Health care providers and social service nonprofits have a hard time working together. Doctors often do not know about social service programs that might benefit their patients, and nonprofits find it difficult to keep health care providers up to date about their services.
When they try to work together, communicating about a patient’s health history or access to services is nearly impossible. These barriers drive up health care costs, and leave patients to navigate the social safety net on their own.
The most vulnerable people in Houston need help with basic needs like food, housing, and employment. When these basic needs aren’t met, it takes a toll on their health. And because health care and social services don’t work together as well as they could, many patients do not get the care or services they need until there is an emergency.
One of the biggest barriers to collaboration is data. Health care providers and nonprofit organizations use different IT systems. They name their data fields differently. And they are vigilant about patient privacy, so they don’t want to share data unless they know it’s safe.
In order to coordinate patient care across doctors and social workers, the Network of Behavioral Health Providers (NBHP) developed a care coordination model and launched the Greater Houston Community Coordination of Care (C3) pilot program. It provided a pathway for these organizations to work together.
Over the last two years, we helped NBHP evaluate the care coordination pilot program. This summer, we released the final program evaluation. Check it out for some fun data visualizations.
This post is a summary of the report, and our key findings after evaluating the pilot program.
The C3 Pilot Program
In 2018, the Network of Behavioral Health Providers (NBHP) brought together medical, behavioral health, and social service organizations, governmental entities, and academic experts to identify the major barriers to care coordination in Greater Houston. Together, they developed recommendations for a care coordination pilot program.
As a result, the C3 Pilot Program launched in Fall 2019. It brought together a network of healthcare and social service providers to “treat the whole patient.”
The program was modeled, in part, after the Pathways Community HUB model. It had three pillars: (1) a dedicated staff of “care coordinators” who assessed clients and organized their care across organizations, (2) twelve participating service providers who helped clients, and (3) a unified IT system and data platform for sharing and tracking client information, referrals, and service histories across partners.

The C3 network included some of the largest and most impactful health care and social service agencies in Houston.
The C3 pilot program was open to adults and children who live with their families, have a diagnosable mental illness, a need for at least one social service, and have a household income at or below 200% of the Federal Poverty Level. Clients also had to live in specific neighborhoods with high social needs that were close to the network of providers.
Over the course of two years, 91 clients enrolled and received services. These clients represented the race-ethnic diversity of the Houston region (40% Hispanic/Latino, 34% Black/African American, and 24% Non-Hispanic White).
The pilot program ended in May 2021. The lessons learned are a critical part of future care coordination efforts in the region. Our evaluation report draws on client histories and assessments recorded in the system, as well as a landscape of insights across original surveys, focus groups, and interviews with clients, care coordinators, and participating partner organizations.
1. Dedicated staff as “care coordinators” are essential.
During the course of our evaluation, we heard repeatedly that the program was succeeding because of the care coordinators. They were the face of the C3 program. They were face-to-face with clients, helping them identify resources and programs that met their complex health care and social service needs. Then they would help clients get help by navigate confusing processes.
The care coordinators were a bridge between their clients and the organizations that provided help. For clients, care coordinators were allies and accountability partners as well as community navigators and advocates who helped them plan and access the services. For providers, care coordinators kept track of their clients, making it much easier to provide treatment.
“Sometimes you need a little push so that you don’t fall off track and get distracted and do something different. My care coordinator was there to make sure that I stayed on track… she made it easier to find a job and a place to live because we worked together.” – C3 Client
Healthcare and social service organizations are strapped for time and resources and need help. An investment in care coordinators eases this burden and improves community health outcomes.
2. The Pathways model was successful and should be used in future efforts.
The C3 pilot program was based upon the Pathways Community HUB model, which is an evidence-based model that connects clients with up to 20 different prescribed “Pathways” of services that address their health and social service needs.
In the C3 program, care coordinators worked with clients to identify different Pathways of care. The chart below summarizes the Pathways for C3 clients during the course of the program. Out of 483 total Pathways opened, more than half were for Social Services Referrals. These included referrals for food, financial support, employment, and educational services.

Over half of the Pathways opened were for social service referrals, highlighting the urgent need for help navigating services.
These Pathways took the guesswork out of figuring out treatments. Care coordinators could reference the Pathway in the shared system and monitor their clients’ progress. In some cases, like applying for food stamps, CHIP, and Medicaid, care coordinators were able to open and close Pathways in one day. In other cases, Pathways were open longer and the care coordinators were able to keep an eye on how their clients progressed towards treatment goals.
Pathways also provided a clear idea of what success looks like. And when success is clarified, it’s also possible: 69% of Pathways were completed successfully during the pilot program. An additional 12% were partially successful.
For the community, Pathways allow for easier analysis and community-wide benchmarks. Future care coordination efforts that use the Pathways model could compare the Pathways opened and success rates in their program with those of the C3 pilot program, as well as other Pathways programs around the country.
3. Covid-19 put a strain on C3 clients and the program.
The Covid-19 pandemic started halfway through the pilot program and its effects were immediate.
Prior to Covid-19, providers were receiving a broad range of referrals connecting clients to social services across the community. After March 2020, there was an abrupt shift to basic needs of food, rent, and employment. Many clients were struggling and on the verge of homelessness.
This was clear in the data. The chart below shows spikes in the number of client needs related to employment services, food, housing, government programs, and legal services in the weeks after the start of the pandemic.

At the beginning of the pandemic, there was a spike in basic needs assistance, such as food, housing, and employment.
The Covid-19 pandemic also had a significant impact on partner organizations. The pandemic stopped most of the face-to-face contact with clients. In some cases, help was provided by phone or virtually. But this represents a dramatic change from services before the pandemic. It also became much more difficult to identify potential new clients for the C3 program.
These changes caused C3 partner organizations to experience financial strain and staffing shortages. In some cases, C3 representatives at a partner organization were furloughed or let go, leading to confusion and gaps in institutional knowledge about the pilot program.
Covid-19 also led to cuts in services and programs, or changes in how they were delivered. Some organizations were no longer offering the same services, and their ability to engage with clients changed. Some organizations were focused on trying to set up telehealth services for the first time.
4. Expanding care coordination could reduce ER visits and save money.
Care coordination is supposed to reduce overall costs in the health care system. By working together to address social service and health needs “upstream,” the goal is to provide treatment before chronic health conditions become an emergency.
But estimating cost savings and making the business case is difficult. For example, it is hard to calculate the impact of C3 on the likelihood of a client receiving emergency treatment. It is also difficult to estimate the costs of ER visits and hospital treatments that were prevented.
As a first step, we focused on the reduction in ER visits based on survey responses we collected from C3 clients. On average, clients reported 0.5 fewer ER visits in the six months after enrolling in the pilot program compared with the six months prior to enrolling.
To estimate the cost of those ER visits, we used data on emergency services cost estimates from the Texas Department of Insurance’s website https://texashealthcarecosts.org/. These data show the average claim amounts for emergency services in the Houston region (770** zip codes) based on 2019 claims data. The data is also broken down by the severity of the ER visit (Low, Moderate, and High severity).
The table below shows the results from our cost-savings analysis. When we applied the model to the C3 client population (91 total clients), the estimated cost savings related to a reduction in ER visits ranged from $21,000-$55,000, depending on the severity.

Expanding care coordinator in Houston could to 1,000 patients could save over half a million dollars.
Expanding the program could yield significant savings.
In the second and third columns, we show the potential savings if the program had reached its original goal of 300 clients and if it was expanded to include 1,000 clients, assuming a similar average reduction in ER visits. In the 1,000-client scenario, the cost savings range from $236,000 to nearly $600,000.
The total cost savings from care coordination are likely much larger. Paying a premium for emergency care visits increase the overall costs of health care and health insurance for everyone.
5. Data interoperability is how Houston will level up
To facilitate care coordination among partners, the C3 program provided partner organizations with access to a shared client data platform called ClientTrack. In general, care coordinators and providers liked the platform and found it helpful in serving their clients.
Still, even with this system, the C3 program was not able to overcome one of the central challenges to care coordination: duplicate data entry that discourages providers from sharing data.
No matter how helpful ClientTrack was for the C3 program, participating organizations all have their own client databases and systems that they still need to maintain. When serving a C3 client, partners had to record client information, interactions, and services in two (or more) systems.
Ideally, any care coordination effort should reduce the amount of data entry. Future care coordination platforms and efforts need to overcome this barrier by funding providers to invest in the technical resources to make their systems interoperable.
In addition to double data entry, partner organizations also have different data standards and classifications that need to be reconciled. For instance, organizations record race-ethnicity of clients differently based on different funder requirements. They also have different names and categories for the many activities and services they provide.
Harmonizing these data across providers is no easy task. Offering organizations a technical resource to write data transformation scripts would significantly reduce the effort organizations need to regularly prepare data for ingestion. For partner organizations with compatible systems, a technical resource could help connect their internal system directly to the care coordination platform using an API.