Study design:
The study utilized a methodological research design (Cross-sectional study with a longitudinal component for reliability).
Settings:
Data were collected from 16 hospitals (out of 17 hospitals) in Port-Said city, Egypt, accepted to participate in the study. Eight hospitals are affiliated to the Ministry of Health, two follow Health Insurance Hospitals, and six are private hospitals. Data collected from beginning of October 2018 to the end of March 2019.
Sampling method and techniques:
The target population included all nurses in the identified hospitals at the time of data collection (N = 1331). Current study included nurses who had experienced more than six months in the current work setting. The nurses who were in labor, sickness and childcare vacations during the data collection period were omitted.
The sample size was estimated depending on the item sample ratio method. Per rule of thumb, it is instructed to recruit a minimum of 10 participants in each item of the scale [17]. The total participants needed to be 240. Assuming a 10% dropout rate, an additional 24 participants were added to be 264. This increased to 335 to achieve sufficient power and precision for the study [18]. Due to incompleteness, seven questionnaires were excluded. Hence, a sample of 328 nurses was recruited for the study. The sample size was distributed proportionally according to total nurses in each hospital. A systematic random sample from each hospital was selected from the list of all nurses available at the hospital administration.
Measures:
The human resource practices (HRP) scale
The HRP scale was developed by Villajos et al. [9], in English to assess perceptions of employees of human resource practices. It is containing 24 items covering two bundles. The first is performance enhancement bundle and the second is employee support bundle. Performance enhancement bundle include five practices are training and development, contingent pay and rewards, performance appraisal, recruitment and selection, and a competitive salary. Employment security, work-life balance, and exit management in the employee-support bundle. The 24 items measuring along a 5-point Likert-type scale as follows: not at all (1), very little (2), some (3), a fair amount (4) and a lot (5). Demographics questionnaire sheet was attached to the scale to collect data related to gender, age, level of education, position, and years of experience in nursing.
Criterion scales:
Baldrige dimension scale [19] and six dimension scale of nursing performance (6-D) [20] were used as external criterion for measuring concurrent validity. Baldrige dimension scale was validated in Arabic by Badri et al. [21] to assess Baldrige Health Care Criteria for performance excellence. The scale composed of 115 items measuring seven categories; leadership, information and analysis, strategic planning, human resource development and management, process management, organization performance results, focus on and satisfaction of patients and other stakeholders. Those seven categories subdivided to 28 dimensions. The responses of participants have collected a long 7-point Likert-type scale.
Six dimension scale of nursing performance (6-D) validated in Arabic by Miloud [22]. It composed of 52 items under six dimensions namely leadership, critical care, teaching/collaboration, planning/evaluation, interpersonal relations/communication, and professional development. Each item was scored using a 4-point scale.
Procedures:
The translation and culture adaptation of the HRP scale followed international guidelines of cross-cultural adaptation of health questionnaires [23]. The process was performed in five stages: forward translation, synthesis of the translated versions, back translation, expert committee, and test of the pre-final version. Psychometric testing (validity and reliability) was calculated to the final Arabic version of HRP scale (Fig. 1).
Stage I: forward translation
The HRP scale was translated from the original language, English; into Arabic by two separate bilingual translators whose mother tongue is Arabic. One translator was a fluent English nurse manager who is familiar with the terminology and construction of HRP in healthcare. The other was a certified translator. This step generated two forward-translated versions of the instrument (T1.T2).
Stage II: synthesis of the translated versions
The two translators and the first author evaluated the two translated versions (T1 & T2), creating an agreed a synthetic Arabic version of the HRP scale (T1-2).
Stage III: back translation
The synthetic Arabic version of the HRP scale (T1-2) was back translated into English by two other independent qualified translators, one of them was bicultural. The two translators were blind to the original English version. As a result, two back-translated versions (BT1& BT2) were produced.
Stage IV: expert committee
The HRP scale was cross-culturally adapted by a multidisciplinary bilingual expert panel consisting of the four translators involved in the forward and back-translations, one nursing administration university professor, and one nursing director with a master degree in nursing administration. The committee was asked to review produced material (T1.T2, T12, BT1, and BT2) and the original version. The necessary modifications were introduced by consensus to realize semantic, experiential, idiomatic, and conceptual equivalence between the original version and the target one. At the end of this stage, researchers produced a consolidated pre-final version of the HRP scale in Arabic.
Stage V: test of the pre-final version (pilot study)
In this stage, 30 nurses (not included in the full-scale study) tested the pre-final Arabic version of the HRP scale. After that, the cognitive debriefing process was utilized, each nurse who completed the scale was interviewed about the meaning of each item and the chosen response. A minor modification was done at this stage and final Arabic version was produced.
Psychometric testing of the Arabic version of the HRP scale
Validity
Face, content, construct, and concurrent validity were used to validate the HRP scale into Arabic. Face validity is the degree to which the Arabic version of the current scale measures superficially HRP. Eleven experts valued the scale readability, feasibility, formatting and style consistency, and language clarity [24]. All experts assured that the proposed Arabic version has good face validity for assessing HRP.
Content validity evaluates the degree to which the Arabic version of the current scale includes all the key elements related to HRP [25]. Eleven experts in the field of nursing administration, business administration and human resources management examine the content validity for the Arabic version of HRP scale. The experts were asked independently to review each item using 4 points ordinal scale (1 = disagree, 2 = need modification, 3 = agree, 4 = highly agree) [26].
The content validity index was estimated at the item level (I-CVI) and scale level (S-CVI). To obtain the content validity index at the item level (I-CVI), the number of experts judging the item as relevant or clear (rating 3 or 4) were divided by the total number of experts. The item will be suitable if the I-CVI exceeds 0.79. It needs to be revisited if the I-CVI is between 0.70 and 0.79. It is removed if the I-CVI is less than 0.70. The content validity index for the entire scale (S-CVI) was assessed by the S-CVI with the average approach, by summing all I-CVI for relevancy divided by items number. The scale as a tool was considered to be valid if S-CVI ≥ 0.90 [25].
The construct validity examines the degree to which the Arabic version of HRP scale in line with an established theory or hypothetical construct. The Arabic version of HRP scale construct validity was evaluated by confirmatory factor analysis (CFA). The criteria for conducting the CFA were evaluated using the Kaiser-Meyer-Olkin (KMO) test of sampling adequacy with value > 0.6 and the Bartlett Test of Sphericity with the statistical significance level P < 0.05. CFA was conducted in terms of item loading and goodness of fit estimation using maximum likelihood (ML). Item loading should be greater than 0.30 [27].
Goodness of fit was assessed for two models. The first (model A) was first-order confirmatory factor analysis consisted of the 8 practices and 24 items. The second was a second-order confirmatory factor analysis combining the practice into two bundles (model B). Goodness of fit evaluated using comparative fit index (CFI), the normed fit index (NFI), and root mean square error of approximation (RMSEA), considering the usual values for comparative fit index (CFI) and normed fit index (NFI) ≥ .90 and root mean square error of approximation (RMSEA) < 0.08 [17].
To test the concurrent validity, a correlation coefficient computed between the Arabic version of HRP scale and some external criterion. The valid criterions were Baldrige dimension scale to assess performance excellence, and six-dimension scale to assess nurses' performance (6-D). This is based on the literature’s depiction of an existing relationship between HRP and achieving performance excellence in healthcare organizations [12] and increased level of nurses' performance [28]. The Arabic version of HRP scale considered valid if its scores correlate highly with scores on the criterion. Coefficients of .70 or higher are desirable [29].
Reliability
The reliability of the Arabic version of HRP scale means the scale ability to yield consistent consistency results. To investigate reliability, internal consistency and test-retest were applied. Internal consistency was examined by Cronbach’s alpha reliability coefficients. Cronbach’s alpha value of 0.50–0.70 was acceptable while 0.70 or higher shows good homogeneity among the items [30]. Two weeks' test-retest reliability was conducted with a subsample of 85 nurses and was tested by intra-class correlation coefficient (ICC). Correlation coefficient (r) values are considered good if r ≥ 0.70 [31].
Ethical consideration
- Permission for the translation, adaptation and applying psychometric testing of the scale was obtained from the originators of the scale.
- Nurses’ informed consent was taken after an explanation of the study aim.
- Nurses recruited in this study were confirmed about the confidentiality of the information collected and that they can refuse or withdraw without penalty at any time.
Statistical analysis:
Data were analyzed with SPSS version 24 (IBM corporation, IL, Chicago, USA) and IBM AMOS V.22.0 for CFA. The content validity index was calculated at the item level (I-CVI) and scale level (S-CVI). Pearson’s correlation coefficient (r) was calculated to measure the correlations between the Arabic version of HRP scale performance excellence and nursing performance. CFA with maximum likelihood estimation was used for validation. Goodness of fit was evaluated using comparative fit index (CFI), the normed fit index (NFI), and root mean square error of approximation (RMSEA). Inter-class correlation (spearman r) was used to measure intra-rater correlation. Cronbach's alpha was calculated to measure the internal consistency between items. P ≤0.05 was considered statistically significant.