Development of the NHS Job Evaluation Tool: A Quantitative Study of Nurses in China

Aim: Based on the National Health Service job evaluation scheme, we aimed to construct a quality indicator for Chinese nurses. Background: Because most nurses’ low job satisfaction has led to a high turnover rate, establishing a job evaluation system that truly reects nurses' values has become the core of Chinese nursing work reform. Methods: Based on the literature review and Delphi method, a questionnaire containing 30 items was designed. By stratied sampling, nurses from six tertiary public hospitals in Beijing were selected. We collected 701 valid questionnaires. SPSS 24.0 was used to extract factors and analyze the questionnaire’s reliability and structural validity. Results: We retained 26 factors and divided them into four dimensions: Knowledge and Skills, Responsibility, Work/Environment, and Task Characteristics according to principal component analysis. The Cronbach’s α of four dimensions ranged from 0.890 to 0.923, indicating high reliability. The Kaiser-Meyer-Olkin value was 0.960, demonstrating good structural validity. Conclusions: The improved job evaluation system has strong reliability and validity and reects the nurse’s job characteristics scientically. Implications for


Background
Nurses are crucial in achieving universal health coverage and global sustainable development goals (Taylor et al., 2020). Nursing staff account for approximately 59% of health professionals and are the largest professional group in health sector. They are playing an integral part in aging care and hospice programs. However, the shortage of nurses has become a global problem (Sermeus et al., 2011).
The number of nursing staff shortages will reach 5.7 million by 2030, most of which are concentrated in low-and middle-income countries(2020a). In addition, nurses' turnover rate is as high as 40%, and the global population over 60 years old will increase to 1.4 billion by 2030. This series of problems makes the nursing industry face severe challenges (Marc et al., 2019). Classical economic theory interpreted income level as a signi cant factor in employee turnover (Drennan & Ross, 2019). The National Federation of Nurses Associations in Taiwan also found that about 60% of nurses' resignation reasons are related to "salary and bonus", "heavy workload", and "work-life imbalance" (Chen et al., 2020). However, the income of nurses in most countries is lower than the average salary in that country (WHO, 2016). A survey from China shows that 83.7% of nursing staff believe that salary is the most important motivating factor in their work, but 76.5% of nurses have a monthly income lower than the average salary in rst-tier cities (735$), which directly leads to the low quality and the decline in patient satisfaction. How to distribute salaries and retain nurses has become a challenge faced by almost all countries(2020a).
Job evaluation is considered to be the basis for scienti cally formulating the salary distribution system and maintaining fair competition in the organization, and it plays a key role in human resource management (Zhang et al., 2021). Job evaluation refers to the design of a reasonable salary structure by comparing the job's relative value. It is mainly divided into four methods: ranking, classi cation, job component, and point system. The point factor method is more commonly used. The point factor method decomposes speci c job tasks into four aspects: skills, responsibilities, effort, and working environment based on organizational strategy and job characteristics. In 1990, the United Kingdom adopted the point factor method to design the National Health Service (NHS) job evaluation system. The system includes ve evaluation dimensions and 16 speci c indicators with different weights. Compared with other systems, the NHS has strengthened the importance of knowledge factors, with a total score of up to 240 points, accounting for 24 of all factors. Furthermore, responsibility factors accounted for 36%, authority factors accounted for 6%, skill factors accounted for 24%, and effort and environmental factors accounted for 10%. Therefore, the NHS job evaluation system is a universal system with high authority and feasibility.
Establishing a job evaluation system that truly re ects nurses' values has become a central topic in China's nursing work reform. The Chinese government issued a notice on 2 September 2020, requiring reasonable adjustments of nursing service prices and nursing service costs accounting, gradually straightening out the relationship between nursing service prices and mobilizing nurses' enthusiasm (2020b). At present, public hospitals in China implement a job performance wage system. The nursing staff's wages include job wages, salary scale wages, performance wages, and allowances and subsidies, but the evaluation system has not been effectively implemented. Most public hospitals in China are run by the government and belong to public institutions. Therefore, nurses' job evaluation in public hospitals mainly focuses on job salary, and less considers job skills, job environment, and responsibility risk factors in nursing work and cannot truly re ect the value and transfer of nurses' labor services-the work enthusiasm of nurses (Xin & Lijuan, 2020). An empirical study of 25 hospitals found that 53.4% of nurses believe that the current salary distribution method is unreasonable (Liping et al., 2011). Moreover, there is a widespread phenomenon of "different pay for the same work" in hospitals. Nursing staff with a permanent job and nursing staff with contract jobs have the same job but the salary difference is Job evaluation models from experienced countries were peachy guides for reforming the juxtaposition of Chinese nursing staff. The auspiciousness of the NHS job evaluation system is solving the internal inequity in salary distribution of medical staff. Because of the differences in each country's economic level and organizational culture, we revised the NHS job evaluation system, selected relevant evaluation factors through literature research and the Delphi method, and formulated a questionnaire suitable for evaluating nurses in public hospitals in China. We tested whether the questionnaire is reliable and valid to draw statistical and logical signi cance.

Questionnaire design
Because of the evident merits of the 'Point factor method', we decided to use it to build the nurses' job evaluation model. The Delphi method was used to select the preliminary assessment elements and indicators ground on an extensive literature review. We adopted 'Job evaluation, point factor method, job analysis model' as keywords by searching in the PubMed, Web of Science, Google, CNKI, Wan Fang, CBM and OVID database. Then we collected and classi ed all the 'points', plus their corresponding frequency.
Considering that the national conditions of each country are different and that China is the most populous country in the world, it faces a more complicated situation when evaluating work than the United Kingdom, so we retained 16 factors from the NHS job evaluation system and included another 14 factors from other job evaluation models drafted by government or advisory body (2008, 2010, 2017Johnson et al., 2015). After that, we invited 20 Chinese experts to conduct two rounds of Delphi consultations, helping us winnow the above elements further. Experts recognized these 30 indicators and further explained these factors' connotation to design the scale question later. These connoisseurs showed great representativeness and reliability; the expert authority coe cient for all items was higher than 0.8. Ultimately, the nal versions of the questionnaire contained 30 items, and each was rated by a 5-point Likert scale, running from 'extremely signi cant' (5 points) to 'totally insigni cant' (1 point). Fifty nurses from a tertiary public hospital in Beijing participated in the pretest to determine the questionnaires' reliability and validity. The result shows that the overall Cronbach's α was 0.955, and the test-retest reliability coe cient was 0.960, demonstrating satisfactory reliability. The Kendall coordination coe cient for all items and the Bartlett's Test of Sphericity were signi cant(p < 0.05). Furthermore, Kaiser-Meyer-Olkin was valued at 0.960. Consequently, the questionnaire had great structural validity.

Sample selection
Samples of this study were selected by a multistage sampling design, which included the following two steps.
First step: According to the statistical yearbook, the six districts with the densest public hospitals in Beijing were selected, and one tertiary public hospital was selected in each district through cluster sampling, and six hospitals were nally determined as the research objects. Each of them holds 800 to 1500 beds, plus 1200 to 3500 staff.
Second step: A total of 800 nurses were selected by random sampling from the above six public hospitals to participate in the questionnaire survey.

Data processing and statistical analysis
After collecting the questionnaire, we utilised Microsoft Excel 2016 for basic data processing. The questionnaire's exclusion criteria include three principles: 1) Three or more items were not answered; 2) The answers to all items were the same; 3) The response showed a prominent circular mode. Moreover, after entering all the data, we selected 5% of the data to assure the data input accuracy. SPSS 24.0 was used for statistical analysis. We primarily carried out principal component analysis with a varimax rotation to extract features and reduce dimensionality. Then we saved the factors whose Eigenvalue was greater than one but extracted the items whose factor load was less than 0.5 from the analysis. In addition, we applied the Cronbach's α to estimate the reliability of the scale and analysed the correlation among the items from the identical dimension and different dimensions. The Kaiser-Meyer-Olkin value was used to test the questionnaire' validity.

Demographics of respondents
We distributed 800 questionnaires and recycled 721 among them, which means the response rate was 90.1%. Besides, after data preprocessing, we identi ed 701 valid questionnaires with a validity rate of 97.2%. The demographic characteristics of the samples were depicted in Table 1. Because female nurses accounted for the majority in China, 79.6% of our respondents were female, and 45.9% were under 30, while 33.1% were between 30 to 39 years old. Compared with physicians, 60.6% of selected nurses did not have a Bachelor's Degree, and 59.6% were attending nurses. Beyond half of them had been working within ten years. Regarding departments, 96.3% of respondents were from clinical divisions and almost equally distributed among six sampling hospitals. There were no signi cant differences between the exploratory and con rmatory samples(p > 0.05), illustrating the data had gratifying conformance.

Factor analysis and dimensions
Through further analysis of the data, the Bartlett's Test of Sphericity was signi cant (p < 0.05), and the Kaiser-Meyer-Olkin value was 0.960, showing the rationality of principal component analysis for extracting key factors. We ultimately retained the principal components that eigenvalues were greater than one and de ned them as four dimensions of nurses' job evaluation model. They accounted for 65.07% of the total variance cumulatively.
Then we exploited varimax rotation to convert the above components into easily interpretable factors. All items with factor loadings over 0.5 were included, so we deleted four items (Innovation skills, Awareness of quality and safety, Responsibilities for patient/client care, thinking environment), leaving 26. Due to the de nitions of the remaining items under each dimension, we named four dimensions as Knowledge and skills (10 items), Responsibility (6 items), Effort/Environment (5 items) and Task characteristics (5 items). Table 3 summarized the factor loadings for each item.

Reliability and validity analysis
The Cronbach's α of the raised four dimensions were between 0.890 to 0.923, and each items' coe cient was displayed in Table 4. Moreover, according to the corrected item-total correlation, the system's items and dimensions' both had a good reliability. Apart from that, the correlation coe cients among items within the same dimensions were all over 0.593, which meant all these factors presented great consistency. As shown in Table 5, there was also exceptional consistency between the four dimensions, as their correlation coe cients were greater than 0.526. Task and Effort/Environment were the most correlated (r = 0.738). However, Effort/Environment and Responsibility were the least correlated (r = 0.526). As for validity, the results of factor analysis proved that our model had quali ed structural validity. Based on mature job evaluation schemes and Delphi method, the content validity can be well determined. As a result, the nurse job evaluation system possessed both reliability and validity.

Discussion
Based on the above research, we nally constructed a job evaluation system with four dimensions and 26 basic projects to scienti cally and equally evaluate nurses' work and performance in Chinese public hospitals. Compared with the previous uniformly distributed system, our study had a satisfactory reliability and validity with Cronbach's α all above 0.8.

Feature of Chinese nurse job evaluation
Chinese nurse job evaluation scheme combines the nursing industry's latest development trends and can re ect the characteristics of nurses' work. First, people's increasingly diversi ed health needs and the accelerated change of knowledge place higher demands on nursing staff's Knowledge and skills. In view of Chinese nurses' weak ability to translate theory into practice (Jinnan, 2016), China issued a standardized training program for new nurses in 2016 (Zhuangzhi et al., 2020). These characteristics can be re ected in the knowledge/skills dimension (10 factors). Second, to ensure patient safety, the Ministry of Health puts forward the concept of responsible holistic care, requiring nurses to provide patients with comprehensive and continuous services (Xia & Pengjuan, 2016). Besides essential nursing work, nurses are also responsible for policy implementation and economics of medical consumables usage. These characteristics can be re ected in the responsibility dimension (6 factors). Third, with the rapid increase in the number of doctors in China and the particularity of nursing work, nurses are under increasing pressure in terms of physiology, psychology and emotion . Research shows that work exibility is the primary source of occupational pressure for nurses (Changrong et al., 2011). These characteristics can be re ected in the effort/environment dimension (5 factors). The frequent occurrence of new infectious diseases and emergency patients' increase put forward higher requirements for nurses' task processing ability. These characteristics can be re ected in the task dimension (5 factors).
Chinese nurse job evaluation scheme re ects the prospective requirements responsibilities to the dimension of responsibility to ensure that nursing students and patients receive comprehensive guidance from nurses.
The application of Chinese nurse job evaluation scheme As for applying this work system, it can be used to evaluate the job characteristics of nurses because its four dimensions, namely "knowledge and skills, responsibilities, effort/environment and task characteristics," relatively completely cover 701 recipients. All the factors most concerned by the interviewer re ect their expectations. In addition, the system can also be used to determine reasonable post salary, personnel selection and appointment, formulate performance evaluation standards, and build career development plans (Yang et al., 2015). At present, two hospitals in China are already using this system, and the speci c effect needs further evaluation.

Limitations
This study also contains some limitations. Our survey only investigated nurses from public hospitals in Beijing, which had a limited geographical range. In the following research, we should expand regions and improve nurses' representative from the whole country. Second, the established nurses' job evaluation system has only been implemented in two hospitals of Beijing. Thus, it is necessary to further verify its applicability through practice.

Conclusion
This study designs a model containing four dimensions and 26 speci c factors to evaluate nurses' work performance in Chinese public hospitals. The model is a reliable and valid system that re ects nurses' position characteristics relatively scienti cally. Our research lls up the current research and practice gaps and provides the essential basis for the comprehensive management of nursing human resources to achieve "equal pay for equal work, better pay for better work".

Implications For Nursing Management
Nursing managers can use this job evaluation model to comprehensively evaluate nurses' job value, improve nurses' enthusiasm for work, and reduce their turnover rate. The information age poses more challenges to the work of nurses. Applying this model to nurses' performance assessment will improve nurses' overall quality in an all-around way so that patients can enjoy more satisfactory services. In addition, the method adopted in this study also has speci c reference signi cance for other countries to establish reasonable work evaluation models. China in 2016, our study did not need formal ethics approval because this survey did not deliver any intervention to research participant and involved no more than minimal risk of harm to participants. The questionnaire surveys were anonymous and voluntary and written informed consent had been obtained from all study participants. All methods were performed in accordance with relevant guidelines and regulations.

Consent for publication:
Not applicable.

Data availability:
The datasets used during the study are available from the corresponding author on reasonable request.

Con ict of Interest:
No competing interests. Author contribution JG contributed to the study design, data analysis and drafting the major portion of the manuscript. DZ contributed to the conceptual framework and data collection. FY and RJH contributed to data analysis and writing the manuscript. All authors critically reviewed the manuscript and approved the nal version.