Analysis of the correlation between ICU nursing human resource allocation and nursing outcome indicators: A multicenter observational study

Abstract

This study aims to analyze the correlation between nursing manpower structure index and outcome index data in intensive care unit (ICU), so as to provide a basis for reasonable allocation of nursing manpower and improvement of nursing quality. This retrospective multicenter observational study used data on ICU nursing manpower structure and nursing outcome indicators from 49 ICUs in 43 tertiary hospitals reported to the National Nursing Quality Data Platform between 2019 and 2023. Descriptive statistics, rank-sum tests were used to assess annual trends, and Spearman correlation analysis was performed to examine the associations between nursing manpower structure indicators and nursing-sensitive outcome indicators. The average daily nurse–patient ratio was negatively correlated with the unplanned extubation rate of gastric tube, the incidence rate of central related blood stream infections, and the incidence rate of catheter-associated urinary tract infections. The proportion of nurses with ≥ 10 years of experience and the proportion of nurses in charge or above was negatively correlated with the incidence rate of ventilator-associated pneumonia, central related blood stream infections, and catheter-associated urinary tract infections. Rational allocation of nursing human resources in ICU may contribute to improved nursing quality and patient outcomes; however, as a retrospective ecological analysis based on aggregated data from a single province, the findings are subject to residual confounding and cannot establish causality.

1. Introduction

With the release of “Handbook of Nursing-sensitive Quality Indicators (2016 Edition),”^[1] more and more medical institutions use nurse–patient ratio to measure the effective staffing of nurses. At the same time, in addition to the number and scale of nursing teams, managers also pay attention to the allocation of nurses’ ability structure. Appropriate nurse–patient ratio and reasonable ability structure, on the 1 hand, can enable patients who need care to obtain adequate nursing services, ensure patient safety and quality of nursing services,^[2] on the other hand, can save the human cost of medical institutions, the balance between human cost and patient outcome is a challenge for managers at present. Therefore, studies on the correlation between nursing manpower structure indicators and outcome indicators have been constantly emerging.^[3–6] Studies have shown that intensive care unit (ICU) nurse–patient ratio and nurses’ education level affect the length of ICU stay of patients, and the mortality of ICU patients is related to the nurse–patient ratio.^[7,8]

At present, most domestic studies are based on data from a single ICU or medical institution, and there is still a lack of large-scale, multicenter research on the correlation between ICU nursing human resource allocation and standardized nursing-sensitive quality indicators across different regions, hospitals and ICU types in China.

Moreover, recent umbrella reviews have shown that nurse staffing and patient outcome research from low- and middle-income countries remains scarce compared with high-income settings, underscoring the need for context-specific evidence.^[9] This study retrospectively collected the data of nursing manpower structure indicators and outcome indicators of 49 ICUs in 43 tertiary hospitals from 2019 to 2023, and analyzed their correlation, to provide evidence support for the rational allocation of ICU nursing manpower and the improvement of patient safety.

From the perspective of healthcare quality evaluation, this study was guided by Donabedian classical structure-process-outcome model, which conceptualizes structure, process and outcomes as 3 interrelated dimensions of quality.^[10,11] In this framework, ICU nurse–patient ratio, nurses’ years of experience and professional title represent key structural attributes, whereas rates of unplanned extubation and device-associated infections constitute nursing-sensitive outcome indicators that reflect the end results of care.^[12]

Based on this model and prior evidence linking better nurse staffing to improved patient outcomes, we hypothesized that ICUs with more favorable nursing manpower structures-characterized by higher nurse–patient ratios and higher proportions of experienced and senior nurses-would have lower incidences of nursing-sensitive adverse events such as ventilator-associated pneumonia (VAP), catheter-related bloodstream infections (CRBSIs), catheter-associated urinary tract infections (CAUTIs) and unplanned extubation.^[13]

2. Methods

2.1. Research design and patients

This study adopted a retrospective, multicenter, multi-year observational design at the ICU-level. The data used in this study came from the original data set of ICU nursing-sensitive quality indicators reported by medical institutions in Hebei Province from 2019 to 2023 on the National Nursing Quality Data Platform. Inclusion criteria: tertiary hospitals; ICUs participating in the report of the national nursing quality data platform for 5 consecutive years from 2019 to 2023; The data reported by ICU were complete and accurate every quarter. A total of 49 critical care departments in 43 tertiary hospitals in Hebei province met the inclusion criteria. This study used aggregated ICU-level nursing-sensitive quality indicator data reported to the National Nursing Quality Data Platform, and no individual patient records or identifiable information were collected. The study protocol was reviewed and approved by the Ethics Committee of The Second Hospital of Hebei Medical University, and the requirement for individual informed consent was waived because the analysis was based on de-identified routinely collected quality monitoring data without direct patient contact.

2.2. Variables and indicators

The variables collected included the number of duty nurses in ICU day-shift, the number of duty nurses in night-shift, the number of nurses with different professional titles, the number of nurses with different working years, the number of patients with day-shift nursing, the number of patients with night-shift nursing, the number of patients newly admitted to ICU, the number of cases of unplanned extubation of endotracheal tube, the total number of days of endotracheal tube indwelling, the number of cases of unplanned extubation of urinary catheter, and the total number of days of catheter indwelling And so on. Guidelines for Data Sets of Nursing Quality Indicators (2022 Edition).

According to the “Guidelines for Data Sets of Nursing Quality Indicators (2022 Edition),”^[14] the raw data were calculated to obtain the required index values. Among them, the ICU nursing manpower structure indicators included: the average daily nurse–patient ratio, the proportion of nurses with ≥ 10 years of experience, and the proportion of nurse-in-charge and above nurses. The nursing outcome indicators included: unplanned extubation rate of patients with catheter, incidence of catheter-related infection, incidence of stage 2 and above hospital pressure injury. The calculation formula for each indicator is as follows.

$${\displaystyle \begin{array}{cc}& {\displaystyle }Averagenurse-patientratioperday\left(1:X\right)\\ & {\displaystyle }=\frac{\begin{array}{cc}& {\displaystyle }Thetotalnumberofday-shiftandnight-shiftnursing\\ & {\displaystyle }patientsinthesameperiodeveryday\end{array}}{\begin{array}{cc}& {\displaystyle }Thetotalnumberofdayandnightduty\\ & {\displaystyle }nursesineachperiodwascounted\end{array}}\end{array}}$$

$$\begin{array}{c}{\displaystyle }\begin{array}{cc}& {\displaystyle }Theproportionofnurseswithcertainworkingyears\\ & {\displaystyle }=\frac{\begin{array}{cc}& {\displaystyle }Thenumberofnursepractitionerswith\\ & {\displaystyle }acertainworkingyearsduringthesameperiod\end{array}}{\begin{array}{cc}& {\displaystyle }Thenumberofnursepractitionersinmedical\\ & {\displaystyle }institutionsduringtheperiodwascounted\end{array}}\times 100\%\end{array}\end{array}$$

$$\begin{array}{c}{\displaystyle }\begin{array}{cc}& {\displaystyle }Theproportionofnurseswithaprofessionaltitle\\ & {\displaystyle }=\frac{\begin{array}{cc}& {\displaystyle }Thenumberofpracticenurseswith\\ & {\displaystyle }aprofessionaltitleinthesameperiod\end{array}}{\begin{array}{cc}& {\displaystyle }Thenumberofnursepractitionersinmedical\\ & {\displaystyle }institutionsduringtheperiodwascounted\end{array}}\times 100\%\end{array}\end{array}$$

$$\begin{array}{c}{\displaystyle }\begin{array}{cc}& {\displaystyle }Rateofunplannedextubationinpatientswithcatheterization\\ & {\displaystyle }=\frac{\begin{array}{cc}& {\displaystyle }Thenumberofunplannedextubationcases\\ & {\displaystyle }ofsuchcathetersduringthesameperiod\end{array}}{\begin{array}{cc}& {\displaystyle }Thetotalindwellingdaysofacertaintype\\ & {\displaystyle }ofcatheterduringtheperiodwerecounted\end{array}}\times 1000\end{array}\end{array}$$

$$\begin{array}{c}{\displaystyle }\begin{array}{cc}& {\displaystyle }Incidenceofcatheter-relatedinfection\\ & {\displaystyle }=\frac{\begin{array}{cc}& {\displaystyle }Thenumberofcatheter-related\\ & {\displaystyle }infectionsinthesameperiod\end{array}}{\begin{array}{cc}& {\displaystyle }Thetotalindwellingdaysofacertaintypeof\\ & {\displaystyle }catheterinhospitalizedpatientsduringtheperiodwerecounted\end{array}}\times 1000\end{array}\end{array}$$

$$\begin{array}{c}{\displaystyle }\begin{array}{cc}& {\displaystyle }Incidenceofstage2andabovepressureinjuryinhospitalizedpatients\\ & {\displaystyle }=\frac{\begin{array}{cc}& {\displaystyle }Numberofnewcasesofpressureinjury\\ & {\displaystyle }ofstage2ormoreinhospitalinhospitalized\\ & {\displaystyle }patientsduringthesameperiod\end{array}}{\begin{array}{cc}& {\displaystyle }Thetotalnumberof\\ & {\displaystyle }(patientsadmittedatthebeginningofthecycle\\ & {\displaystyle }+newpatientsadmittedduringthecyclewerecounted)\end{array}}\times 100\%\end{array}\end{array}$$

The unplanned extubation rate of patients with catheterization was subdivided into unplanned extubation rate of tracheal catheter, unplanned extubation rate of central venous catheter, unplanned extubation rate of gastric tube and unplanned extubation rate of urinary catheter. The incidence of catheter-related infection was subdivided into the incidence of VAP, CRBSI, and CAUTI.

2.3. Data quality control

The national nursing quality data platform embedded data logic verification rules and upper and lower thresholds. If the data were reported by ICU in violation of the data logic verification rules, the platform would prompt that the data could not be reported. After data collection, the researchers cleaned the data, and eliminated the abnormal values according to the working hours of white and night duty nurses were greater than the total working hours of ward nurses, and the total days of tracheal catheter indwelling were greater than or equal to the days of invasive mechanical ventilation, so as to ensure the accuracy of the data.

2.4. Statistical analysis

Excel software was used for data cleaning, including range checks, logical consistency checks and exclusion of records with missing or implausible values in key nursing manpower structure and nursing outcome indicators, and SPSS 22.0 software was used for data analysis. Frequency, rate and median (interquartile range) were used for descriptive statistics. The nursing manpower structure index and outcome index were tested by rank-sum test by year. Because the data set consisted of aggregated ICU-level indicators and did not include patient-level information on diagnosis, severity of illness or comorbidities, potential confounders related to ICU case mix could not be adjusted for. Spearman correlation analysis was used to explore the relationship between the nursing manpower structure index and outcome index data. Therefore, the reported Spearman correlation coefficients represent unadjusted ICU-level associations and may be influenced by unmeasured differences in acuity and case mix across units. P < .05 was considered statistically significant.

3. Results

3.1. General information

In this study, 43 hospitals covered 11 cities in Hebei Province, and the number of hospitals in each city was distributed as shown in Table 1, including 41 general hospitals (95.35) and 2 specialized hospitals (4.65%). There were 38 public hospitals (88.37%) and 5 private hospitals (11.63%). There were 40 tertiary hospitals (93.02%), 1 tertiary hospital (2.33%), and 2 tertiary uncertain hospitals (4.65%). Among the 49 ICUs, there were 42 comprehensive ICUs (85.71%), 1 surgical ICU (2.04%), 1 neurology ICU (4.08%), 1 cardiology ICU (2.04%), and 3 emergency ICU (6.12%).

Table 1.

Regional and quantitative distribution of 43 hospitals (n = 43 hospitals).

Region	Number of hospitals	Proportion (%)
Shijiazhuang city	7	16.28
Tangshan city	7	16.28
Baoding city	6	13.95
Cangzhou city	5	11.63
Langfang city	4	9.30
Xingtai city	4	9.30
Handan city	3	6.98
Hengshui city	2	4.65
Qinhuangdao city	2	4.65
Zhangjiakou city	2	4.65
Chengde city	1	2.33

Data are presented as number of hospitals and percentage of the total (n = 43).

3.2. Changes of nursing manpower structure indicators and nursing outcome indicators in each ICU

From 2019 to 2023, the average nurse–patient ratio per day in ICU was 2.28 (1.76, 2.89), the proportion of nurses with more than 10 years of experience was 29.63% (19.85%, 42.16%), and the proportion of nurse-in-charge and above nurses was 31.25% (20.54%, 40.46%). The unplanned extubation rate of endotracheal tube was 0.00‰ (0.00‰, 0.43‰), the unplanned extubation rate of central venous catheter, gastric tube and urinary catheter was 0.00‰ (0.00‰, 0.00‰), and the incidence of VAP was 1.29‰ (0.18‰, 3.44‰). The incidence of CRBSI was 0.33‰ (0.00‰, 0.84‰), the incidence of CAUTI was 0.66‰ (0.00‰, 1.43‰), and the incidence of stage 2 or above was 0.15% (0.00%, 0.60%). The annual changes of nurse–patient ratio and nursing outcome indicators in 49 ICUs are shown in Table 2. Among them, the proportion of nurses with ≥ 10 years of experience and nurse-in-charge and above increased year by year (P < .05), and the unplanned extubation rate of tracheal tube, the incidence of VAP and the incidence of stage 2 and above hospital pressure injury decreased year by year (P < .05), and the differences were statistically significant. Notably, for several device-related unplanned extubation indicators and in-hospital pressure injury, the median incidence over the 5-year period was 0.00‰ or 0.00%, indicating extremely low event rates and suggesting potential floor effects of these surveillance indicators as well as possible variability in detection or reporting practices across ICUs.

Table 2.

Situation of 49 ICU nursing manpower structure indicators and nursing outcome indicators M(QR).

First-level indicators	Second-level indicators	2019	2020	2021	2022	2023	H value	P value
Structure indicators	Average nurse–patient ratio per day (1 ∶ X)	2.23 (1.81, 3.08)	2.21 (1.75, 2.95)	2.30 (1.65, 2.89)	2.35 (1.74, 2.91)	2.28 (1.83, 2.80)	0.879	.928
	Proportion of nurses with ≥ 10 years of experience (%)	27.40 (15.54, 35.13)	27.08 (16.86, 38.35)	27.87 (22.50, 41.96)	31.11 (21.60, 43.43)	37.07 (23.71, 46.12)	10.867	.028
	Proportion of nurse-in-charge and above nurses (%)	26.55 (16.70, 33.33)	28.00 (19.80, 36.61)	31.37 (24.08, 40.00)	35.59 (26.74, 44.09)	41.75 (29.93, 53.75)	30.956	<.001
Outcome indicators	Unplanned extubation rate of tracheal tube (‰)	0.25 (0.00, 0.89)	0.00 (0.00, 0.56)	0.00 (0.00, 0.44)	0.00 (0.00, 0.00)	0.00 (0.00, 0.78)	25.018	<.001
	Unplanned extubation rate of central venous catheter (‰)	0.00 (0.00, 0.18)	0.00 (0.00, 0.13)	0.00 (0.00, 0.06)	0.00 (0.00, 0.00)	0.00 (0.00, 0.00)	6.124	.190
	Unplanned extubation rate of gastric tube (‰)	0.00 (0.00, 0.00)	0.00 (0.00, 0.31)	0.00 (0.00, 0.00)	0.00 (0.00, 0.00)	0.00 (0.00, 0.00)	11.116	.055
	Unplanned extubation rate of urinary catheter (‰)	0.00 (0.00, 0.32)	0.00 (0.00, 0.00)	0.00 (0.00, 0.00)	0.00 (0.00, 0.00)	0.00 (0.00, 0.00)	1.324	.857
	Incidence of VAP (‰)	3.11 (0.41, 6.77)	1.98 (0.49, 4.38)	0.91 (0.40, 3.31)	0.94 (0.00, 1.55)	0.96 (0.00, 2.05)	18.690	<.001
	Incidence of CRBSI (‰)	0.49 (0.00, 1.15)	0.40 (0.00, 0.82)	0.24 (0.00, 0.93)	0.19 (0.00, 0.58)	0.33 (0.00, 0.74)	6.484	.166
	Incidence of CAUTI (‰)	1.06 (0.34, 2.06)	0.90 (0.26, 1.52)	0.63 (0.29, 1.41)	0.64 (0.00, 1.09)	0.55 (0.00, 1.15)	9.467	.050
	Incidence of stage Ⅱ and above in-hospital pressure injury (%)	0.59 (0.12, 1.10)	0.18 (0.00, 0.81)	0.00 (0.00, 0.48)	0.00 (0.00, 0.49)	0.00 (0.00, 0.45)	23.640	<.001

Data are presented as the median (interquartile range). H value and P value are from Kruskal–Wallis tests across years. Rates expressed in ‰ denote events per 1000 device-days (for example, ventilator days, central venous catheter days or urinary catheter days, as defined by the national nursing-sensitive quality indicator manual), and percentages (%) denote proportions of patients. Values of 0.00 indicate that the median and lower quartile were zero across ICUs for that year, reflecting extremely low event rates and potential floor effects.

CAUTI = catheter-associated urinary tract infection, CRBSI = catheter-related bloodstream infection, H = Kruskal–Wallis H statistic, ICU = intensive care unit, OR = odds ratio, P = P value, VAP = ventilator-associated pneumonia.

3.3. Correlation analysis of nursing manpower structure indicators and nursing outcome indicators

The results showed that the average daily nurse–patient ratio was negatively correlated with the unplanned extubation rate of gastric tube, the incidence of CRBSI, and the incidence of CAUTI, as shown in Table 3. The proportion of nurses with ≥10 years of experience and the proportion of nurse-in-charge and above were negatively correlated with the incidence of VAP, CRBSI, and CAUTI, as shown in Tables 4 and 5. Based on 245 ICU-year observations, these ICU-level associations ranged from small to moderate in magnitude. The average daily nurse–patient ratio showed weak negative correlations with the unplanned extubation rate of gastric tube (r = −0.186, 95% confidence interval [CI] −0.304 to −0.062), the incidence of CRBSI (r = −0.143, 95% CI: −0.264 to −0.018) and the incidence of CAUTI (r = −0.142, 95% CI: −0.263 to −0.017) (Table 3). The proportion of nurses with ≥10 years of experience was moderately negatively correlated with the incidence of VAP (r = −0.203, 95% CI: −0.320 to −0.080), CRBSI (r = −0.142, 95% CI: −0.263 to −0.017) and CAUTI (r = −0.130, 95% CI: −0.251 to −0.005) (Table 4), and similar associations were observed for the proportion of nurse-in-charge and above nurses (VAP: r = −0.233, 95% CI: −0.348 to −0.111; CRBSI: r = −0.144, 95% CI: −0.265 to −0.019; CAUTI: r = −0.152, 95% CI: −0.272 to −0.027) (Table 5).

Table 3.

Correlation between nurse–patient ratio and outcome indicators in 49 ICUs (2019–2023).

Dependent variables	r value	95% CI of r	P value
Unplanned extubation rate of tracheal tube	−0.061	−0.185 to 0.065	.340
Unplanned extubation rate of central venous catheter	−0.054	−0.178 to 0.072	.401
Unplanned extubation rate of gastric tube	−0.186	−0.304 to −0.062	.003
Unplanned extubation rate of urinary catheter	0.110	−0.016 to 0.232	.087
Incidence of VAP	−0.148	−0.268 to −0.023	.073
Incidence of CRBSI	−0.143	−0.264 to −0.018	.027
Incidence of CAUTI	−0.142	−0.263 to −0.017	.026
Incidence of stage Ⅱ and above in-hospital pressure injury	0.052	−0.074 to 0.176	.416

Analyses are based on 245 ICU-year observations from 49 ICUs over the period 2019 to 2023 and represent unadjusted exploratory ICU-level associations between nursing manpower structure indicators and nursing-sensitive outcome indicators; no risk adjustment for patient acuity, case mix or hospital-level factors was performed. Data are presented as Spearman correlation coefficients (r) with 95% confidence intervals (CI) and P values.

CAUTI = catheter-associated urinary tract infection, CI = confidence interval, CRBSI = catheter-related bloodstream infection, ICU = intensive care unit, r = Spearman correlation coefficient, VAP = ventilator-associated pneumonia.

Table 4.

Correlation between proportion of nurses with ≥ 10 years of experience and outcome indicators in 49 ICUs (2019–2023).

Dependent variables	r value	95% CI of r	P value
Unplanned extubation rate of tracheal tube	−0.117	−0.239 to 0.008	.067
Unplanned extubation rate of central venous catheter	−0.182	−0.3 to −0.058	.094
Unplanned extubation rate of gastric tube	−0.061	−0.185 to 0.065	.338
Unplanned extubation rate of urinary catheter	−0.062	−0.186 to 0.064	.333
Incidence of VAP	−0.203	−0.32 to −0.08	.001
Incidence of CRBSI	−0.142	−0.263 to −0.017	.028
Incidence of CAUTI	−0.130	−0.251 to −0.005	.041
Incidence of stage Ⅱ and above in-hospital pressure injury	−0.049	−0.173 to 0.077	.441

Analyses are based on 245 ICU-year observations from 49 ICUs over the period 2019 to 2023 and represent unadjusted exploratory ICU-level associations between nursing manpower structure indicators and nursing-sensitive outcome indicators; no risk adjustment for patient acuity, case mix, or hospital-level factors was performed. Data are presented as Spearman correlation coefficients (r) with 95% confidence intervals (CI) and P values.

CAUTI = catheter-associated urinary tract infection, CI = confidence interval, CRBSI = catheter-related bloodstream infection, ICU = intensive care unit, r = Spearman correlation coefficient, VAP = ventilator-associated pneumonia.

Table 5.

Correlation between proportion of nurse-in-charge and above nurses and outcome indicators in 49 ICUs (2019–2023).

Dependent variables	r value	95% CI of r	P value
Unplanned extubation rate of tracheal tube	−0.159	−0.279 to −0.034	.084
Unplanned extubation rate of central venous catheter	−0.098	−0.221 to 0.028	.125
Unplanned extubation rate of gastric tube	−0.088	−0.211 to 0.038	.171
Unplanned extubation rate of urinary catheter	0.060	−0.066 to 0.184	.346
Incidence of VAP	−0.233	−0.348 to −0.111	<.001
Incidence of CRBSI	−0.144	−0.265 to −0.019	.026
Incidence of CAUTI	−0.152	−0.272 to −0.027	.017
Incidence of stage Ⅱ and above in-hospital pressure injury	−0.145	−0.266 to −0.02	.181

Analyses are based on 245 ICU-year observations from 49 ICUs over the period 2019 to 2023 and represent unadjusted exploratory ICU-level associations between nursing manpower structure indicators and nursing-sensitive outcome indicators; no risk adjustment for patient acuity, case mix, or hospital-level factors was performed. Data are presented as Spearman correlation coefficients (r) with 95% confidence intervals (CI) and P values.

CAUTI = catheter-associated urinary tract infection, CI = confidence interval, CRBSI = catheter-related bloodstream infection, ICU = intensive care unit, r = Spearman correlation coefficient, VAP = ventilator-associated pneumonia.

4. Discussion

4.1. Status quo of ICU nursing human resource allocation

Compared with the bed-nurse ratio, the nurse–patient ratio can dynamically reflect the matching relationship between the number of hospitalized patients needing care and the nursing manpower. With the improvement of medical information, more and more managers began to use the nurse–patient ratio as an index to measure the rationality of nursing human resources. California in the United States is the earliest administrative region in the world to enforce the minimum nurse–patient ratio. Act No. 394 clearly stipulates that the minimum nurse–patient ratio in the ICU is 1:2, and the minimum nurse–patient ratio in the trauma ICU is 1:1.^[15] In Germany, the nurse–patient ratio is divided into day-shift nurse–patient ratio and night-shift nurse–patient ratio according to Regulations on the determination of the lower limit of nurses in Sensitive Areas of Hospital Nursing, and the lower limit of the day-shift nurse–patient ratio in ICU is 1:2.5, and the lower limit of the night-shift nurse–patient ratio is 1:3.5.^[16] In 2012, China put forward the “nurse-patient ratio of 2.5 to 3:1 in ICU.”^[17] According to the 2021 National Medical Service and Quality and Safety Report. Nursing Professional Volume, the median nurse–patient ratio of 908 tertiary general hospitals ICU was 1:2.11.^[18] 2.28, which met the requirement of “nurse-patient ratio of 2.5 to 3:1 in ICU” in China, indicating that medical institutions pay attention to the allocation of nursing human resources in ICU, which is in line with the current policy requirements and not much different from the national level.

On the basis of satisfying the quantity of ICU nursing manpower, we should also pay attention to the reasonable allocation of nursing staff’s ability. The indicators commonly used to evaluate the ability of ICU nurses include educational background, working years, and health technical titles. Due to the accessibility of data, this study selected working years and health technical titles as the evaluation indicators of ICU nurses’ ability. At present, there is no clear requirement for working years and the proportion of professional title allocation in China. The World Health Organization recommends that the composition ratio of senior, intermediate and primary nurses’ professional title structure is 1:4:1.^[19] Reports show that the median number of nurses with ≥10 years of experience in comprehensive ICUs in China is 24.16%, and the median number of nurse-in-charge and above is 18.44%.^[18] This study shows that the median of ICU nurses with ≥10 years of experience in Hebei Province is 29.63%, and the median of nurse-in-charge and above is 31.25%, which is higher than the national level and shows an increasing trend, but has not yet reached the proportion recommended by the World Health Organization.

Under the current medical and health system reform, human resource cost is an important part of hospital operating cost, and some studies show that human resource cost of some hospitals has accounted for 60% of the total cost.^[20] In this context, it is suggested that nursing managers should reasonably allocate the number of nursing staff according to workload, difficulty and risk, and build effective talent echelon according to different ability evaluation dimensions, especially in departments with high risk and difficulty to help ensure patient safety and support safe, high-quality care.

4.2. Influence of ICU nursing manpower structure indicators on nursing outcome indicators

At present, there are few studies on the correlation between nursing manpower structure and nursing outcomes in China, most of which are single ICU data or cross-sectional surveys, and lack of regional longitudinal studies using standardized ICU nursing-sensitive indicators. Drawing on 5 consecutive years of data from 49 ICUs in 43 tertiary hospitals in Hebei Province, this provincial-level multicenter study provides longitudinal evidence on the associations between ICU nursing manpower structure and device-related nursing-sensitive adverse events in the Chinese context, complementing previous single-center or cross-sectional work.

4.3. Influence of nurse–patient ratio on nursing outcome indicators

Most studies on unplanned extubation in ICU focus on tracheal intubation. There is no study on the correlation between unplanned extubation of gastric tube, central venous catheter and urinary catheter and nursing manpower allocation at home and abroad. On the other hand, studies on unplanned extubation focus on risk factors such as patients themselves, catheter types, fixation methods, and effective restraints.^[21–26] A 9-year retrospective analysis^[23] found that standardized endotracheal tube management protocols were the safest and most effective associated with reductions inunplanned extubation. Yang Leiet al^[4] showed that ICU nurse–patient ratio had nothing to do with unplanned extubation of tracheal intubation, which was consistent with the results of this study. According to the data mining of 210 cases of unplanned extubation of gastric tube, it was found that night time was the risk factor for unplanned extubation of gastric tube, which was related to the reduction of the number of nurses in night-shift compared with that in day-shift,^[27] which was similar to the results of this study.

A retrospective study of 353 Class Ⅲ Grade A hospitals showed^[28] that the nurse–patient ratio was the influencing factor of CAUTI, and the incidence of CAUTI would increase by 9% for every increase in the number of patients cared by nurses. Another single ICU retrospective study^[4] showed that the ICU nurse–patient ratio was correlated with the incidence of CAUTI, which was the same as the results of this study. The difference is that the incidence of CRBSI in this study is not related to the nurse–patient ratio, which may be related to the different working environment, disease spectrum and severity of patients. Another cross-sectional survey took the actual bed-nurse ratio as the indicator of nursing human resource allocation, and the results showed that ICU bed-nurse ratio was positively correlated with the incidence of CRBSI and CAUTI,^[29] which was similar to the results of this study.

4.4. Influence of nurses’ working years and professional title allocation on nursing outcome indicators

ICU is the most concentrated nursing unit for critically ill patients, with the most advanced function and the largest number of instruments and equipment. The patient’s condition changes rapidly, requires close observation of the condition, and is ready for rescue at any time, so the nurses are required to master a lot of knowledge and operation skills. With the increase of working years, nurses will gradually accumulate clinical knowledge and skills, and cultivate keen observation and resilience. With the increase of health technical titles, nurses assume the tasks of teaching, quality control and scientific research, and their theoretical knowledge is more comprehensive and abundant. They can identify and predict the change trend of the disease, take intervention measures to prevent various complications and improve the outcome of patients. Studies have shown that the incidence of VAP is significantly negatively correlated with the proportion of ICU nurses with 5 to 10 years of work experience.^[30] A study involving the allocation of nursing human resources in 57 ICUs showed that^[31] the proportion of ICU nurses with <5 years of work experience was positively correlated with the incidence of VAP and CRBCI, and the higher the proportion of nurses with intermediate professional titles or above, the lower the incidence of CRBSI. Another cross-sectional survey showed that the proportion of intermediate professional titles in ICU was negatively correlated with the incidence of VAP.^[29] Liumin Ding et al^[32] studied on ICU of 11 tertiary general hospitals in Xinjiang showed that the proportion of nurses working for >5 years was negatively correlated with the incidence of CAUTI, and the proportion of nurses with senior titles or above was negatively correlated with the incidence of VAP and CRBSI. Therefore, managers should rationally allocate the structure of ICU nursing staff, innovate the human incentive guarantee mechanism, and encourage nurses to continuously improve their professional ability, so as to help ensure patient safety and may be associated with improved patient outcomes.

The working years selected as the human resource structure index in this study are correlated with the health technical title, which has certain limitations as an index to evaluate the ability of nursing staff. In the future, it can be combined with other dimensions such as initial education, highest education, specialist nurse qualification and so on. In addition, this study has several important methodological limitations that should be taken into account when interpreting the findings. First, because our analysis was conducted at the ICU-level using aggregated annual indicators and unadjusted Spearman correlations, we were unable to perform risk-adjusted modeling or to control for key potential confounders such as ICU type (for example, general vs specialty ICUs), patient acuity and case mix, staffing models and scheduling patterns, nurse skill mix and hospital-level management or quality policies; as a result, residual confounding is likely and the observed associations may be partly explained by unmeasured factors. Second, the nursing-sensitive outcome indicators (such as device-related infections and unplanned extubations) relied on routine surveillance and reporting by participating ICUs, so under-reporting, differences in diagnostic thresholds or variability in documentation may have introduced reporting bias, especially for outcomes with very low event rates (0.00‰ or 0.00%). Third, the ecological nature of the data means that the correlations we report describe relationships at the ICU-level and should not be directly extrapolated to individual patients; making inferences about individual risk based on group-level associations would constitute an ecological fallacy, a well-recognized issue in epidemiology and health disparities research. Taken together, these limitations indicate that our results should be interpreted as exploratory ICU-level associations between nursing manpower structure and nursing-sensitive outcome indicators rather than as evidence of a causal effect of specific staffing configurations on patient outcomes.

5. Conclusion

This study analyzed 5-year ICU nursing manpower structure indicators and nursing outcome indicators from 49 ICUs in 43 tertiary hospitals in Hebei Province. Higher average daily nurse–patient ratios and higher proportions of experienced and senior nurses were associated with lower incidences of VAP, CRBSI, and CAUTI at the ICU-level. These findings suggest that optimizing ICU nursing manpower structure may help support patient safety and may be associated with improved nursing-sensitive quality indicators, but they are based on unadjusted observational data from a single province and should be interpreted as exploratory associations rather than evidence of a causal effect of specific staffing configurations. Future multicenter studies incorporating risk adjustment for patient acuity, case mix and organizational factors are needed to provide more robust evidence to guide ICU staffing decisions.

Author contributions

Conceptualization: Jing Li, Bin Zhao.

Data curation: Jing Li.

Formal analysis: Jing Li.

Investigation: Jing Li.

Project administration: Ni Chen.

Resources: Bin Zhao, Jie Liu.

Software: Ni Chen.

Supervision: Bin Zhao.

Visualization: Jing Li

Writing – original draft: Jing Li.

Writing – review & editing: Bin Zhao, Ni Chen, Jie Liu.