Development and validation of the Scale for Staff–Family Partnership in Long‐term Care (SSFPLC)

1 INTRODUCTION

With rapid aging of the worldwide population, the older people in need of long-term care are also increasing due to chronic diseases such as dementia and stroke. Although the demand for both service providers and service users has increased owing to the quantitative increase in long-term care facilities, there are difficulties in providing these services to meet the needs of older adult families and to improve the health and quality of life of older adults living in the facilities (McGilton et al., 2016).

Most of the older adult nursing home residents are highly dependent, and as a result, the staff experience excessive physical and mental burnout, lowering their job satisfaction and increasing job turnover (Rajamohan et al., 2019). A previous study identified work overload, inadequate staffing and interpersonal conflicts as the causes of burnout among facility staff (White et al., 2020). Particularly, caregiving is a type of human service involving frequent emotional interactions; thus, emotional work plays a critical part (Yeatts et al., 2018). One of the major interactions required for nursing home staff is that with the families of nursing home residents. Even after admission, families frequently contact nursing home staff and play the role of an advocate and watcher for the admitted older adults (Shippee et al., 2017), during which they may face conflicts with the staff. However, families of nursing home residents may provide important information about the resident's life, habits, preferences and care needs (Reid & Chappell, 2017); thus, family member's participation in care is essential for the residents' well-being (Puurveen et al., 2018). Families of older adults living in the facility serve as a customer and a resource, and they care for the resident. As such, there is a complex interaction among the resident, the older adult's family, and the staff in partnerships for caring the older adult in the facility (Bauer & Nay, 2003).

Since the World Health Organization declared promoting the health of everyone in 1978 as a major goal (World Health Organization, 1978), the concept of partnership has been used as a collaborative relationship between healthcare professionals and clients with a greater focus on the patient's health status and enhancement of health management skills (Gregory et al., 2018). In nursing studies, partnership-related research has been conducted in various aspects, including for meaning and concept analysis of partnerships (Lee, 2007), the development of a partnership model (Coyne & Cowley, 2007; Wiggins, 2008), and the application of parent and family participation programs such as interventions for family involvement care (Mackie et al., 2018).

Despite the increased awareness of and emphasis on the need and importance of partnership, the development of a standardised instrument to assess partnership is at an inchoate stage, and existing tools have some limitations. Measurement scales for partnership developed so far only assess limited concepts, such as treatment alliance (Kim et al., 2001), trust (Jones and Barry, 2011) and relationship (Kiriake & Moriyama, 2016), and studies encompassing the core property of partnership are rare. Jones and Barry (2011) stated that trust is one of the most important factors in the effective functioning of partnerships and developed the trust measurement tool that only measures the trust and mistrust dimension in a partnership. Alliance is often used synonymously with partnerships. The Kim Alliance Scale tool measures the quality of therapeutic alliance in the dimensions of collaboration, integration, empowerment and communication (Kim et al., 2001). However, the evaluation of the tool was performed with a small number of participants; therefore, acceptable validity and reliability could not be shown.

As reported in previous studies, partnership formation leads to family satisfaction with the facilities and a decrease in conflicts with employees (Bidmead & Cowley, 2005; Gallant et al., 2002; Hook, 2006). Effects on employees include increased job satisfaction, reduced conflict and stress and improved quality of care (Bidmead & Cowley, 2005). Moreover, the effects on the older adult living in facilities include maintenance of well-being and health and improvement of quality of life (Dupuis et al., 2016). As such, the partnership between facility staff and their families pursues the common goal of improving the health and quality of life of the older adult, but it is formed through different influences in different contexts (Jang, 2020). In addition, there was a difference in the partnership attributes of facility staff and their families as reported in a previous study (Jang, 2020), and even in one attribute, they had different perspectives on partnership, indicating that the indicators were different. Therefore, to accurately measure the partnership between facility staff and their families, it is necessary to develop a tool that reflects each viewpoint.

The Partnership Care Delivery Model (Wiggins, 2006, 2008) emphasises partnership for patient-centred care and explains that partnership among the patient, family and health care provider has a positive effect on patient safety, quality of care, satisfaction, outcome and job performance. Therefore, in this study, we intended to develop a tool for measuring the partnership between the nursing home staff and families of nursing home residents targeting nursing home staff and verify its reliability and validity.

2 METHODS 2.1 Study design

This is a methodological study aimed to develop and psychometrically test an instrument applicable to nursing home staff to assess their partnership with residents' families.

2.2 Developing the scale

The development and validation of the instrument were performed according to the guidelines proposed by DeVellis (2016) and comprises the following four stages: (1) generation of an item pool, (2) estimation of content validity, (3) a preliminary survey and (4) testing of validity and reliability.

2.2.1 Generation of an item pool

The components of partnership were identified in the author's previous study (Jang, 2020), which analysed the concept of partnership between facility staff and family. The study conducted by Jang (2020) used the hybrid model reported by Schwartz-Barcott and Kim (2000). By integrating theoretical analysis through a systematic literature review with an empirical process that reflects the situation in the field through focus group interview (FGI), the dimension and attributes of the concept were identified.

Based on the two dimensions (interpersonal and environmental dimensions) and seven attributes (relationship, information sharing, shared decision-making, professional competence, negotiation, involvement in care and shared responsibility) as reported in a previous study (Jang, 2020), the components of the item were confirmed, and two of our researchers developed the initial items based on literature and FGI data (Appendix S1).

We developed 32 self-reported preliminary items in Korean. Each item has a 4-point rating scale with responses ranging from 1 (strongly disagree) to 4 (strongly agree). Higher scores indicate a higher level of partnership. To prevent fixed response patterns, reverse coding items were included, and items were rearranged non-consecutively.

2.2.2 Estimation of content validity

Content validity was tested to verify whether each item is appropriate per the operational definition. A panel of experts was invited and it included five nursing professors, three nursing home directors, and two nurses with at least 3 years of employment at a nursing home. The preliminary items were tested for the Item-level Content Validity Index (I-CVI). A ratings of 4 (very relevant) and 3 (relevant) were scored as 1 and the rest were scored as 0. All the preliminary items had I-CVIs exceeding the cut-off of value of 0.78 (Polit et al., 2007).

2.2.3 Preliminary survey

A preliminary survey was performed with 10 staff members each at the facility with ≤29 beds, the one with 30–99 beds, and the one with ≥100 beds, a total of 30 participants were participated. The participants were 2 men and 28 women, the average age was 53.1 years, and average working period was 63.4 months. Eleven were college graduates or had a higher education level.

Participants were asked to respond to the readability, comprehensibility and clarity of the items. It took between 8 and 10 min for them to complete the preliminary survey. There were no problems with readability, comprehensibility, clarity, time required to complete and appropriateness of length. Therefore, the main survey was carried out with 32 items.

2.3 Samples and setting

The participants were nursing home staff. The inclusion criteria were as follows: (1) direct care providers who were involved in the care of older adult nursing home residents, (2) those who consented to participate in this survey. Based on an appropriate sample size of 150–200 for exploratory factor analysis (EFA) (Hinkin, 1998) and a sample size of 150 or more for confirmatory factor analysis (CFA) (Anderson & Gerbing, 1988), the sample size was set at 350. Data were collected from 365 staff working at nursing homes in Seoul, Gyeonggi, Chungnam, Gangwon and Gyeongbuk provinces in Korea. 19 questionnaires were excluded due to missing data; thus, a total of 346 questionnaires were analysed. Samples were randomised to the EFA (n = 173) and CFA (n = 173) using the IBM SPSS/WIN 23.0 (IBM Corp) program feature for random case sampling, as Hinkin (1998) suggested to use different sample sets for EFA and CFA.

2.4 Data collection

Data were collected from June to October 2018. In order to obtain approval and cooperation for data collection, researchers visited the nursing homes and explained the purpose of the study and the method of data collection to the head staff of nursing homes. A self-filled questionnaire was distributed after written informed consent was obtained from staff that were willing to participate in the study. The completed questionnaire was sent to researchers by mail. For the test–retest, an additional survey was conducted two weeks after the initial survey.

2.5 Instrument 2.5.1 Sociodemographic characteristics

Sociodemographic characteristics included age, gender, educational level, perceived economic status, perceived health status, perceived stress status, size of facilities, work position, working experience and satisfaction at current workplace.

2.5.2 Attitude toward family checklist

Criterion validity was tested using the attitudes toward family checklist based on the evidence that staff shows positive attitudes toward families when they have a good collaborative relationship with families (Maas et al., 2004; Park, 2010). This tool measures staff attitudes toward families using three subscales (disruption, family as partners and family relevance). The Cronbach's α was 0.70 in the previous study (Park, 2010) and 0.73 in this study.

2.6 Statistical analysis

Data were analysed using the IBM SPSS/WIN 22.0 and AMOS/WIN 22.0 software. Participants' general characteristics were analysed with descriptive statistics, and differences in the characteristics between the CFA and EFA groups were analysed using χ2 tests and independent two-sample t-tests. Items were analysed for each item score, skewness and kurtosis, and item-total correlation coefficients of ≥0.3 (Field, 2013) were selected. For the EFA, factors were extracted using principal component analysis with Oblimin rotation. The fit indices used for model fitness for the CFA were χ2 (p) (p < .05), normed χ2 (χ2/df) ≤ 3, goodness of fit index (GFI) ≥ 0.80, adjusted GFI (AGFI) ≥ 0.80, normed fit index (NFI) ≥ 0.90, comparative fit index (CFI) ≥ 0.90, root mean square residual (RMR) ≤ 0.05 and root mean squared error of approximation (RMSEA) ≤ 0.10 (Hair et al., 2010). The criteria for convergent validity were as follows: factor loading (FL) ≥ 0.50, critical ration (C.R) ≥ ±1.97 (p < .05), average variance extracted (AVE) ≥ 0.50 and composite construct reliability (CCR) ≥ 0.70. The discriminant validity was tested with AVE > Ф2. (Yu, 2016) For criterion validity, concurrent validity was tested with Pearson's correlation analysis with attitudes toward families. Reliability was tested with item-total correlation (ITC) and Cronbach's α. Test–retest reliability was tested with intra-class correlation coefficient (ICC).

2.7 Ethical consideration

This study was approved by the institutional review board (IRB No. 17–085–1). After informing the participants about the purpose and procedure of the study, a written consent was obtained. The researcher explained about the anonymity of participation, voluntary participation, ability to withdraw and confidentiality during data processing and analysis.

3 RESULTS 3.1 General characteristics of the participants

The mean age was 54.46 ± 9.80 years, and 310 (89.6%) were women. Although 206 (59.6%) perceived themselves to be in good health, 257 (74.3%) perceived themselves to have low level of stress. Regarding the size of workplace, 173 (50.0%) worked in a 30–99 bed facility. The mean length of work experience in the current position was 5.00 ± 5.26 years, and the workplace satisfaction score was 6.51 ± 1.94. There were no significant differences in the general characteristics between two groups (Table 1).

TABLE 1. General characteristics of participants (N = 346) Characteristics Categories Total Group A for EFA (n = 173) Group B for CFA (n = 173) t or χ2 p n (%) or M ± SD Age (years) 54.46 ± 9.80 54.50 ± 9.21 54.40 ± 10.38 0.09 .930 Gender Female 310 (89.6) 160 (92.5) 150 (86.7) 3.10 .056 Male 36 (10.4) 13 (7.5) 23 (13.3) Educational level ≤Middle school 44 (12.7) 18 (10.4) 26 (15.0) 2.01 .367 High school 153 (44.2) 76 (43.9) 77 (44.5) ≥College 149 (43.1) 79 (45.7) 70 (40.5) Perceived economic status Good 16 (4.6) 9 (5.2) 7 (4.0) 0.26 .876 Moderate 282 (81.5) 140 (80.9) 142 (82.1) Poor 48 (13.9) 24 (13.9) 24 (13.9) Perceived health status Good 206 (59.6) 96 (55.4) 110 (63.6) 3.53 .171 Moderate 134 (38.7) 75 (43.4) 59 (34.1) Poor 6 (1.7) 2 (1.2) 4 (2.3) Perceived stress status Low 257 (74.3) 132 (76.3) 125 (72.3) 0.74 .389 High 89 (25.7) 41 (23.7) 48 (27.7) Size of facilities ≤29 beds 77 (22.3) 37 (21.4) 40 (23.1) 2.86 .240 30–99 beds 173 (50.0) 81 (46.8) 92 (53.2) ≥100 beds 96 (27.7) 55 (31.8) 41 (23.7) Position Nurse & assistant nurse 66 (19.1) 35 (20.2) 31 (17.9) 3.68 .159 Healthcare worker 193 (55.8) 88 (50.9) 105 (60.7) Social worker 87 (25.1) 50 (28.9) 37 (21.4) Working experience in current position (year) 5.00 ± 5.26 5.08 ± 5.01 4.93 ± 5.51 0.25 .801 Satisfaction of current workplace (range: 0–10) 6.51 ± 1.94 6.69 ± 1.95 6.33 ± 1.93 1.73 .084 Abbreviations: CFA, confirmatory factor analysis; EFA, exploratory factor analysis. 3.2 Item analysis

Item analysis was performed for 32 preliminary items. Each item was included in the calculation of the mean score, standard deviation, Z-score, skewness and kurtosis value in order to test the appropriateness of the collected data. The skewness (−0.46–0.87) and kurtosis (−0.64–2.27) of each item satisfied the assumption of normality (Yu, 2016). The Z-score was <±3.0 for all items (Yu, 2016). The mean scores for individual items ranged from 2.74 to 3.39 out of a score of 1–4, with a standard deviation of 0.44 to 0.75. In order to examine the contribution rate of the items, item-total correlation coefficients were calculated. After deleting 10 items with an ITC value of below r = |.30| (Field, 2013), 22 items were left in the tool.

3.3 Construct validity

To verify the construct validity, EFA and CFA were performed and convergent validity and discriminant validity were tested.

3.3.1 Exploratory factor analysis

Prior to the EFA, we performed the Kaiser–Meyer–Olkin (KMO) and Bartlett's test of sphericity. The value of KMO was 0.91 and Bartlett's sphericity test value was χ2 = 2759.27 (p < .001), indicating that the sample was appropriate for factor analysis.

Principal component analysis and factor rotation were performed to extract factors. As a result, two items with a commonality ≤0.40 (#3, 5), and one item with an FL value <0.40 in the structure and pattern matrix (#30) (Hair et al., 2010) were deleted. The remaining 19 items were analysed with EFA, and the FL of all items was ≥0.60. The number of factors was set up as four by the scree graph, eigenvalue, explanatory power of factors, and explained cumulative variance. Four factors showed eigenvalues of ≥1.0. There were four significant factors shown as elbow points on the scree graph (Appendix S2). Furthermore, the explanatory power of the factors ranged from 21.3% to 24.6%, and the explained cumulative variance of factors was 91.7% (Table 2).

TABLE 2. Item analysis and factor analysis of SSFPLC (N = 346) Factor/Item contents M ± SD Factor loadings Communality Explained Variance (%) ITCa Cronbach's α if item deleteda Cronbach's αa ICC (95%CI) (n = 30) a 1 2 3 4 Factor 1 – Encouragement to participate in care 27. I encourage the families to visit the facility. 3.05 ± 0.63 0.79 0.04 −0.03 0.01 0.68 24.6 0.59 0.76 0.82 0.85 (0.68–0.93) 28. I positively support family involvement in providing care (e.g. conversation, taking a walk, meal assistance, etc.). 3.23 ± 0.55 0.75 0.13 0.07 0.11 0.68 0.62 0.76 26. I inform the families about the condition or changes in the condition of the older adults residing in the facility. 3.16 ± 0.59 0.63 −0.07 −0.26 0.09 0.62 0.62 0.80 29. I welcome the families when they visit the facility. 3.27 ± 0.60 0.59 0.26 0.04 0.16 0.60 0.64 0.79 9. I think families and facility staff are responsible for the care of the elderly residing in the facility. 3.23 ± 0.59 0.53 −0.05 0.01 .19 0.39 Factor 2 –Family's trust and support 2. Families abide by the rules and the policies of the facility well. 2.88 ± 0.57 −0.06 0.84 −0.01 −0.03 0.67 21.3 0.44 0.74 0.79 0.52 (0.01–0.77) 1. Families are reassured about the life of the older adults residing in the facility. 3.14 ± 0.51 0.09 0.72 0.18 0.13 .55 0.46 0.77 4. Families are grateful for my care for the older adults residing in the facility. 3.11 ± 0.55 −.13 .66 −.18 .06 .54 0.47 0.76 7. Families trust the information provided by the facility staff for their decision-making. 3.07 ± 0.54 0.21 0.64 −0.10 −0.08 0.57 0.56 0.75 6. Families actively participate when I (the facility staff) ask for cooperation regarding the older adults residing in the facility. 3.04 ± 0.61 0.09 0.62 −0.22 −0.13 0.53 0.49 0.76 Factor 3 – Collaborative relationship and communication 17. Families and I cooperate with each other in caring for the older adults residing in the facility. 3.02 ± 0.52 0.02 0.09 −0.73 0.11 0.67 23.1 0.63 0.76 0.82 0.69 (0.34–0.85) 16. Families and I communicate smoothly regarding caring for the older adults. 2.92 ± 0.58 0.12 0.08 −0.70 0.03 0.64 0.61 0.78 18. Families and I share a common goal in caring for the older adults residing in the facility. 2.98 ± 0.54 −0.10 0.10 −0.68 0.22 0.61 0.58 0.78 21. Families and I respect each other's knowledge and experience with regard to caring for the older adults residing in the facility. 2.99 ± 0.54 −0.03 0.19 −0.63 .15 0.60 0.62 0.79 32. I involve the families when planning care for the older adults residing in the facility. 2.68 ± 0.62 0.36 −0.07 −0.48 −0.14 0.40 Factor 4 – Professional care 13. I provide appropriate care on the condition of the older adults residing in the facility. 3.20 ± 0.44 0.12 0.02 0.01 0.78 0.71 22.7 0.58 0.75 0.82 0.70 (.38-.86) 31. I am sensitive to changes in the state of the older adults residing in the facility. 3.16 ± 0.59 0.10 0.02 0.08 0.75 0.61 0.48 0.82 23. I encourage the older adults residing in the facility to eat or exercise by themselves regularly as much as possible. 3.27 ± 0.48 0.04 −0.08 −0.23 0.73 0.69 0.59 0.79 25. I provide care while maintaining the dignity of the older adults residing in the facility. 3.25 ± 0.45 0.04 0.03 −.018 0.71 0.65 0.61 0.78 Total 3.09 ± 0.55 91.7 .90 .96 (.91-.98) KMO = 0.91, Bartlett's test: χ2 = 2759.27 (p < .001) Abbreviations: ICC, intra-class correlation coefficient; ITC, item-total correlation. aResults excluding items 9 and 32. 3.4 Confirmatory factor analysis

Confirmatory factor analysis was conducted to test the construct validity for the 19 items under four factors identified through EFA. The factors were named encouragement to participate in care (factor 1), family's trust and support (factor 2), collaborative relationship and communication (factor 3), and professional care (factor 4). We checked whether the items have a standardised FL of ≥0.50 and significance (C.R.) of ≥±1.97 (p < .05), and items 9 and 32 did not meet the criteria and were deleted. The model fit for 17–items were χ² = 186.25 (

留言 (0)

沒有登入
gif