The paper provides an in-depht explorative analysis of the challenges and requirements to make Integrated Practice Units IPU Porter, 2013 work, relying on evidence from a research on Breast Cancer Integrated Care Pathways ICPs. The latter involved 4 provinces in Italy and the unit of analysis has been the breast cancer ICP, where “integrated” means yet the services’ integration, from preventative screening to end of life care, yet integration between the province Local Health Authority LHA and the public hospital where autonomous from the LHA, by setting up inter-organisational multiprofessional groups for designing a common ICP. Provinces have a middle size with on average 400k to 500k inhabitants. In all provinces, Breat Units BU according to EUSOMA are in place and, in two provinces, two BUs coexist. Indeed, in some cases there are private accredited hospitals with recognised BU. It is largely recognised that BU today represents the paradigmatic example of IPU. The study, based on quantative and qualitative analyses of ICPs, according to the PHM, provides key lessons over the efficacious roles of IPUs to deliver ICPs. The predominant understanding tackles out that the IPU model can not work efficaciously outside Accountable Care Organisations, featured by an integrated commissioning function over all services required by the ICP across different providers either public or private and by the capacity to steer and nudge patients’ choices. This understanding holds a crucial role for the Italian NHS, wherein LHAs are gaining increasing value as Health Population organisations, enlarging the population served and value chain of services provided. In this line, the study provides four lessons. Firstly, IPU as BUs actually does not include all services required as for screening or radiology services: the professional responsibilities and services are fragmented and the BU is not always able to manage effects as drop-out or patients shopping around. The solution of inter-organisational tumor boards are not largely spread-out. Secondly, this fragmentation largely belongs to the context setting, yet from the geography or the urbanisation level of LHA, which can influence the IPU’s efficacy; yet form the local path-dependency in setting-up agreements of integration/collaboration among providers and/or the LHA’s commissioning capacity and power. Third, whereas patient choice is a value and would be recognised, in most cases what can be observed is that most of the women shop around for accelerating some diagnostics more than seeking for centres of excellence, and in any case the latter usually are the most literate persons who maintain a contact with local services. Finally, the IPU model has enhanced the competition with the private providers, but given the mission of the IPU, new competitive patterns are emerging: namely a competition for the market but within targeted-services, and not the overall ICP. Private IPU would not be able to provide full IPC in a quasi –market as the Italian NHS. Indeed, the requirement for public providers to set up IPU/BU hinders any collaborative-competitive pattern unless it will not be promoted directly by the LHA, as commissioner, according to the ACO model.