Collaboration and Satisfaction About Care Decisions in Team questionnaire—Psychometric testing of the Norwegian version, and hospital healthcare personnel perceptions across hospital units

Abstract Aim To translate “The Collaboration and Satisfaction About Care Decisions in Team” questionnaire (CSACD‐T) into Norwegian and test it for psychometric properties. The further aim was to describe and compare healthcare personnel's collaboration and satisfaction about team decision‐making (TDM) across hospital units. Design A cross‐sectional study. Methods The questionnaire was translated into Norwegian. A total of 247 healthcare personnel at two hospitals responded to the questionnaire. An explorative factor analysis was performed to test the factor structure of the questionnaire, while a Cronbach's alpha analysis was used to test for internal consistency. A one‐way ANOVA analysis and a Kruskal–Wallis test were applied to test for differences between hospital units. Results The results demonstrate that the Norwegian version of the CSACD‐T has promising psychometric properties regarding construct validity and internal consistency. The mean score of the CSACD‐T was significantly higher in the maternity ward group than in the emergency room group.

people who interact dynamically, interdependently and adaptively towards a common and valued goal, who have each been assigned specific roles or functions to perform and who have a limited life span membership" (Salas, Dickinson, Converce, & Tannenbaum, 1992, p. 4).
Decision-making and problem-solving are important parts of everyday practice for hospital healthcare personnel, including physicians, nurses and allied healthcare personnel (Levenson, 2010), with all professions being members of interprofessional teams from time to time (Weinberg, Cooney-Miner, Perloff, Babington, & Avgar, 2011). Although physicians play an essential role in patient treatment decisions (Farnan, Johnson, Meltzer, Humphrey, & Arora, 2008;Levenson, 2010), nurses and allied healthcare personnel hold patient information that is important in planning, managing and making decisions about patient care and should therefore be involved in the decision-making process (Marshall, West, & Aitken, 2011). TDM involves both the group which shares information and the team leader who integrates the information and makes a final decision. TDM is important in hospital units, which are often characterized by rapidly changing environments and time pressures (Lipshitz, Klein, Orasanu, & Salas, 2001;Reader, 2017).
Previous research on decision-making in health care has focused on physicians making decisions about patient treatment (Farnan et al., 2008;Hausmann, Zulian, Battegay, & Zimmerli, 2016), whereas studies on interprofessional collaboration in decision-making having mostly been about nurse-physician collaboration (DeKeyser Ganz, Engelberg, Torres, & Curtis, 2016;Maxson et al., 2011;Nathanson et al., 2011). Physicians report the most positive perceptions of collaborating in a team; in contrast, nurses are often less satisfied. A limited number of studies have investigated decision-making in larger hospital teams, beyond nurse-physician teams (Lancaster, Kolakowsky-Hayner, Kovacich, & Greer-Williams, 2015;Zwarenstein, Rice, Gotlib-Conn, Kenaszchuk, & Reeves, 2013). The results of these studies show that care decisions most often take place in isolation by physicians and that the decisions are rarely made collectively.
Previous research of multi-professional TDM across different hospital units has been limited.
Multiple instruments have been developed to measure collaboration in teams (Valentine, Nembhard, & Edmondson, 2015), but not many specific for measuring TDM. The "Collaboration and Satisfaction About Care Decisions in Teams" (CSACD-T) was designed to measure healthcare personnel's perceptions of collaboration and satisfaction with decision-making in healthcare teams and is based on the original "Collaboration and Satisfaction about Care Decisions" (CSACD) questionnaire (Baggs, 1994). The original CSACD was developed to measure collaborations between nurses and physicians (Baggs, 1994) and has been used in multiple studies Maxson et al., 2011;Nathanson et al., 2011;Papathanassoglou et al., 2012) and has also been linked to patient outcomes (Baggs et al., 1999;Boev & Xia, 2015). When the original CSACD was tested for psychometric properties, an EFA was applied only on the first six items of the questionnaire and found the six items to be one factor (Baggs, 1994). When the team version (CSACD-T) was developed, only minor changes were made in the items with different wordings to capture a broader healthcare team (Fox & Heineman, 2002). The team version of the nine-item questionnaire (CSACD-T) has not previously been psychometrically tested.
The theory base of the original CSACD questionnaire was drawn from the work of Thomas (1976), which described a model of collaboration and coordination in complex organizations. According to that model, collaboration is necessary in situations where two or more persons have common interests and the stakes are high. In complex organizations, regarded as dynamic and unpredictable, collaborative solutions provide maximum satisfaction for all parties concerned (Thomas, 1976). Baggs and Schmitt (1988) broadened that model to cover collaborations in decision-making in health care. Through an extensive literature review, they identified five critical attributes of collaboration: assertiveness, planning, shared decision-making, open communication and coordination (Baggs, 1994).
An important aspect of TDM in health care is patient participation in decision-making. The involvement of the patient as a member of the healthcare team is increasingly recognized as a key component of healthcare processes and is advocated as a means to improve patient results and patient safety (Epstein & Gramling, 2013;Longtin et al., 2010;WHO, 2013). An extra item was therefore added to the survey for this study.
We did not find a Norwegian questionnaire that measures healthcare personnel's perception of collaboration in decision-making in teams. Because CSACD-T is a brief and simple questionnaire and not profession-specific, we chose this questionnaire for our study. To the best of our knowledge, no previous studies have investigated TDM among multi-professional healthcare personnel teams across different hospital units. The aim of the study was to translate "The Collaboration and Satisfaction About Care Decisions in Teams" questionnaire into Norwegian and test it for psychometric properties. The further aim was to describe and compare healthcare personnel's perceptions of collaboration and satisfaction about team decision-making across hospital units.

| Design
This study was designed as a cross-sectional study. The nine-item CSACD-T questionnaire was translated into Norwegian according to a translation-back-translation procedure (Brislin, 1970), with further details given in Section 3.4. It was then distributed as a survey to test its psychometric properties.

| Setting and sample
In total, 624 healthcare personnel (registered nurses, postgraduate nurses, midwives, occupational therapists, physical therapists, assistant nurses and physicians) were invited to participate in the study. The respondents were from two hospitals in two different hospital trusts in Eastern Norway: 436 from hospital A (110 beds) and 188 from hospital B (167 beds). The healthcare personnel from hospital A (N = 436) were from the emergency room (ER), intensive care unit (ICU), operating room (OR)/anaesthesia unit (AN), maternity ward and the medical/surgical (med./surg.) wards, whereas the healthcare personnel from hospital B (N = 188) were from medical wards only. All healthcare personnel from the included units were invited to participate in the study. A total of 247 healthcare personnel from the two hospitals responded to the survey.

| The CSACD-T questionnaire
The nine-item CSACD-T questionnaire has response options on a Likert scale ranging from 1-7. The first six items measure attributes of collaboration, ranging from 1 (strongly disagree) -7 (strongly agree). The seventh item measures the level of global collaboration and ranges from 1 (no collaboration) -7 (complete collaboration).
The last two items consider satisfaction with decisions and have response options ranging from 1 (not satisfied) to 7 (very satisfied; Baggs, 1994). The CSACD-T questionnaire was obtained from its creator, Professor Judith Baggs, and permission was obtained to translate the questionnaire into Norwegian.
Due to the importance of patient participation in care decisions, an extra item (item 10) was developed and added by the research group: "Do patients participate in decision-making relating to their own care?" The response options ranged from 1 (no participation at all) -7 (complete participation). This item was not included in the psychometric testing of the questionnaire. The study included the following background data: sex, age, profession, unit type and time employed in the unit.

| The translation process
The Brislin Model (Brislin, 1970)  3. Back-translation-A bilingual professional translator, who was blinded to the original English version, back-translated the Norwegian version into English.

4.
Compare-The research group compared the back-translated version with the original version and found no differences in meaning.
Thus, the Norwegian questionnaire was approved for pilot testing.

| Data collection
The survey was distributed as a paper version in November 2015. Two e-mail-based reminders were administered, with the assistance of managers, during a data collection period of 3 weeks. Completed surveys were sealed in return envelopes and placed in boxes in the units.

| Data analysis
Data analyses were performed with SPSS version 24 (IBM). An explorative factor analysis (EFA) was conducted to test the factor structure of the questionnaire. The aim of the EFA was to test the factor structure of the group of items. In addition to testing the factor structure of the total questionnaire (items 1-9), an EFA was used to analyse items 1-6, due to the intension of comparing our results to those of the original CSACD (Baggs, 1994). Prior to the EFA, we assessed the suitability of our data for factor analysis.
This included a correlation matrix for displaying the relationships between the items, as well as correlation coefficients between 0.30-0.70 considered significant (Polit & Beck, 2017;Tabachnick & Fidell, 2013). A Kaiser-Meyer-Olkin (KMO) test was performed to measure sample adequacy. Within the KMO range of 0-1, a value of 0.60 and above was considered suitable for EFA (Pett, Lackey, & Sullivan, 2003;Tabachnick & Fidell, 2013). A principal component analysis (PCA) was chosen for factor extraction. We applied the "eigenvalue rule," which only allows items of 1.0 or more to be retained for further investigation (Polit & Yang, 2016). A Cronbach's alpha was performed for items 1-9 to check for internal consistency and for items 1-6 to also compare with the original CSACD. A Cronbach's alpha for each item removed was calculated for items 1-9 (Pett et al., 2003).
Descriptive statistics were used to describe the characteristics of the sample and to analyse the results of the CSACD-T scores. A between-group one-way ANOVA, with a Tukey post hoc test, was conducted to compare for differences between unit groups on the total mean score of the healthcare personnel's perceptions of TDM (Polit & Beck, 2017). A Kruskal-Wallis test was conducted to test for differences between groups regarding item 10 (patient participation in decision-making; Polit & Beck, 2017). A two-tailed significance level of p-value <0.05 was used for all tests (Polit & Beck, 2017).

| Translation and psychometric testing of the CSACD-T questionnaire
The total of 247 healthcare personnel that responded to the survey represented an overall response rate of 40%. Table 1 shows the distribution of response per healthcare profession.
Among these individuals, 156 (response rate 36%) were at hospital A and 91 (response rate 48%) at hospital B. Characteristics of the sample are shown in Table 2. Registered nurses constituted most respondents, followed by postgraduate nurses and then midwives.
Most of the respondents were female, with 75% of the sample from hospital wards (maternity ward and med./surg. wards). Leaving more than 50% of items blank, two respondents were excluded from further analysis.
Inter-item correlations ranged from 0.45-0.81 for the total questionnaire (items 1-9) and from 0.51-0.81 for items 1-6. All correlations were significant (<0.001). The KMO values were 0.93 for the total questionnaire (items 1-9) and 0.89 for items 1-6. The PCA identified one factor for the total questionnaire (item 1-9), with an eigenvalue of 6.154 on one factor, explaining 68% of the variance and eigenvalues <1.0 on the remaining factor solutions.
The PCA on item 1-6 also identified one factor for these items, with an eigenvalue of 4.294 on one factor, explaining 72% of the variance and with eigenvalues <1.0 on the remaining factor solutions (Table 3). PCA factor loadings for the total questionnaire (item 1-9) ranged from 0.72-0.87 and factor loadings for items 1-6 ranged from 0.77-0.89 (Table 4). The Cronbach's alpha was 0.94 for items 1-9 and was not improved when each item was removed (ranged from 0.93-0.94). The Cronbach's alpha value for items 1-6 was 0.92.

| Collaboration in team decision-making across different hospital units
The entire sample of healthcare personnel's perceptions of TDM showed a total mean score of 5.14 (SD = 0.95), as measured by the CSACD-T (Table 5). Single-item mean scores for the total sample ranged from 4.82 (SD 1.24) ("Coordination of decision-making among team members") to 5.35 (SD 1.13) ("Shared responsibilities for decision-making") (see Table 5). The added item (item 10), which was healthcare personnel's perceptions of "patient participation in decision-making," had a mean score of 4.63 (SD 1.25) in the total sample ( Table 5).
The results of the one-way ANOVA revealed statistically significant differences in the total mean score of the CSACD-T across unit groups (F(4, 240) = 4.1, p = 0.003). The effect size, calculated using eta squared, was 0.06 (medium effect). Post hoc comparisons, using the Tukey post hoc test, showed a significantly higher score in the maternity group than in the ER group (p = 0.001). A Kruskal-Wallis test of item 10 revealed a statistically significant difference between the unit groups (χ 2 (4) = 11.77, p = 0.001) with the highest score in the maternity ward group and the lowest score in the ER group.

| Translation and psychometric testing of the CSACD-T questionnaire
The purpose of translating a questionnaire is to obtain an instrument in a new language that is equivalent to the instrument in the original language (Sousa & Rojjanasrirat, 2011). The translationback-translation method used in this study is recommended as a reliable method for translating research instruments (Brislin, 1986;Jones, Lee, Phillips, Zhang, & Jaceldo, 2001). The translation process was thorough; both medical and nursing professionals participated to assure the content and cross-cultural validity of the questionnaire (Polit & Yang, 2016). No major problems occurred in the translation or back-translations steps. The respondents in the pilot study provided adequate responses to the items, which suggested that the translated questionnaire was well understood.
The results of the EFA showed one factor for all nine items, and all loadings were above 0.40, which was considered good and the structural validity was thereby supported (Polit & Yang, 2016).
Only the first six items were tested and found to be a one-factor scale in the validation study of the original CSACD (Baggs, 1994).
However, when Sapnas, Ward-Presson, and Monzeglio (2006) conducted a psychometric test to evaluate the original CSACD in diverse hospital unit types, their EFA identified one factor for all nine items of the CSACD questionnaire when tested in diverse hospital unit types, as we found in our study. In any case, since the EFA can only identify clusters of tests that measure the same things, there is no assurance that these "same things" are primary dimensions. Consequently, a factor analysis alone is insufficient to confirm that a factor corresponded directly to the "real" dimension of the construct measured (Polit & Yang, 2016). The interpretation of the results is just as important, a good PCA must make sense (Polit & Yang, 2016 included satisfaction with the decisions as an important part of the collaboration in decision-making (Baggs, 1994). Nonetheless, if the satisfaction part of the questionnaire had included more than two items, the EFA might have resulted in a two-factor solution.
The CSACD-T had a Cronbach's alpha value above the desirable 0.80 (Polit & Beck, 2017) and demonstrated a good internal consistency. Result from the previous study of the nine items CSACD showed alpha value over 0.90 (Sapnas et al., 2006). Tavakol and Dennick (2011) argue that the maximum value to be recommended is 0.90, which means that the alpha value of item 1-9 was maybe too high.
Regarding sample size, the recommended sample size for EFA in validation studies is disputed and no consensus exists (Polit & Yang, 2016). Some suggest a minimum of 300, but emphasize that if there is strong correlations and few distinct factors, a smaller sample is adequate (Tabachnick & Fidell, 2013). Others offer guidance on the number of respondents per items, ranging from 5-40 or 50 per item, with the most common recommendation as a minimum of 10 cases per item (Polit & Yang, 2016

| Collaboration in team decision-making across different hospital units
After translating and testing the psychometric properties of the Norwegian version of the CSACD-T, we aimed to describe and compare healthcare professional's perceptions of collaboration and satisfaction with TDM across different hospital units. The results showed that the mean CSACD-T scores were at the same level or slightly higher than those reported in previous studies of nurse-physician collaboration in decision-making in ICUs (Nathanson et al., 2011;Papathanassoglou et al., 2012) and in paediatric teams (Jankouskas et al., 2007), as measured by the original CSACD. The explanation for why the healthcare personnel from the maternity ward reported a significantly higher score than the healthcare personnel in the ER may be due to having a more team-based approach to their work (Gregory et al., 2017). Nonetheless, many factors influence TDM such as the amount of work load, time stress and culture (Gregory et al., 2017), as in other types of hospital units.
The lower score in the ER group might be explained by more inefficient teamwork, which may be due to the way clinical work is organized for the sub-acute patients in ERs in Norway. A nurse or a physician triages the patients and then assigns the patient to a dedicated nurse. By the time the patient's physician arrives, the dedicated nurse has often moved on to attend to the next patient (Krogstad, ERs as in other hospital units (Reader, 2017).
The added item "patient participation in decision-making" had the lowest mean score in the ER group. Although it is well-known that patients should be included in care decisions, they are still not always included in practice (Williams, Fleming, & Doubleday, 2017).
They should be included because they might have valuable information about their own health condition (WHO, 2013). Patient participation may contribute to help healthcare personnel make the right decisions and to minimize decision errors (Zavala et al., 2018).
Although healthcare personnel are striving for patient participation in decision-making to increase the quality of care and patient-centred outcomes in the ER (Grudzen, Anderson, Carpenter, & Hess, 2016), many patients cannot participate in decisions because of their critical condition (Joseph- Williams, Elwyn, & Edwards, 2014). In addition, decision-making in the ER is complex with rapid assessments and decisions to make and includes transferring patients to the next level of care. But for the healthcare team, TDM is possible, as well as being a contribution to safe care, also in the context of the ER (Reader, 2017).

| Limitations
Some limitation of the study must be stated. The response rate from physicians was low, which is a common problem in research on healthcare personnel in hospitals (Cunningham et al., 2015).
Furthermore, the participants from hospital B were only from medical wards, and except for the med./surg. ward group, unit groups were relatively small. Although this study displayed a relatively low overall response rate, the sample size of 245 respondents was sufficiently large to conduct a factor analysis of the nine-item questionnaire (Polit & Yang, 2016).

| CON CLUS ION
The results of the study demonstrate that the Norwegian version of the CSACD-T is a questionnaire with promising psychometric properties regarding construct validity and internal consistency. The CSACD-T questionnaire can be used in assessing collaboration and satisfaction with TDM in hospital teams, in quality improvement, in continuing education endeavours for healthcare personnel and in research. Moreover, the results showed that the levels of collaboration in care decisions in healthcare teams varied across hospital units, with significantly higher scores in the maternity ward group than in the ER group. Further studies are needed with representative samples from diverse hospital units. Additional studies are also needed to confirm our results of the psychometric testing of the questionnaire.

ACK N OWLED G EM ENTS
We would like to extend our sincere appreciation to the healthcare personnel who participated in the study. We also want to thank statistician, Jari Appelgren, Karlstad University, Sweden, for consulting with us on the data analysis.

CO N FLI C T O F I NTE R E S T
All authors declare no conflict of interest.

AUTH O R CO NTR I B UTI O N
All the authors made substantial contributions to the conception and design of the study, the analysis and interpretation of data, and the drafting of the manuscript. All the authors critically revised the manuscript for important intellectual content, and all have given their final approval of the version to be published.