How to get those databases talking

Among the lofty promises of connecting data systems are faster response times, deeper insights into connected information, and a farewell to duplicate data entry. Alluring as these rewards may be, many attempts to integrate data systems fall far short of their ambition. In this post we’ll share a data integration success story and extract from it the ingredients you’ll need to get your databases talking.

Pre-integration pain

Connected Families New Hampshire (CFNH) is a Care Management Entity that lives up to its name. Specializing in Wraparound services, the focus of their clinical work centers around creating a single plan of care for families of children with the most complex behavioral health needs in the state. These families initially have myriad providers and few natural supports. The Wraparound coordinator tilts that balance toward natural supports as it brings them together around the family in the form of a team that coheres around a single, family-driven plan of care. Unfortunately, the information flow between CFNH and its various partner organizations was decidedly less coordinated (see figure 1 below).

Referrals to CFNH land at the door of the NH Bureau for Children’s Behavioral Health (BCBH) who conduct the Wraparound eligibility process. BCBH records information about the family in a database before passing the information along to CFNH, who in turn must capture the same information in their database. Things only get more complicated as the family begins Wraparound and information needs to flow between CFNH and various other organizations serving the family, including NFI North (Youth Move peer support) and the NH chapter of the National Alliance on Mental Illness (NAMI NH; family peer support). These organizations, of course, each have their own platforms for recording contact notes and tracking service delivery.

Figure 1. Pre-integration data flow

These disconnected systems forced duplicate data entry and created information silos. This in turn contributed to service delays and conflicting information, raising the risk of families falling through the cracks. We knew we needed to get these data systems talking, but given the diversity of reporting requirements, thin resources, logistical barriers, etc. the question was “How?”

Streamlining the systems

In early 2021 BCBH was tracking intake data in wide, colorful spreadsheets only they could access. We’d cultivated a strong working relationship with BCBH through years of Wraparound program evaluation collaboration, yielding trust in our understanding of the practice and technical expertise. Still, abandoning these painstakingly maintained spreadsheets required reimagining them as a relational database that would enhance searchability, pave the way for real-time dashboards, data-responsive email notifications, and – most critically – “push” data from their future database directly to CFNH’s.

We used Quickbase to construct a database for BCBH, the same secure platform-as-a-service used by CFNH, laying the foundation for data integration. We elucidated how the BCBH-to-CFNH referral process operated, identified the pain points, and created a shared vision for how it ought to operate. During the development process we aligned the data collection methods employed by BCBH and CFNH, ensuring critical fields were always captured with the same (or easily translatable) response options. This involved cross-referencing over 400 fields for compatibility and occasionally coaxing compromises. With the hardest part of the work complete, we then leveraged Quickbase’s robust application programming interfaces (APIs) to allow CFNH to copy data from BCBH’s app into theirs with a button click.

With these systems connected, CFNH now receives real-time email notifications from BCBH when Wrapround eligibility is determined. CFNH no longer wastes time duplicating intake data in their database, nor risks committing data entry errors. BCBH has dashboard widgets that track the families as Wraparound unfolds via a cross-database relationship. Both practices have substantially aligned their efforts and work more closely to ensure quality service for Wraparound families. We also facilitated the integration of CFNH’s database with those of NFI North and NAMI NH to similar success. We’ve maintained these collaborative relationships over several years now, adapting to changes in practice and data collection methods as they arise.

Figure 2. Post-integration flow

Key ingredients for integrating data systems

Connecting data systems is no easy feat. The checklist below contains key ingredients to data integration success. May they help you succeed in your own integration efforts.

Ingredient	Challenge rating	Description
Technological capacity	Easy	Without connective features, it is not possible to automate data flows. Most modern platforms have public-facing APIs designed for external connectivity.
Privacy & security	Medium	End-to-end encryption, authentication protocols, least necessary access policies. Privacy and security are prerequisites for system integration.
Buy-in on both ends	Hard	Perhaps the most underestimated ingredient and the place where many efforts fail. Both parties must be sufficiently incentivized to collaborate. The best technology is not enough on its own.
Time, money, communication skills	Medium	Secure these resources before starting. Integration is a slog under the best conditions; the last thing you need is to be under-resourced.
Practice expertise	Hard	Deep understanding of the practice and user workflows. Critical to finding the highest leverage points for integration to increase efficiency and support practice fidelity/quality.
Technical expertise	Medium	Sometimes this can be outsourced, but you’re best served by having someone in-house who can implement the technical aspects of integrating data systems. You may need multiple people depending on your use-case and the systems involved.
Plan B	Easy	Data systems will change unexpectedly, bugs will surface, integrations will break. It’s best practice to have a protocol for when the integration malfunctions to minimize pain. Ideally the systems would be able to operate independently but work best connected.

---

John Erdmann is a Lead Analyst at the Behavioral Health Improvement Institute (BHII) at Keene State College.