Unite Us connects health and social service providers together to deliver better care

Intelligent name matching is key to maintaining universal client profiles for a 360 view of social and medical needs

Executive Summary

Vulnerable populations often need multiple services across the care continuum, such as housing assistance, employment services, and clinical services. While the needs of at-risk populations are connected the organizations that meet those needs are usually not. Recognizing this pattern, first among U.S. veterans, then in other communities, Unite Us addresses the disconnect with technology that links health and social services providers, provides visibility into the entire health journey, and tracks 100% of outcomes. When names are one of the few unique identifiers, intelligent name matching from Rosette text analytics helps providers find the correct record faster thus reducing duplicate records and time wasted looking for a record.

Key Highlights
  • Building care networks connects ensures smooth transitions between health and human service providers for better health and financial outcomes
  • Names are the only universally available identifier to create or find universal records of each individual client
  • Complaints from providers about duplicate entries have significantly dropped since Unite Us deployed Rosette’s intelligent name matching

The Challenge: Connecting Services To Deliver Holistic Care

Former service members Daniel Brillman and Taylor Justice know how difficult it is for American veterans to adjust to life back home. Even though there are numerous nonprofit and government organizations helping veterans (find jobs, housing, healthcare, counseling and more), they work in silos that prevent them from addressing the co-occurring needs of veterans and other vulnerable populations.

Addressing health and social needs holistically instead of piecemeal helps secure the future stability of individuals, and frees up the service providers they work with to assist even more individuals.

Connecting distributed services decreases the likelihood that clients will need further assistance in the future. Logically, this makes sense. If an individual receives assistance to secure a new job, but becomes homeless, they may not be able to hold that job. This hurtful cycle for the client is also an expensive one for the providers in terms of resources and time.

Together, Andrew Price, Brillman and Justice founded Unite Us to create integrated technology networks that link related veteran service providers. After successful deployments within multiple veteran networks, Unite Us expanded to connect other health and social service providers across regional and local communities.

Getting To One Universal Profile for Each Client

The key to participating providers being able to easily refer clients to each other, is maintaining one master database of client records.

Unite Us networks contain vetted providers across the spectrum of human, health, and social services. Unlike other directory-like systems that scrape the internet to find service organizations, Unite Us provides a distributed case management system. These community networks are built by taking key community providers through strategy sessions, socialization, and training to ensure that all providers are accountable and committed to collaborating through referrals and data exchange. This system ensures that all community partners are on the same page and addressing all the needs of their community.

To work, Unite Us has to maintain a single client record for each individual so that all providers on the network, have a complete history of each client. This means that the outpatient therapy office knows the client’s medical history and the housing agency knows to recommend homes in neighborhoods and price ranges that correlate to the client’s new job. Individuals are better served when each agency has all the relevant data.

Tracking social data trends

Maintaining profiles not only improves the experience of individual the client, it also allows Unite Us and its network partners to track trends of how and when individuals receive various services, thus learning about the community and its needs on a macro level, and building processes and interventions for the network to better meet those needs.

Looking forward, the Unite Us team is hoping that starting at this granular level will enable them to power proactive change. It may be easy to see that a community has a housing crisis, but it is difficult to understand why without comprehensive data.

Names: A Data Linchpin

Unfortunately, maintaining a single profile across a distributed network of providers is not an easy task. In other industries, individuals make profiles linked to an email or phone number, however, not all members of an at-need community have regular access to internet or cell phone service.

Instead, Unite Us determined that names, with date of birth for verification, would serve as the cornerstone of client profiles. Regardless of financial circumstances or willingness to share personal information, everyone has a name and birthdate and people are used to self-identifying by them when receiving social services. It’s the lowest common denominator of personally identifying information.

“If someone raises their hand and says ‘I need help,’ but they’re concerned about sharing something as potentially sensitive as their social security number, we never want to turn them away,” says Unite Us CTO, Michael DeLorenzo. “We had to find a less confidential identifier that was also universal. Names–with birthdates for verification–met that need.”

Name Search Powered By Rosette Finds The Right Records

Unite Us quickly discovered several challenges to using names as their cornerstone data type. Nicknames, typos, misspellings and more were keeping the right records from being found. For example, a search for “Joseph Anderson” would not return a record under “Joe Anderson.” Instead of finding and adding to an existing client record, providers were creating duplicates, compromising the value of their data and the network.

“…the problem is only going to get bigger. Data is spread all over the place. It’s a hard problem and it will only get trickier as we add more data and more users.”

At first, the Unite Us team tried suggesting to users that they use shorter search terms, like only the last name “Anderson” and then manually selecting the right record from those results, however that was counterintuitive for users. They continued to miss existing records and create duplicates, an expensive mistake. According to the Pew Charitable Trusts, reconciling a single duplicate record can cost as much as $1000.

Unite Us began investigating fuzzy name matching methods that they could integrate into their database search tool. “Even if it couldn’t be perfect, we wanted to at least give users a better chance of finding someone who’s already in the system” says DeLorenzo. “Plus, the problem is only going to get bigger. Data is spread all over the place. It’s a hard problem and it will only get trickier as we add more data and more users.”

They moved their database onto Elasticsearch, but although some fuzzy matching was supported, it didn’t cover all the potential causes of missed matches. They needed an intelligent name matching system, and found it in Rosette’s name matching plugin to Elasticsearch. Rosette solves 12 common name matching phenomenon, and also can fuzzy match dates, the second data type Unite Us uses to verify clients’ identities.

Rosette matches names based on algorithms that take into account the origins and structure of names from different languages and cultures. Its fuzzy matching recognizes differences such as phonetic variation (“Caitlin”, “Katelyn”, “Kaitlyn”); nicknames (“John”, “Johnny”, “Jack”); swapped word order (“Zhang Wei” and “Wei Zhang”); and more.

Finding an out-of-the-box solution for their name matching allows Unite Us engineers to focus on the technology that is core to their business. “We’re not a search company.” says DeLorenzo. “While search is important, that’s not something that we excel at building ourselves, or would want to spend the time and resources building. Rosette does everything we’re looking to do long term.”

Success In The Field

With Rosette name matching, Unite Us has already had incredible success in improving operations and making connections for their user networks. Service providers in Pittsburgh’s PA Serves network have received 3,570 referrals through Unite Us since integrating the system in 2016. PA Serves also cut the time spent on administrative tasks by 85%.

Pittsburgh Mercy, another services network that integrated Unite Us called them “human services for the 21st century.”

“We’ve stopped getting complaints from providers about duplicate entries since we integrated Rosette. We even try entering wacky searches to trick the system and the right names always seem to come up.”

Unite Us has even started to see different networks connecting to one another for even more comprehensive coverage. For example, a client from Durham, VA that moves to San Diego can be transferred from a Virginia to a California network so that their profile and history travels with them.

“We’ve stopped getting complaints from providers about duplicate entries since we integrated Rosette. We even try entering wacky searches to trick the system and the right names always seem to come up.” says DeLorenzo. “When the user knows that a client record already exists, they are now consistently able to find it.”

Looking To The Future

Building on their success, Unite Us has ambitions to create more intelligent networks that serve clients better than either a static directory, or the case management and referral system they offer. Because of the disparate service providers they interact with, Unite Us is able to see a more complete picture of clients’ lives individually and as a population. Utilizing that insight to further improve clients outcomes is the next step.

In addition to static field data like names and birthdates, Unite Us networks include hundreds of thousands of documents of unstructured notes from case managers. Developers are also experimenting with Rosette base linguistics for search enablement to derive value from that unstructured text. Understanding and analyzing what is contained in those notes can allow Unite Us to track patterns in client behaviors and ultimately create a predictive system to flag at-risk behaviors and preemptively provide care to clients before they reach a crisis state.