Bottom Upside Down: Professionals in Inter-Organizational Partnerships in Primary Care

Abstract

In primary health care, with a wide range of different disciplines, added value is seen in working together with other disciplines, especially in the management of chronic conditions. Therefore, a trend is observed toward collaboration between small mono-disciplinary practices, so-called inter-organizational partnerships. This chapter focuses on the role of primary care professionals in 69 inter-organizational partnerships in the Netherlands and the relationship with quality of care and project success. While collaboration in primary care is often initiated by managers and policymakers, our study shows that the advantages and disadvantages as perceived by the executing professionals are important. Their perceptions, in relation to the type of project, are important for the success of the inter-organizational partnerships and the improvement in the quality of care delivered.

1. Introduction

For years, in many high-income countries, primary care was provided in small mono-disciplinary practices [1, 2]. Nowadays, in primary care, with a wide range of different disciplines, added value is seen in working together with other disciplines, especially in the management of chronic conditions that require help from different professionals [3, 4, 5]. Therefore, a trend has been observed toward increasing the scale of primary care organizations. Key examples are the establishment of larger mono-disciplinary group practices, multidisciplinary health centers, and care groups for disease management [6, 7]. A further and parallel development that has occurred next to the growth of these larger primary care organizations is the collaboration between small mono-disciplinary practices, the so-called inter-organizational partnerships. Since the 1990s, the development of inter-organizational partnerships is seen in primary care [8].

Inter-organizational partnerships are increasingly created to tackle complex problems [9, 10]. They are seen as a way to share resources, and perhaps more importantly to facilitate knowledge transfer [11, 12]. Also, in health care they are used to overcome the silos that emerge from highly specialized professions and organizations [13].

In the literature, three main drivers for the development of inter-organizational partnerships in primary care are seen [8, 14]. The first driver is the increasing demand on primary care, because of demographic trends such as living longer in combination with living on your own, and the increase of highly complex patients who have many different needs, e.g., because of a combination of mental and physical problems [13, 15]. Second, there is a shift in treatments from hospital care and nursing home care to primary care, promoted by governments as a response to problems with access, quality, and continuity of services [8, 16]. And third, collaboration is increasingly seen as efficient to control the growth of governmental budgets for health care in many countries [14]. Therefore, inter-organizational partnerships in primary care have been promoted and often subsidized by governments as a means to improve health care.

Despite the expected positive results, studies show that collaboration remains low and provide ambiguous results [17, 18]. In literature reviews, only few inter-organizational partnerships in care generate positive results [19, 20].

Looking at the expected results, such as higher quality of care for complex patients, effects seem to specifically depend on collaboration at the professional level. That is, professionals are the key players in delivering care as they actually see patients [21]. However, this role of the professional in inter-organizational partnerships is often forgotten or not well understood [5, 19, 22].

When it comes to inter-organizational partnerships, small qualitative case studies are most commonly published [19]. Larger and quantitative research on inter-organizational partnerships and the role of professionals is lacking [23], especially when it comes to the role of the professional over time and their influence on the outcomes of the partnership [19]. Therefore, broader research studying professionals in inter-organizational partnerships, and their influence on quality of care, is needed and will contribute to the literature.

In investigating the role of professionals, literature refers to different enablers:

• Personal commitment to the collaboration is crucial. If professionals are more involved in the collaboration, outcomes for patients are better [13, 22]. Their involvement in the collaboration is therefore important.

• A sense of urgency to collaborate among professionals reflects that they are more willing to invest time and effort and thus the project is more likely to succeed [24].

• Professionals are motivated to participate and add value to a project if they see advantages in the project [25]. Vice versa, if professionals mainly see disadvantages in a project, they will be demotivated to contribute. This especially applies in primary care practices where work pressures are high; when professionals are not convinced of benefits for patients on the quality of care, a collaboration will not become a success [22, 26].

In short, the ambiguous and limited results of partnership projects in primary care can be due to a number of factors as mentioned before, i.e.: professionals not being involved, lack of sense of urgency, or not seeing advantages/only seeing disadvantages [19, 21, 22, 26]. Interesting in this respect is the change in the quality of care when partnerships are created, and the way this motivates or demotivates the professional in the partnership over time.

Literature on inter-organizational partnerships in primary care points out at management or leads and their role in creating the necessary conditions for the collaboration between professionals [27]. Little research is done on both the level of the project lead and professionals. Investigating both, therefore, contributes to the literature and a theoretical understanding on inter-organizational partnerships in primary care.

In this chapter, we present the results of a multi-project, multi-actor, and multi-level research, to investigate the role of the professional as key to the success of inter-organizational collaboration in primary care. The main aim of our research is to disentangle to what extent the experiences and perceptions of professionals determine the outcomes of collaboration projects. This is done by researching how professionals perceived and experienced the project, and if they perceived as to how the project has actually improved the quality of care.

The first questions we address are:

1. How do the sense of urgency, involvement, and (dis-)advantages—as perceived by professionals in the inter-organizational collaboration—relate to the change in the quality of care they perceived between the beginning and the end of the project? And is this relationship influenced by taking into account basic characteristics of the professionals and the type of project?

   Next, we investigate if the professionals’ perceptions are also related to the final success of the project, but now measured from a project management perspective:

2. To what extent do the sense of urgency, involvement, and (dis-) advantages—as perceived by professionals in the inter-organizational collaboration—relate to the final success of the collaboration as perceived by the project leaders? And is this relationship influenced by taking into account basic characteristics of the professionals and the type of project?

   To, finally, elaborate on the third question:

3. To what extent is the final success of the project as perceived by the project leaders related to factors at the professional level on the one hand, and project factors on the other?

2. Methods

To study collaboration in depth, longitudinal data were collected by interviews with project leaders and survey among professionals in 69 inter-organizational partnership projects in Dutch primary care (see Box 1). These multi-level and multi-actor data provided the opportunity to analyze multiple projects, thus enabling us to study the determinants of project success from different perspectives and over time.

An opportunity to investigate the role of healthcare professionals within a primary care collaboration was created by the Primary Focus program of the Netherlands Organization for Health Research and Development (ZonMw). The Primary Focus program (2010–2014) funded 69 projects in primary health care. The purpose of the program was: to strengthen the organization of collaboration in primary care in the Netherlands. The projects in the program varied in type, size, subject, and expected outcomes. They were accompanied by research to gain insights into the critical success factors that hamper or facilitate collaboration. The data of this research are used in this chapter.

Box 1.

The primary focus program and its projects.

Two data types were combined in the analysis. At the project level (N = 69), data were collected by interviews with multiple project leaders and project employees per project—at the beginning of the project, roughly halfway, and at the end of the funding of the project. This resulted in 621 interviews. The interview results were used to classify structure and type of the project at the beginning and measure the perceived success at the end of the project.

At the professional level, per project, survey data were collected from among the professionals who were involved and actually delivered care to the patients. The surveys were based on earlier research and validated questions [28, 29]. Questionnaires were sent to the professionals at the beginning and end of the project (T1 and T2). The questionnaire contained 20 questions and was kept short to encourage professionals to complete the questionnaire and taking the least possible time. Not for all 69 projects was the questionnaire for professionals conducted. Four projects that were prematurely terminated were excluded, as the T2 survey could not be sent out.

2.1 Measurements

Figure 1 shows an overview of the variables and levels included in the analyses to address the three research questions, as presented in the previous section.

Figure 1.

The variables and levels used in the analyses.

The change in the quality of care—due to the inter-organizational collaboration project and as perceived by the professionals—–is the dependent variable in the first model. The perceptions of professionals at T1 and T2 are the independent variables, while the professionals’ gender and contract, as well as two characteristics of the project, are included in model 1 as control variables.

In the second model, the project success according to project leaders at T2 is the dependent variable. Here (as in model 1), the perceptions of professionals at T1 and T2 are the independent variables, while their gender and contract, as well the type of project are included as control variables. In addition, the dependent variable of model 1 (change in the quality of care as perceived by the professionals) is included as an explanatory (independent) variable.

The two levels (i.e., 473 individual professionals (1) and 69 projects (2)) are also demonstrated in Figure 1. Because all 473 professionals were participating in the 69 projects, the third research question can be answered, analyzing the influence of predictors at the professional and project levels as dependent variables. More information about all variables can be found in Appendix A.

2.2 The dependent variables

The two dependent variables in the analyses are measured on two different levels, namely “the change in quality of care due to the project according to professionals” (model 1) and “the success of the project according to project leaders” (model 2).

2.2.1 Change in the quality of care

The change in the quality of care was measured by asking professionals in the surveys: ‘Do you think the quality of care was increased, decreased, or remained the same due to the collaboration project?’. This (perceived) change was measured just after the start (T1) and just before the end (T2) of the funding of the project. Because of a skewed, more positive distribution of the answers, the variables were changed into a dichotomy scale: professionals answering ‘decreased or stable’ and professionals answering ‘increased’.

2.2.2 The success of the project

The success of the project according to project leaders was rated by interviewers who were commissioned to perform an independent monitoring and evaluation of the Primary Focus program. At T2, the interviewers addressed questions with the project leaders and multiple project members were: “Are the intended results of the project achieved?” and “Which parts of the project are successfully achieved?”. The interviewers then assessed the final project success by scoring on a 5-point Likert scale, from: ‘very successful, successful, neutral, unsuccessful, not successful at all’. Because none of the projects were rated as unsuccessful or not successful at all, the answers were converted into two categories ‘neutral or unsuccessful’ and ‘successful’.

2.3 The independent variables

The sense of urgency, involvement, (dis)advantages, and quality of care change due to the inter-organizational collaboration project, as perceived by the professionals, could obviously only be measured if professionals were at least aware of their participation in a collaboration project. This implied that a proportion of the professionals were not directed to these questions in the survey, either at T1 or at T2.

2.3.1 The sense of urgency

The sense of urgency of inter-organizational collaboration project was measured at T1 and T2 among those professionals who were aware of the collaboration. It was measured by the question: “Are you convinced of the importance of the project?” It was measured by a 5-point Likert scale, ranging from ‘totally unconvinced’ to ‘totally convinced’.

2.3.2 The involvement

The involvement of the professionals in the inter-organizational collaboration project was measured at T1 and T2 by the question: “In what way are you involved in project?” The three possible answers were: ‘only involved in delivering care’, ‘also in the organization’, and ‘also as an initiator of the project’.

2.3.3 The (dis)advantages

The respondents answered 27 predefined advantages and disadvantages of the inter-organizational collaboration project at T1 and T2 (see Appendix B). Each item was presented on a 5-point Likert scale: ‘totally disagree’ to ‘totally agree’. To aggregate these measurements of the perceived advantages and disadvantages, they were subjected to a principal component analysis. The principal component analysis revealed the presence of two components. All the 14 items formulated as experienced advantages fitted into one scale, and the 13 items that experienced disadvantages could also be joined together. The analysis can be found in Appendix B. The two components are thus included as two different (independent) variables in the analyses.

2.4 The control variables

2.4.1 The characteristics of professionals

The professionals’ gender and type of contract were included in models 1 and 2 to control for characteristics of the professional. The two categories for type of contract were ‘self-employed’ or ‘salaried’. Both gender and type of contract were derived from background questions that were included in the survey among professionals at T1.

2.4.2 The type of project

Two variables that classify the type of project were used as control variables in the analyses, i.e., the type of patient population involved and the type of collaboration. The type of patient population involved was coded by two categories, whether the population was ‘disease/diagnose specific’ or ‘generic’. The second variable ‘type of collaboration’ was coded into two categories: ‘collaboration within primary care’ and ‘collaboration between primary care and other sectors e.g. welfare, or mental health’. These variables were derived from the interviews with the project leaders at T1.

2.5 Analyses

Data were analyzed using Stata 13.1 and MLwiN version 2.30. The differences between T1 and T2 were valued by a Student’s t-test. For the testing of the outcomes, a multilevel model was used to account for the hierarchical structure of the data. The data were structured at two levels, because the units of analysis were at the level of the professionals (n = 474 at T1 and n = 395 at T2), nested within the level of the project (n = 69). A binominal logistic model (second-order professional quality of care (PQL) for quality of care and first-order Monitoring Query Language (MQL) for project success) was used to account for dummy-dependent variables. A significance level of p < 0.05 was used for all statistical tests.

3. Results

3.1 Descriptive results

The number of responding professionals per project varied between 3 and 148. The mean response was 64% at T1 and 46% at T2, with a range from 16 to 100% among the different projects. For 65 projects, a success score was constructed, in terms of the changes in the quality of care according to professionals (for model 1), and project success as perceived by the project leaders and classified by the project interviewers (for model 2).

3.1.1 The characteristics of professionals and projects

Most of the responding professionals were female (73%). Almost 40% of the respondents was self-employed. Eighty-one percent of the projects focused on a specific patient population. The goal of these projects was to improve care, e.g., for elderly, patients with chronic obstructive pulmonary disease (COPD), or pregnant women. Half of the projects were concerned with inter-organizational collaboration within primary care, whereas the other half of the projects were aimed to initiate or improve collaboration between primary care organizations and organizations in at least one other sector.

3.1.2 The role of professionals

At the beginning of the project, 21% of the professionals was unaware of the collaboration. At the end of the project, more than 15% of the respondents was still unaware that they were involved in a project. There were no projects where none of the respondents was aware; however, the percentage of professionals who were aware of the project at T2 ranged from 19 to 100%. The professionals who were unaware of the projects were excluded from further analysis. Descriptives of the variables can be found in Appendix, Table 4.

As shown in Table 1, the mean scores on involvement and urgency are both lower at T2 than T1, but the difference is not significant. The mean score on advantages increased significantly over time (p < .05). The score on disadvantages also increased, but not significantly. A correlation matrix is presented in Appendix C, Tables 8 and 9 and indicates that no factors are intercorrelated to cause multicollinearity problems in the multivariate models 1 and 2 (see the next section).

		T1	T2
Involvement	Mean	2.00	1.98
	sd	0.83	0.82
	N	514	414
Urgency	Mean	4.25	4.16
	sd	0.72	0.76
	N	514	457
Advantages	Mean	3.51	3.64
	sd	0.68	0.78
	N	490	469
Disadvantages	Mean	2.59	2.62
	sd	0.76	0.69
	N	467	411
Quality of care	Mean	0.63	0.63
	sd	0.48	0.48
	N	474	395

Table 1.

Differences between T1 and T2, a t-test of dependent and the perceived change of quality of care due to the project.

3.2 Explanatory results

3.2.1 The factors that determine the outcomes of the project

First, the outcome of the analysis with estimated quality of care change (due to the collaboration project and according to professionals) as the dependent variable (i.e., earlier presented as model 1) is presented in Table 2. A first result is that, as measured by surveying the professionals, the quality of care remained the same over time (T1–T2). Table 2 also shows the relationship of this dependent variable with the sense of urgency, involvement, advantages, and disadvantages as perceived by the professionals. A higher perceived quality of care at T1 (B = 0.78, p < .05) and T2 (B = 1.112, p < .05) is related to professionals who felt a high urgency of the project. Also, advantages experienced by professionals are positively related to a higher score on perceived quality of care at T1 (B = 1.402, p < .05) and T2 (B = 1.415, p < .05). Disadvantages experienced by professionals are only related to a lower perceived quality of care at T1 (B = −0.581, p < .05).

	Perceived quality of care by professionals at T1		Perceived quality of care by professionals at T2
	B	(S.E.)	B	(S.E.)
Fixed part
Cons	−6.114	(1.264)	−9.329	(1.626)
Project level
Generic patient population3	0.341	(0.348)	−0.506	(0.451)
Within primary care4	0.066	(0.273)	−0.529	(0.331)
Professional level
Female1	−0.473	(0.326)	−0.289	(0.337)
Wage earning2	−0.251	(0.283)	0.389	(0.313)
Involvement T1	0.242	(0.165)
Urgency T1	0.780*	(0.224)
Advantages T1	1.402*	(0.230)
Disadvantages T1	−0.581*	(0.196)
Involvement T2			0.119	(0.178)
Urgency T2			1.112*	(0.225)
Advantages T2			1.415*	(0.274)
Disadvantages T2			0.247	(0.232)
Random part
Project level
Cons/cons	0.242	(0.192)	0.699*	(0.327)
ICC5	0.069		0.175
N project	52		45
N professionals	427		364

Table 2.

Multilevel logistic regression on the quality of care realized by the project according to professionals (second-order professional quality of care (PQL)).

¹

Ref cat: male.

²

Ref cat: self-employed.

³

Ref cat: specific population, e.g., diabetes type 2.

⁴

Ref cat: a collaboration of primary care and other sectors.

⁵

ICC = Intra Class Correlation.

*

Significant at p < .05.

The Intra Class Correlation (ICC) is 0.0685 at T1 and 0.1752 at T2, meaning that in both models most of the variance exists at the professional level. However, at T2 the influence of the project variables is substantial (ICC = 17.5%). This means that in explaining these outcomes, the differences between projects (i.e., project characteristics) should be taken into account.

In Table 3, the results of the multilevel analysis of the project leaders’ perceived success of the project as the dependent variable (earlier presented as model 2) are shown. The analysis does not show significant effects of any of the explanatory variables. Most of the variance is explained at the project level (ICC = 0.576 in the first model, ICC = 0.592 in the second model), meaning that the differences between the projects determine the largest part of this dependent variable, being the ‘project success’ as perceived by the project leaders. While in the first model projects aiming at a generic patient population show a higher score on project success, this result is not significant in this second model.

	Professionals (T1), project success (T2)		Professionals (T2), project success (T2)
	B	(S.E.)	B	(S.E.)
Fixed part
cons	0.833	(1.286)	0.817	(1.578)
Project level
Generic patient population3	2.131*	(0.851)	1.452	(0.913)
Within primary care4	0.004	(0.312)	−0.064	(0.326)
Professional level
Female1	0.071	(0.307)	0.021	(0.330)
Wage earning2	0.087	(0.282)	0.048	(0.318)
Involvement T1	−0.030	(0.171)
Urgency T1	−0.053	(0.222)
Advantages T1	−0.017	(0.216)
Disadvantages T1	−0.087	(0.186)
Quality according to professionals T1	0.095	(0.303)
Involvement T2			−0.006	(0.182)
Urgency T2			0.004	(0.226)
Advantages T2			0.024	(0.247)
Disadvantages T2			−0.074	(0.237)
Quality according to professionals T2			0.009	(0.329)
Random part
Level: Project
cons/cons	4.463*	(1.099)	4.769*	(1.275)
ICC5	0.576		0.592
N project	48		42
N professionals	400		320

Table 3.

Multilevel logistic regression on the success of the project at T2 based on the interviews with project leaders (first-order monitoring query language (MQL)).

¹

Ref. cat: male.

²

Ref. cat: self-employed.

³

Ref. cat: specific population, e.g., diabetes type 2.

⁴

Ref. cat: a collaboration of primary care and other sectors.

⁵

ICC = Intra Class Correlation.

*

Significant at p < .05.

4. Discussion

A key question in the many studies on collaborative processes between organizations in health care remains how collaboration actually occurs and why it works out or not [19, 20]. Moreover and especially in health care, it is necessary to pay attention to the role of the professional in this collaboration [17, 18, 19]. Just like other studies on inter-organizational partnerships [19], this chapter shows that the role of professionals is critical to take into account. This concerns the professionals’ sense of urgency and the extent to which they perceive it to be beneficial to cooperate in collaboration projects. Next to this, the role of the project leader is of importance. This is addressed by many leadership theories [30] and where mainly formal leadership was researched, recently there was a shift toward more research into informal leadership. Informal leadership requires influence at different levels, which is important for inter-organizational collaborations. Particularly in (primary) health care, informal leadership is important to strengthen inter-organizational partnerships by setting the right preconditions, e.g., at the level of steering committees [30, 31]. While collaborations between primary healthcare organizations with multiple professions ask for leadership [30], still little attention is paid to both the role of professionals and leaders.

Looking at the results of our study, it appears that for the success of inter-organizational partnerships, the role of both project leaders and professionals is intertwined. Professionals have an independent effect which depends not only on the quality of care change of collaboration projects, but also on the type of project. The steering of inter-organizational partnerships might be a case of formal and informal leadership, as boards and professionals need to work together. To overcome boundaries between stakeholders, project steering needs to be done together as in ‘shared leadership’. In this way, our study complies with the study by Schot [19], in which professionals actually have a role for themselves in overcoming typical coordination problems in inter-organizational partnerships. While this seems to be obvious, often how to involve professionals and to actually collaborate at different levels is still a struggle in many inter-organizational partnerships. Given that professionals in the 69 projects of our study were not always aware of the project, it is necessary during projects to have in mind who should be involved where, when, and how. In our study, a high percentage of 16% of the professionals was unaware of participating in a project to improve collaboration, even after some years. Although this is a remarkable result, it could be due to changes in the professionals in those projects. Dow [32], for example, shows that most networks of collaborating professionals change rapidly over time.

In addition to shared leadership, characteristics of projects are also important to success. In our second model, projects oriented on generic patient population showed higher project success. It could be that primary care collaboration specifically contributes to the care for more complex patients with multiple diagnoses. To support this, it goes beyond the topic of this chapter, i.e., further research on patient populations is needed.

Strengths of this research include the longitudinal data and multilevel analysis. A large dataset with multiple projects and data on different levels was used, i.e., management level and the level of professionals delivering care in these projects. It provided the opportunity to do multilevel analysis, and thereby account for clustering in the data. In this way, the difference between projects and between professionals within projects was taken into account, including the differences over time. Moreover, professionals did not only judge their own success, but it was also judged by the interviewers of project leaders.

However, there might be a selection bias if more positive professionals and projects participated in the survey. Whereas the negative professionals and the less successful projects did not send out the questionnaire as the project was already terminated, or professionals were less committed to respond. On the contrary, the decrease in the sense of urgency, involvement, and a stable quality of care over time is an indication that not only positive professionals responded.

In the study, we use the data of the interviews with project leaders to measure the success of the project. This could be biased by biased responses, as project leaders were monitored by the funding organization. Still, the interview was done at the end of the project and the public reports based on these interviews did not show results of identified projects. Therefore, the possible bias might be limited. Another limitation is that the project leader and professional could be the same person, a case of so-called hybrid professionalism [31]. Although less than a third of the professionals also is an initiator, the outcomes are judged differently. Therefore, we conclude that the levels are indeed different levels.

5. Conclusions

Different studies show a key role for professionals in inter-organizational partnerships [19, 21]. However, the role of the professional is not yet well understood [19, 22]. This study adds to the understanding of this role and shows that quality of care as perceived by the professional is higher in collaborations where they feel a sense of urgency for the collaboration, and also see advantages in collaborating. The professionals’ perceived quality of care is, however, not related to the final success from a managerial and project perspective. And neither is the professionals’ sense of urgency, involvement, and benefits. This confirms that professionals play a significant role in inter-organizational partnerships, but project success is also dependent on project characteristics. At the project level, the scope of collaboration and type of patient care also determine the quality of care according to professionals as well as the success of the project. To conclude, project level and professional level are both important for outcomes of inter-organizational collaboration projects in primary care and therefore need to be aligned with each other.

5.1 What this study adds

This study is the first in primary care that looks at the role of professionals in inter-organizational partnerships over time and its impact on outcomes. It provides evidence that a sense of urgency by the professional and their experienced advantages from a collaboration are necessary for positive results on the quality of care. This takes up the debate about bottom-up implementations, and shared leadership as a necessary condition. Often, when managers start a collaboration project, this is perceived ‘bottom-up’ by governments and policy advisors. However, this might be a ‘top-down’ implementation in practice from the perspective of professionals. In this chapter, it was shown that inter-organizational collaboration as a strategy to improve primary care should not only be driven by managers in a steering committee, but also by professionals who should be involved at all times.

Conflict of interest

The authors declare no conflict of interest.

A. Appendix

See Table 4.

	T1		T2
	N	%	N	%
Sense of urgency
Totally unconvinced of the importance	1	0.19	0	0
Unconvinced of the importance	6	1.17	10	2.19
Neutral	60	11.67	72	15.75
Convinced of the importance	245	47.67	212	46.39
Totally convinced of the importance	202	39.3	163	35.67
Involvement in the project
Only delivering care	176	34.24	143	34.54
In the organization of the project	160	31.13	137	33.09
Initiator of the project	178	34.63	134	32.37
Advantages
Totally disagree with the advantages	4	0.82	2	0.43
Disagree with the advantages	20	4.08	14	2.99
Neutral	211	43.06	203	43.28
Agree with the advantages	234	47.76	183	39.02
Totally agree with the advantages	21	4.29	67	14.29
Disadvantages
Totally disagree with the disadvantages	33	7.07	15	3.65
Disagree with the disadvantages	166	35.55	159	38.69
Neutral	232	49.68	204	49.64
Agree with the disadvantages	31	6.64	33	8.03
Totally agree with the disadvantages	5	1.07	0	0
Perceived quality of care
Decreased or stable quality of care	174	36.71	145	36.71
Increased with the disadvantages	300	63.29	250	63.29
Success of the project
Neutral or unsuccessful project			24	27.69
Successful project			41	72.31

Table 4.

Descriptives.

B. The principal component analyses

At both T1 and T2, the 27 advantages and disadvantages were subjected to a principal component analysis. Prior to performing this analysis, the suitability of data for the analysis was assessed:

• Inspection of the correlation matrix at both T1 and T2 revealed many coefficients of .3 and above.

• The Kaiser-Meyer-Oklin value was .91 for the items measured at the beginning of the project and .87 at the end, so they exceeded the recommended value of .6 (19).

• And statistical significance was reached for Bartlett’s Test of Sphericity at both T1 and T2 (20).

The data were suitable for principal component analyses. Both the principal component analyses at T1 and T2 revealed the presence of six components with eigenvalues exceeding 1 (Table 5). However, an inspection of the scree plot revealed a clear break after the second component. So, the two components were taken for further analysis.

	T1				T2
Comp	Eigenvalue	Difference	Proportion	Cumulative	Eigenvalue	Difference	Proportion	Cumulative
1	8.57	4.52	0.32	0.32	7.99	4.47	0.30	0.30
2	4.05	2.40	0.15	0.47	3.52	1.64	0.13	0.43
3	1.65	0.39	0.06	0.53	1.88	0.54	0.07	0.50
4	1.26	0.20	0.05	0.57	1.34	0.08	0.05	0.55
5	1.06	0.00	0.04	0.61	1.26	0.24	0.05	0.59
6	1.05	0.19	0.04	0.65	1.02	0.13	0.04	0.63
7	0.86	0.07	0.03	0.68	0.89	0.10	0.03	0.66
8	0.79	0.05	0.03	0.71	0.79	0.01	0.03	0.69
9	0.74	0.10	0.03	0.74	0.78	0.06	0.03	0.72
10	0.64	0.07	0.02	0.77	0.72	0.01	0.03	0.75
11	0.57	0.37	0.02	0.79	0.71	0.11	0.03	0.77
12	0.54	0.05	0.02	0.81	0.60	0.03	0.02	0.80
13	0.49	0.02	0.02	0.82	0.57	0.01	0.02	0.82
14	0.47	0.01	0.02	0.84	0.56	0.05	0.02	0.84
15	0.46	0.01	0.02	0.86	0.50	0.01	0.02	0.86
16	0.45	0.03	0.02	0.88	0.49	0.01	0.02	0.88
17	0.42	0.03	0.02	0.89	0.47	0.10	0.02	0.89
18	0.39	0.02	0.01	0.91	0.38	0.00	0.01	0.91
19	0.37	0.03	0.01	0.92	0.38	0.03	0.01	0.92
20	0.34	0.02	0.01	0.93	0.35	0.02	0.01	0.93
21	0.32	0.02	0.01	0.94	0.34	0.02	0.01	0.95
22	0.31	0.03	0.01	0.96	0.31	0.03	0.01	0.96
23	0.28	0.01	0.01	0.97	0.29	0.04	0.01	0.97
24	0.27	0.04	0.01	0.98	0.25	0.02	0.01	0.98
25	0.24	0.03	0.01	0.99	0.23	0.03	0.01	0.99
26	0.21	0.02	0.01	0.99	0.20	0.02	0.01	0.99
27	0.19	.	0.01	1.00	0.18	.	0.01	1.00

Table 5.

Initial eigenvalues T1 and T2.

To aid in the interpretation of these two components, Oblimin rotation was performed. The principal component analysis revealed the presence of two components, showing that the 14 items formulated as experienced advantages did fit into one scale, and the 13 items that experienced disadvantages could also be joined together (Table 6). The rotated components are shown in Table 7. The results of this analysis support the use of the advantage items and the disadvantage items as separate scales.

	T1			T2
Due to the project…	Comp1^1	Comp2^1	Unexplained	Comp1^1	Comp2^1	Unexplained
My professional knowledge is extended	0.30	0.00	.43	0.27	0.02	.57
I get to know other organizations	0.23	0.01	.66	0.18	0.02	.82
I get to know other disciplines	0.22	0.03	.71	0.20	0.05	.78
Time is saved	0.19	−0.05	.72	0.22	−0.07	.62
The time with patients increases	0.25	−0.01	.58	0.28	0.01	.53
The quality of care is better	0.28	−0.02	.45	0.30	−0.04	.41
A better work environment is provided	0.27	−0.03	.47	0.30	0.00	.44
My work is more fun	0.31	−0.01	.38	0.33	0.01	.36
My work is more efficient	0.27	−0.06	.46	0.25	−0.06	.53
I experience more support and guidance	0.33	0.05	.34	0.31	0.02	.44
I get more appreciation from patients	0.28	0.02	.51	0.32	0.04	.44
I get more appreciation from colleagues	0.29	0.01	.48	0.29	0.03	.53
I’m earning more money	0.15	0.09	.87	0.10	0.01	.95
All in all the advantages are higher than the disadvantages	0.20	−0.04	.71	0.19	−0.06	.71
I spend too much of my time in meetings	0.10	0.22	.74	0.10	0.28	.64
It takes a lot of coordination	0.12	0.24	.69	0.13	0.29	.60
I have annoyances	−0.04	0.29	.44	−0.02	0.26	.59
I have the feeling that my work is being checked	0.02	0.30	.45	−0.01	0.28	.56
My decisions are limited	0.02	0.33	.37	−0.03	0.29	.52
There is a lot bureaucracy	−0.03	0.27	.48	−0.01	0.30	.52
There is less time for patients	−0.03	0.29	.41	−0.01	0.31	.45
It requires time and effort till results are shown	−0.10	0.16	.70	−0.07	0.19	.72
Competition arises besides collaboration	0.01	0.25	.62	−0.00	0.23	.72
The workload increases without compensation for it	0.02	0.29	.51	0.02	0.26	.66
I am pushed into a different way of working	−0.02	0.32	.34	0.01	0.33	.41
I am losing some of my independence	0.03	0.32	.40	−0.04	0.29	.48
All in all the disadvantages are higher than the advantages	−0.10	0.24	.47	−0.13	0.22	.51

Table 6.

Rotated components per variable.

¹

Rotation: orthogonal oblimin (Kaiser off).

	T1				T2
Component	Variance	Difference	Proportion	Cumulative	Variance	Difference	Proportion	Cumulative
1	6.41	0.21	0.24	0.24	6.10	0.67	0.23	0.23
2	6.20	0.0	0.23	0.47	5.42	.	0.20	0.43

Table 7.

Rotate components T1 and T2.

Rotation: orthogonal oblimin (Kaiser off).

C. Correlations

See Tables 8 and 9.

	Involvement T1	Urgency T1	Advantages T1	Disadvantages T1	Perceived quality T1
Involvement T1	1
Urgency T1	−0.3121	1
Advantages T1	0.1641	−0.4874	1
Disadvantages T1	−0.0291	0.3241	−0.2254	1
Perceived quality	−0.179	0.3954	−0.4242	0.2605	1

Table 8.

Correlations at T1.

	Involvement T2	Urgency T2	Advantages T2	Disadvantages T2	Perceived quality T2	Success of the project T2
Involvement T2	1
Urgency T2	−0.2807	1
Advantages T2	0.1313	−0.4379	1
Dis-advantages T2	−0.0522	0.3872	−0.3574	1
Perceived quality T2	−0.0531	0.3524	−0.3679	0.247	1
Success of the project T2	0.0296	0.092	0.088	−0.01622	−0.0901	1

Table 9.

Correlations at T2.

Funding

This work was supported by the Netherlands Organisation for Health Research and Development (ZonMw). ZonMw did not play any role in the design, data collection, analysis, and conclusions of this chapter [Grant number 154013001].