**5. Conclusions**

There is much in Ahrne and Brunsson's account of meta-organizations that is recognizable in our case study. We note the formalized character of the Partnership and the effort made to produce a 'decided order', and the Partnership as 'an attempt' to reshape and influence the field of interest; and we note the work of the Partnership both to structure the contributions of and interactions among members, and to promote collective presence, voice, and action. The theory of meta-organization has validity as a means of asking questions of, 'seeing' and describing, analytically, the character and dynamics of the Partnership as a meta-organization. However, our analysis of the accounts of the Partnership nevertheless raises questions for the theory of meta-organization.

First, the theory of meta-organization takes a strong position on the characteristics that define the form, and Ahrne and Brunsson (2008, p. 92) sugges<sup>t</sup> that where difference and complementarity are the basis for engagemen<sup>t</sup> an associational form would not be a natural choice; rather, they would expect a network. The case study of PiP suggests two possibilities for consideration. Either the partnership, both in its varied membership composition and in its pursuit, as a primary aim of collective action might be a hybrid, for example, a 'whole network' in Provan et al. (2007) terms, formally constituted and bounded, but with accepted variety. Alternatively, the Partnership may have evolved through phases from association to network and back towards association, in response to experience in and signals from its environment, but also seeking to follow the will of its members and to distinguish the Partnership clearly from its members' capabilities. This would be the explanation of co-evolutionary scholars. A third possibility looks to the theory: Meta-organization theory may find it important to respond to empirical studies and to other, relevant theoretical traditions. In doing so, it might 'unbound' itself from its more stringent assumptions, and tolerate a more diverse set of characteristics of the form. Berkowitz and Dumez (2016) also sugges<sup>t</sup> the criterion of similarity may need further thought.

Second, since the theory of meta-organization focuses largely on internal dynamics, it makes sharp distinction between what is inside—the member organizations—and what is outside—all other organizations. Our case study, however, suggests that beyond this 'governance boundary' of formal membership, the environment remains highly salient. In Rodrigues and Child (2008) terms, co-evolution continues: Indeed, the differences between institutional environment and internal organization are often less readily discernible and are more closely intertwined than Ahrne and Brunsson suggest. Pfeffer and Salancik point out, " ... *the boundary is where the discretion of the organization to control an activity is less than the discretion of another organization or individual to control that activity*" (Pfeffer and Salancik 1978, p. 32). The Partnership is a decided order, but it is also an emergen<sup>t</sup>

order. It is associational in form and practice; however, it also promotes and is understood by members as catalyzing and delivery changes to the production systems of members, and, by implication too, of non-members who form part of the system of children's services. Ahrne and Brunsson admit such 'collaboration' to the list of reasons for creating meta-organization, but they are pessimistic about its success.

This paper has examined the changing membership composition of a Partnership, which has many of the characteristics of a meta-organization as set out by Ahrne and Brunsson (2008, 2011). It traces the growth and stabilization of membership, enabling and limiting factors, and focuses on the pattern of membership composition over time—we term this 'compositional dynamics'. Ahrne and Brunsson's account stipulates an essential similarity of identity of members as a defining characteristic of a meta-organization. The paper explores what similarity might mean, how that criterion might break down, and what consequences compositional diversity might have for the meta-organization. Following the membership of the case study, Partnership, over some considerable time (17 years), we have been able to explore some effects of membership change on the focus and structure of the meta-organization, what it hopes to reach, and what it does. The paper has also suggested that the boundaries between the meta-organization and its environment and between members and non-members needs further attention in meta-organization theory. We note that the institutional environment remains of significance as a system of meanings that help to define similarity or difference between existing and potential members. Further, active non-members could be recognized as playing a potentially important role, if not in the governance, then in the energy and action radius of the meta-organization. Both questions about the boundaries of meta-organization in theory and practice would merit further case and comparative case research. The relationship between membership composition and the pattern of meta-organization activity over time is also an important area in the assessment and continuing elaboration of this novel and intriguing theory. Such detailed empirical studies will also help in appraising the arguments in theory about the distinctive character in meta-organization and its relationship to other (adjacent) types of inter-organizational entity.

**Author Contributions:** The authors contributed equally to this work.

**Funding:** The Strategic Research Council at the Academy of Finland, project CORE (313013 + 313017), made the contribution by Sanne Bor to the writing of the paper possible.

**Conflicts of Interest:** Sanne Bor declares no conflict of interest. Steve Cropper is Academic Advisor to Partners in Paediatrics and has been since 1998. He is a member of the Core Group of the Partnership and attends the Board of Members as Academic Advisor. He received an annual fee from Partners in Paediatrics from 2000-2011, but since 2011 has undertaken the role *gratis*. He has published about the Partnership and its work on Managed Clinical Networks before, naming the Partnership with its consent, and will be preparing a history of the Partnership for publication, with past officers of the Partnership, again with the Partnership's approval and support.

### **List of Annual Reports Partners in Paediatrics. (www.partnersinpaediatrics.org):**

