5.67 ± 1.97 to 5.59 ± 2.76; t = 0.21, p = .834) and increased in the control group (5.53 ± 1.78 to 5.80 ± 1.17; t = −0.97, p = .334), but between-group change was not significant (−0.08 ± 2.66 vs 0.27 ± 2.01; t = −0.76, p = .451). Consecutive night shifts slightly rose in the experimental group (2.67 ± 0.68 to 2.78 ± 0.64; t = −1.03, p = .308) and remained unchanged in the control group (2.82 ± 0.48 to 2.82 ± 0.52; t = 0.00, p = 1.000). Between-group change (0.12 ± 0.82 vs 0.00 ± 0.85; t = 0.71, p = .477) was also not significant. Thus, no statistically significant differences were found within or between groups (Tables 3–5).

Table 3.

Comparison of work characteristics, health status, work-life balance between groups (N = 102).

Variables	Groups	Pre test	Post test	t	p
Variables	Groups	M ± SD	M ± SD	t	p
Maximum consecutive work days	Exp. (n = 51)	4.63 ± 1.00	4.31 ± 0.73	−1.81	.077
Maximum consecutive work days	Cont. (n = 51)	4.61 ± 0.60	4.61 ± 0.57	0.00	1.000
Numbers of shifts
Day shift	Exp. (n = 51)	6.14 ± 3.71	5.80 ± 3.22	0.51	.610
Day shift	Cont. (n = 51)	5.92 ± 3.16	5.82 ± 2.85	0.15	.879
Evening shift	Exp. (n = 51)	6.02 ± 3.02	6.18 ± 3.20	−0.28	.778
Evening shift	Cont. (n = 51)	6.82 ± 3.39	6.39 ± 2.62	0.71	.478
Night shift	Exp. (n = 51)	5.67 ± 1.97	5.22 ± 2.77	1.23	.226
Night shift	Cont. (n = 51)	5.47 ± 1.71	5.49 ± 1.77	−0.06	.949
Working days per month	Exp. (n = 51)	19.06 ± 2.19	18.47 ± 1.84	−1.51	.138
Working days per month	Cont. (n = 51)	19.43 ± 2.44	19.10 ± 2.30	−0.73	.467
Numbers of night shifts per month	Exp. (n = 51)	5.67 ± 1.97	5.59 ± 2.76	0.21	.834
Numbers of night shifts per month	Cont. (n = 51)	5.53 ± 1.78	5.80 ± 1.17	−0.97	.334
Numbers of consecutive night shifts per cycle	Exp. (n = 51)	2.67 ± 0.68	2.78 ± 0.64	−1.03	.308
Numbers of consecutive night shifts per cycle	Cont. (n = 51)	2.82 ± 0.48	2.82 ± 0.52	0.00	1.000
Health status	Exp. (n = 51)	3.49 ± 0.57	3.61 ± 0.59	1.40	.167
Health status	Cont. (n = 51)	3.62 ± 0.49	3.56 ± 0.64	−0.45	.652
Work-life balance	Exp. (n = 51)	2.98 ± 0.73	3.21 ± 0.72	2.53	.015
Work-life balance	Cont. (n = 51)	3.12 ± 0.66	2.95 ± 0.54	−1.47	.148

Table 4.

Comparison of work characteristics between groups (N = 102).

Variables		Exp. (n = 51)	Cont. (n = 51)	x²	p
Variables		n (%)	n (%)	x²	p
1 day off after 7 night shifts
Yes	Pre test	41 (80.4)	41 (80.4)	1.01	.603
No		10 (19.6)	10 (19.6)
Yes	Post test	38 (74.5)	38 (74.5)	1.00	.590^a
No		13 (25.5)	13 (25.5)
Between-shift breaks shorter than 48 h after 2 consecutive night shifts
Yes	Pre test	44 (86.3)	41 (80.4)	1.72	.423
No		7 (13.7)	10 (19.6)
Yes	Post test	43 (84.3)	41 (80.4)	0.27	.796^a
No		8 (15.7)	10 (19.6)

^aFisher’s exact test; Con. = control group; Exp. = experimental group.

Table 5.

Differences in main variables between groups (N = 102).

Variables	Categories	Exp. (n = 51)	Cont. (n = 51)	t or x²	p
Variables	Categories	M ± SD	M ± SD or n (%)	t or x²	p
Maximum consecutive work days	Post-pre	−0.31 ± 1.24	0.00 ± 0.80	−1.52	.132
Numbers of shifts
Day shift	Post-pre	−0.33 ± 4.64	−0.10 ± 4.58	−0.26	.797
Evening shift	Post-pre	0.16 ± 3.95	−0.43 ± 4.31	0.72	.474
Night shift	Post-pre	−0.45 ± 2.63	0.02 ± 2.20	−0.98	.328
Working days per month	Post-pre	−0.59 ± 2.79	−0.33 ± 3.25	−0.43	.671
1 day off after 7 night shifts
Yes	Post-pre	79 (77.5)	79 (77.5)	0.51	.636^a
No	Post-pre	23 (22.5)	23 (22.5)	0.51	.636^a
Between-shift breaks shorter than 48 h after 2 consecutive night shifts
Yes	Post-pre	87 (85.3)	82 (80.4)	0.78	1.000
No	Post-pre	15 (14.7)	20 (19.6)	0.00	1.000
No. of night shifts per month	Post-pre	−0.08 ± 2.66	0.27 ± 2.01	−0.76	.451
No. of consecutive night shifts per cycle	Post-pre	0.12 ± 0.82	0.00 ± 0.85	0.71	.477
Health status	Post-pre	0.12 ± 0.62	−0.05 ± 0.85	1.19	.236
Work-life balance	Post-pre	0.23 ± 0.65	−0.17 ± 0.81	2.72	.008

^aFisher’s exact test; Con. = control group; Exp. = experimental group.

Hypothesis 2 test

Hypothesis 2, which predicted that the experimental group working in a unit where the nurse staffing program using RPA is implemented would have better health status than the control group, was rejected. Paired t-test results comparing nurses’ health status before and after the RPA program showed an increase in the experimental group from 3.49 ± 0.57 to 3.62 ± 0.49, which was not statistically significant (t = 1.40, p = .167). In the control group, health status decreased from 3.62 ± 0.49 to 3.56 ± 0.64, also showing no significant difference (t = −0.45, p = .652) (Tables 3 and 4). The difference in change between the two groups was not significant, with the experimental group at 0.12 ± 0.62 and the control group at −0.05 ± 0.85 (t = 0.19, p = .236) (Table 5).

Hypothesis 3 test

Hypothesis 3, which predicted that the experimental group working in a unit where the nurse staffing program using RPA is implemented will show better work-life balance than the control group, was supported. The paired t-test comparing nurses’ work-life balance before and after the RPA program showed a significant increase in the experimental group from 2.98 ± 0.73 to 3.21 ± 0.72 (t = 2.53, p = .015). In the control group, work-life balance decreased from 3.12 ± 0.66 to 2.95 ± 0.54, but this difference was not statistically significant (t = −1.47, p = .148) (Tables 3 and 4). Additionally, the difference in change scores between the two groups was statistically significant, with the experimental group showing 0.23 ± 0.65 and the control group −0.17 ± 0.81 (t = 2.72, p = .008) (Table 5).

Discussion

This study aimed to develop a nurse staffing program using RPA and to evaluate its effects on work characteristics, health status, and work-life balance among nurses. The results supported Hypothesis 1, showing no statistically significant differences in work characteristics between the experimental group and the control group, reflecting that both groups followed the Nurse Shift System Improvement Guidelines and the Guidelines for Nurses’ Night Work. However, the experimental group showed higher compliance with key scheduling criteria. No nurses exceeded three consecutive night shifts; only 2.0% in the control group did so, both reflecting improvements compared to the 8.3% reported by Park et al.²⁹ Approximately 80% of nurses in both groups received at least 48 h of rest after two or more consecutive night shifts, which exceeds the 32.0% reported by Hong et al.,¹ with higher compliance in the experimental group.

Inappropriate nurse scheduling may increase the risk of treatment errors and mortality rates.³⁰ Nurse managers involved in the Nurse Shift System Improvement Pilot Project reported difficulties in manually creating 3-month schedules, particularly when considering individual preferences and institutional constraints.^19,29,31 The RPA-based scheduling system offers a practical and scalable solution for addressing the complexities of nurse staffing while supporting healthcare workforce digital transformation. By efficiently generating fair, guideline-compliant rosters, it enhances operational performance and nurse well-being, suggesting strong potential for broader adoption. Although further research is warranted to evaluate its scalability and adaptability across diverse healthcare settings.

The results did not support Hypothesis 2, indicating no statistically significant difference in health status between nurses in the experimental and control groups. Although health status scores increased in the experimental group and declined in the control group, the difference was not statistically significant. These findings are consistent with a study conducted at a Danish psychiatric hospital, which found no significant differences in behavioral, cognitive, or somatic symptoms when a work schedule reflecting individual preferences was implemented.³² This suggests that shift work itself, rather than the implementation of an automated staffing system, may be a contributing factor to the decline in nurses’ health. Shift work disrupts circadian rhythms, increasing the risk of gastrointestinal disorders, sleep disturbances, fatigue, and musculoskeletal symptoms.³³ Shift-working nurses report poorer sleep quality and higher fatigue compared to those with fixed shifts.³⁴ Moreover, they are significantly more likely to experience work-related incidents, compromising patient safety and the quality of care.³⁵

Improving health outcomes among shift-working nurses requires systemic and policy-level interventions, including minimizing long consecutive shifts, ensuring sufficient rest periods, offering flexible scheduling options, and recognizing the inherent physiological challenges posed by shift work. A comprehensive approach that prioritizes nurse health is essential to maintaining a sustainable and safe healthcare workforce. While the RPA-based scheduling system improved reflection of nurse preferences, enhanced scheduling fairness, and promoted work-life balance, statistically significant health improvements were not observed. This may be because the system primarily focuses on operational efficiency and fairness, rather than direct health-related factors, such as fatigue recovery, mental well-being, or sleep quality.^36,37 Therefore, improving nurses’ health requires additional organizational measures, including adequate staffing, guaranteed rest periods, and resilient shift rotations, along with national policies limiting night shifts and consecutive working days.^38,39

The results supported Hypothesis 3, showing that nurses in the experimental group experienced significantly better work-life balance than those in the control group. Work-life balance scores increased in the experimental group but declined in the control group, a difference that reached statistical significance. This finding aligns with a study conducted at a Danish psychiatric hospital, where nurses reported improved work-life balance when schedules reflected individual preferences.³² Irregular and unpredictable work schedules, such as night shifts and weekend duties, negatively affect nurses’ work-life balance and overall well-being.^32,40 Allowing nurses to choose shift patterns suited to their needs can reduce fatigue and improve sleep quality.³¹

In this study, although staffing constraints limited the degree of shift options, the RPA-based program allowed rosters to be finalized by the 15th of each month and planned 1 to 3 months in advance. This increased predictability likely gave nurses greater control over their work hours, contributing to enhanced work-life balance. These improvements support nurses’ health and job satisfaction while contributing to workforce retention and patient safety. Sustaining such benefits will require organizational and governmental support, including adequate staffing levels and scheduling models that align with the Korean healthcare context.

This study is the first in South Korea to develop and implement an RPA-based nurse scheduling system and assess its impact on workforce outcomes. The findings suggest that RPA may not only improve work-life balance and potentially enhance perceived fairness in scheduling, but also lessen nurse managers’ administrative burden by automating complex scheduling tasks. Nevertheless, several limitations should be acknowledged. The study was conducted at a single site using self-reported data, limiting generalizability. The effectiveness and acceptability of RPA may also vary depending on hospital size, workforce density, digital infrastructure, and institutional policies, indicating the need for further research to evaluate the system’s applicability and scalability across diverse healthcare settings.^41–43 Operational performance indicators such as scheduling efficiency, time saved, and user satisfaction were not assessed; future studies should include these objective metrics to fully capture the system’s broader value. Additionally, the work characteristics survey contained descriptive items similar to demographic questions, making reliability testing inappropriate; this was acknowledged as a limitation.

Successful implementation and wider adoption of RPA within healthcare organizations require consideration of technology acceptance perceptions. According to the technology acceptance model and the unified theory of acceptance and use of technology, perceived usefulness and perceived ease of use are critical determinants of successful technology adoption.^44,45 Future research should quantitatively or qualitatively assess how RPA was perceived during the nurse scheduling process, including satisfaction, perceived usability, and trust among frontline nurses and nurse managers. In this context, RPA should not be viewed as a standalone solution but rather as one component of a comprehensive strategy that includes nursing workforce policies, organizational culture reform, and policy-level interventions to ensure health, fairness, and efficiency.

Conclusion

This study developed and evaluated a nurse staffing program using RPA to improve scheduling efficiency, work-life balance, and compliance with staffing guidelines. As the first empirical study in Korea to implement and assess an RPA-based nurse scheduling system, it demonstrated practical benefits in clinical settings. Although no statistically significant change was observed in health status, work-life balance significantly improved in the experimental group, highlighting the value of predictable and flexible scheduling.

Publishing rosters 1–3 months in advance enhanced schedule predictability, supported personal planning, and reduced work-home conflict. The system produced fair, guideline-compliant schedules, reduced administrative workload of nurse managers, and increased transparency. It also alleviated junior nurses’ discomfort when their requests conflicted with those of senior staff.

By incorporating individual preferences while adhering to national guidelines, RPA provides a scalable, efficient solution to persistent scheduling challenges. Given the relationship between poor scheduling, nurse well-being, and patient safety, healthcare organizations should consider adopting automation technologies. Successful implementation will require institutional support, ongoing evaluation, and engagement from both leadership and frontline staff. Future implementation efforts should consider nurses’ acceptance of technology, organizational readiness, and alignment with workforce policies for ensuring sustained adoption and maximizing system’s benefits.

Supplemental Material

Supplemental Material - Development of a nurse scheduling program using robotic process automation in Korea

Supplemental Material for Development of a nurse scheduling program using robotic process automation in Korea by Sun Young Jung, Hyung-Eun Seo, Sung Hwa Hong, Eun-Young Doo in Health Informatics Journal.

Footnotes

Acknowledgements

The authors thank all the survey participants.

ORCID iDs

Hyung-Eun Seo

Eun-Young Doo

Ethical considerations

This study was approved by the Institutional Review Board of Myongji Hospital (IRB No.: MJH 2023-06-048).

Consent to participate

Participants provided written informed consent after being briefed about the study purpose, confidentiality, voluntary participation, and right to withdraw.

Author contributions

Conception and design, data acquisition, and data analysis and interpretation: SYJ, HES, SHH, and EYD. Manuscript drafting and revision: HES and EYD. Final approval of the manuscript and public responsibility for appropriate portions of the content: SYJ, HES, SHH, and EYD. Agreement to be accountable for all aspects of the work: SYJ, HES, SHH and EYD.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: this research was supported by Myongji Hospital (grant number 2301-04-01).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.*

Supplemental Material

Supplemental material for this article is available online.

References

Hong

Cho

Jung

. Analysis of the status of nursing shifts and nurses’ perceptions of work schedules in general wards based on the government. J Korean Clin Nurs Res 2021; 3: 165–178.

Rerkjirattikal

Huynh

Olapiriyakul

, et al. A goal programming approach to nurse scheduling with individual preference satisfaction. Math Probl Eng 2020; 2020: 2379091.

Moen

Baste

Morken

, et al. Menstrual characteristics and night work among nurses. Ind Health 2015; 53: 354–360.

Uekata

Kato

Nagaura

, et al. The impact of rotating work schedules, chronotype, and restless legs syndrome/Willis-Ekbom disease on sleep quality among female hospital nurses and midwives: a cross-sectional survey. Int J Nurs Stud 2019; 95: 103–112.

Liu

Lee

Chia

, et al. Exploring the association between nurse workload and nurse sensitive patient safety outcome indicators. J Nurs Res 2012; 20: 300–309.

Park

. Factors influencing work-life balance of shift nurses. Doctoral dissertation. Ajou University, 2020.

Cho

Heo

Yeom

. The impact of female nurses’ work intensity and health status on work-life balance: utilizing the 6th Korean working conditions survey. J Next-gen Converg Technol Assoc 2024; 8: 2969–2983.

Lee

Jeong

. Compliance with recommendations on work schedule for shift nurses in South Korea. Saf Health Work 2021; 12: 255–260.

Lim

. A study on shifting nurses' health conditions and job satisfaction. Master’s thesis. Hanyang University, 2002.

10.

Greenhaus

Collins

Shaw

. The relation between work-family balance and quality of life. J Vocat Behav 2003; 63: 510–531.

11.

American Nurses Association . Addressing nurse fatigue to promote safety and health: joint responsibilities of registered nurses and employers to reduce risks. https://www.nursingworld.org/practice-policy/nursing-excellence/official-position-statements/id/addressing-nurse-fatigue-to-promote-safety-and-health/ 2014, accessed 24 February 2025.

12.

Rahman

Mulyani

Rizany

. Correlation between nurse scheduling and satisfaction in patient ward Ulin Banjarmasin hospital. Dunia Keperawatan 2017; 6: 41–48.

13.

Armstrong-Stassen

Freeman

Cameron

, et al. Nurse managers’ role in older nurses’ intention to stay. J Health Organisat Manag 2015; 29: 55–74.

14.

Rizany

Hariyati

RTS

Purwaningsih

. Head nurse function optimization on scheduling based on nurse competencies: pilot study. J Persatuan Perawat Nasional Indonesia 2017; 1: 244–256.

15.

Park

Han

. A predictable pattern shift schedule model for improving nurses’ shift work. Health Soc Welf Rev 2022; 42: 258–276.

16.

Mun

. Perceptions of hospital nurses towards duty schedule using Q methodology. Master's thesis. Hanyang University, 2014.

17.

Jeon

Kim

. A study of nurse scheduling problem using efficient approximation algorithms. J Kor Inst Inform Technol 2016; 14: 159–166.

18.

Song

. A study on the nurse scheduling optimization model for nurse needs-type scheduling automation system. J Korea Soc Comput Inf 2020; 25: 57–64.

19.

D’souza

Rao

Muthana

, et al. ROTA: a system for automated scheduling of nursing duties in a tertiary teaching hospital in South India. Health Inf J 2021; 27: 14604582211001426.

20.

ElAdoly

Gheith

Fors

. A new formulation and solution for the nurse scheduling problem: a case study in Egypt. Alex Eng J 2018; 57: 2289–2298.

21.

Renggli

Gerlach

Bieri

, et al. Integrating nurse preferences into AI-based scheduling systems: qualitative study. JMIR Form Res 2025; 9: e67747.

22.

Smagul

Daineko

. Optimization of administrative processes in the medical center of the university using RPA technology: case of IITU, Almaty. Procedia Comput Sci 2024; 251: 621–626.

23.

Van der Aalst

WMP

Bichler

Heinzl

. Robotic process automation. Bus Inf Syst Eng 2018; 60: 269–272.

24.

Rizany

Hariyati

RTS

Afifah

, et al. The impact of nurse scheduling management on nurses’ job satisfaction in army hospital: a cross-sectional research. Sage Open 2019; 9: 2158244019856189.

25.

Ministry of Health and Welfare . Announcement of partial revisions to the “night shift guidelines for nursing staff,”. https://www.mohw.go.kr/board.es?act=view&bid=0003&list_no=378606&mid=a10501010100&tag= 2023, accessed 24 February 2025.

26.

Ministry of Health and Welfare . Nurse shift system improvement pilot project participating organization competition. https://www.mohw.go.kr/board.es?mid=a10501010100&bid=0003&act=view&list_no=370310&tag=&nPage=1 2022, accessed 24 February 2025.

27.

Kim

. The case for industrial workers by medical data index (MDI): evaluation of the effectiveness of health questionnaire. J Korean Soc Hortic Sci 1979; 4: 7–19.

28.

Kim

Park

. A study on the development of a work-life balance scale. J Leis Stud 2008; 5: 53–69.

29.

Korean Hospital Nurses Association . A study on the improvement of shift work for nurses working in hospital. https://khna.or.kr/home/pds/utilities.php?bo_table=board1 2024, accessed 2 February 2025.

30.

Trinkoff

Johantgen

Storr

, et al. Nurses’ work schedule characteristics, nurse staffing, and patient mortality. Nurs Res 2011; 60: 1–8.

31.

Choi

Kim

Jeon

, et al. The effects of a shift improvement pilot project on nurses’ shift adjustment, work-life balance, and turnover intention: a mixed-methods study. J Korean Clin Nurs Res 2023; 3: 223–237.

32.

Pryce

Albertsen

Nielsen

. Evaluation of an open-rota system in a Danish psychiatric hospital: a mechanism for improving job satisfaction and work-life balance. J Nurs Manag 2006; 14: 282–288.

33.

Caruso

. Negative impacts of shift work and long work hours. Rehabil Nurs 2014; 39: 16–25.

34.

Rosa

Terzoni

Dellafiore

, et al. Systematic review of shift work and nurses' health. Occup Med 2019; 69: 237–243.

35.

Ebrahemzadih

Hossein Halvani

Barsang

, et al. Association between shift work, sleepiness, chronic fatigue, and incident occurrence in nurses: a generalized structural equation modeling approach. Sleep Med Res 2024; 15: 269–277.

36.

Wang

, et al. Effects of factors related to shift work on depression and anxiety in nurses. Front Public Health 2022; 10: 926988.

37.

Petrie

Crawford

LaMontagne

, et al. Working hours, common mental disorder and suicidal ideation among junior doctors in Australia: a cross-sectional survey. BMJ Open 2020; 10: e033525.

38.

Owolabi

OlaOlorun

, et al. Work-related stress perception and hypertension amongst health workers of a mission hospital in Oyo State, south-western Nigeria. Afr J Prim Health Care Fam Med 2012; 4: 1–7.

39.

Min

Hong

. Work schedule characteristics and fatigue among rotating shift nurses in hospital setting: an integrative review. J Nurs Manag 2019; 27: 884–895.

40.

Shin

Choi

Kim

, et al. Comparison of work-life balance, fatigue and work errors between 8-hour shift nurses and 12-hour shift nurses in hospital general wards. J Korean Clin Nurs Res 2018; 3: 170–177.

41.

Carayon

Wooldridge

Hoonakker

, et al. Challenges and opportunities for improving patient safety through human factors and systems engineering. Health Aff 2020; 39: 2091–2097.

42.

World Health Organization . State of the world’s nursing 2020: investing in education, jobs and leadership. WHO, 2020.

43.

Dall’Ora

Saville

Rubbo

, et al. Nurse staffing levels and patient outcomes: a systematic review of longitudinal studies. Int J Nurs Stud 2022; 134: 104311.

44.

Venkatesh

Thong

JYL

. Unified theory of acceptance and use of technology: a synthesis and the road ahead. J Assoc Inf Syst Online 2016; 17: 328–376.

45.

Mauro

Noto

Prenestini

, et al. Digital transformation in healthcare: assessing the role of digital technologies for managerial support processes. Technol Forecast Soc Change 2024; 209: 123781.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.12 MB