Authors:

Abstract

Introduction

Fragmentation of health care services remains widespread, resulting in adverse effects such as inefficient and inequal care. Networks have the potential to address those ‘wicked’ problems. However, current ways of paying seem to impede coordination and collaboration. Health-care providers are predominantly reimbursed through traditional payments models that reward volume rather than value, such as fee-for-service or diagnosis-related groups. It is believed that moving from separate provider reimbursement to joint reimbursement of networks stimulates coordination between providers and encourages flexible use of resources. Ultimately, jointly reimbursing networks through network-level payment models is assumed to lead to improved performance. Aims Objectives Theory or Methods

Our aim was to study how network-level payment models affect the performance of health-care networks. Due to the broad nature of this aim, a scoping review was conducted. Six bibliographical databases were searched, combining various terms that reflected ‘payment model’ and ‘interorganizational network’. Articles were eligible for inclusion if they described networks, payment (models), performance indicators and if it was a peer-reviewed, empirical study set in an OECD country. Network performance was defined as the ability of the network to satisfy the payment model’s objectives. We distinguish four categories of performance: spending, utilization, quality of care, and other consequences.

Highlights or Results or Key Findings

63 studies were included with the majority of studies stemming from the USA (N=58). Payment flows to the network were more common than payment flows in the network (N=57 vs. N=6). Global payment with add-ons (N=49) was the most common model, followed by pay-for-performance (N=7), capitation (N=4) and bundled payment (N=3). In general, the results show that payment models have mixed effects on performance. No single payment model proved able to consistently improve on all four categories of performance. However, in the majority of studies performance in terms of quality and utilization remained stable or improved. Spending was not curbed under the disease-based bundled payment model. Our results also show that the relation between payment models and performance is not necessarily stable, with several factors of importance: timing of the performance assessment, cohort entry year, and scope of services offered by the networks explain differences in performance.

Conclusions

Although network-level payment models are still in their infancy, this review shows that these models have the potential to improve performance of networks. Given the increasing omnipresence of health-care networks, it is fruitful to keep experimenting with joint reimbursement of those networks.

Implications for applicability/transferability sustainability and limitations

It might prove worthwhile to develop more theory-based understanding on the contexts and mechanisms under which payment models lead to certain performance. As alternative payment models gain momentum, this understanding can enhance transferability of knowledge and subsequently support providers in preparing for future (mandatory) payment reform.