Networks studied exhibited a variety of business and operational models. In the paragraphs below we describe key aspects of network governance and partnership models as well as the major functions health center networks undertook.
Formal governance. We observed a relatively consistent formal governance model across the networks we investigated. These networks are typically governed by a Board of Directors comprised largely of leadership from their member health center organizations. As such, the activities of health center networks, under most circumstances, are controlled by health center Executive Directors. In some cases, respondents indicated that larger health centers that contribute more resources to the network wield greater influence on the Board of Directors than other health centers.
In addition to the Board of Directors, some networks use a series of committees, comprised of both network and health center staff dedicated to specific network priorities or functions. For example, networks that worked on EHR implementation such as Boston HealthNet and the Florida Health Choice Network (HCN) tasked clinical committees and workgroups with reviewing and customizing software prior to roll-out. In addition, networks that provide centrally maintained applications usually have an IT committee where support issues are discussed. In addition to making recommendations to the Board and leadership, committees serve as an important forum for health center staff to provide feedback directly to the network, discuss problems and share lessons learned.
Executive leadership. A feature common to most networks visited is the presence of a single, network-level entrepreneur who served, in name or function, as the network’s CIO. This individual, usually someone with an extensive background in IT from larger provider systems or staff model managed care organizations, provided expertise, leadership and vision around IT not accessible to most independent health centers. In most successful cases, this individual works with health centers to develop a program of collaborative activities, to procure and manage IT vendors, to establish a revenue model using a combination of health center contributions and outside funding, and to attract technical staff necessary to support network activity.
Evidence of strong CIOs was found in the most centralized networks including the Oregon Community Health Information Network (OCHIN), Florida’s HCN and the Community Health Action Network (CHAN) in New Hampshire. Health centers in these networks were very familiar with the CIO and often contacted that person directly when they required assistance. They emphasized that the CIO champion was a crucial factor in providing the impetus to move the group to adopt clinical systems. Typically, health center executives who sit on the Board of Directors are responsible for final sign-off on all network investments. However, these Boards rely heavily on the network leadership to present analyses and recommendations that drive these decisions.
Integration model. We found that health center networks varied considerably in the extent they functioned as integrated decision-making bodies where systems decisions affecting individual health centers were made on the network level. For example, at the Health Federation of Philadelphia (HFP) a process to select a single networked practice management application led only to the selection of a “preferred” application with health centers opting to go on their own rather than give up that level of control to the network. Instead, HFP decided pursue a data warehouse project that would integrate data from individual health center systems.
On the opposite end of the spectrum, HCN in Florida was able to achieve agreement from its core members to hand over control of all of their software applications, data and financial operations (including billing and accounting) to the network and operate under a true shared CFO/CIO model. We found that the level of integration that networks were able to achieve relates to a number of factors, including how they are prioritizing objectives of network formation, the type of partners involved and the historical relationship between partnering organizations. Each of these issues is explored below.
Collaborative history. Well integrated networks were often located in communities that had a more collaborative history than others. CHAN’s health centers, for example, credited their success in adopting a robust system to their small size and highly collaborative environment. Health centers in Boston reported a long history of working together and cooperating with safety net hospitals even while they engage in some competition for patients.
In some cases, such as HFP, networks have been unable to forge member consensus around single applications for practice management or EHR or to convince health centers that these applications are best supported on the network level. Some health centers in Philadelphia acknowledged that there is some reluctance to collaborate because of underlying competition among health centers. In these cases, the networks focus on providing general guidance to health centers and fostering collaboration. Such networks also often work to integrate data from across health centers using interfaces or manual data entry.
Horizontal vs. vertical integration. Most of the networks we visited were horizontal collaborations among ambulatory primary care providers, forming around regionally located health centers, many of which have prior relationships with one another. Within these horizontal collaborations, however, there can often be one member that leads in terms of contribution to the network. This was the case for both OCHIN and HCN. In other cases, regional primary care associations were the convening force that brought health centers together and serve as the center of gravity for the network.
Boston HealthNet and the Southeastern Kentucky Community Action Program (SKYCAP) are key exceptions to the mode of horizontal partnerships. Boston HealthNet is the only true vertically-integrated network we visited, with Boston Medical Center (BMC), the major safety net hospital in Boston that formed after the merger of Boston City Hospital and Boston University Medical Center, providing a strong anchor for the network. BMC, given its resources as a large inpatient institution, often shaped network decisions such as adopting the EHR system being used at the hospital. BMC also has substantial incentive to bring together a coalition of health centers and invest in network development as health center referrals represent a large portion of BMC admissions.
SKYCAP represents a different model where both health and social service providers are connected in a loosely organized structure. Although, the SKYCAP program includes both ambulatory clinics and hospital Emergency Departments, we did not find evidence in this case that including multiple provider types increased incentives for network participation or investment.
Public health involvement. As mentioned in previous sections, several networks forged horizontal partnerships with the wider public health community, such as with Departments of Public Health or Medicaid Managed Care insurance providers. OCHIN and SKYCAP demonstrated the strongest ties to public health stakeholders, attempting to exchange data with these stakeholders with the ultimate goals of improving the tracking of disease outbreaks and other public health trends, and coordinating care among various providers including health centers, hospitals and public health departments. While public health stakeholders were often involved in the consortia of such networks as HCN, OCHIN and CHAN, models of streamlined data exchanged between public health and health centers had not yet been achieved.
Overlapping in networks. In investigating the seven health center networks described in Exhibit 3 above, we came across additional networks engaged in similar activities in the same general geographic area. In some cases, these represented sub-networks or smaller collaborations within networks that were based on historical partnerships among a smaller group of health centers (usually two). This occurred, for example, in Oregon, where two OCHIN health centers were funded by the BPHC for collaborative activities under an organization called Community Health Network of Oregon (CHNO) and some Boston HealthNet health centers located in Boston’s Dorchester community shared a single practice management system.
In other cases, these networks operated on a similar scale as the network we investigated, but in a slightly different geographic region and using a different model. For example, in addition to HCN, we noted the presence of the Community Health Center Alliance (CHCA) in Florida, which provides a networked practice management application and data management capacity for health centers slightly north of HCN’s core health centers. CHCA is a less integrated model for collaboration, offering fewer centralized services to health center members. In the case of Virginia and New Hampshire, other health center networks (the Richmond Enhancing Access to Community Health Project and New Hampshire Community Technology Partnership) had missions that were potentially complimentary to the ones studied, but we found little evidence of collaboration among these networks at the time of the case study.