Exchange and Reconciliation of Clinical Decision Support Outputs across Systems for Coordinated Quality Improvement: Results and Future Direction from an Implementation in a U.S. Population Health Partnership

Background: In a population health partnership, shareable clinical decision support (CDS) can reduce service duplication and promote patient wellness by presenting consistent information to all members of a cross-functional, distributed care team. However, existing health IT standards present challenges in the exchange of CDS outputs from multiple systems. Objectives: To exchange and reconcile CDS outputs across systems in a partnership between an integrated health system (Deaconess Health System) and a population health services organization (Evolent Health). Methods: We developed a bi-directional HL7-based interface for CDS outputs between Deaconess’s Electronic Health Record (EHR) and Evolent’s population health services platorm (PHSP). The mapping of CDS logic between systems enabled this interface to automatically reconcile inconsistent CDS outputs. Fifteen quality measure (QM)- based CDS rules to identify care gaps were selected for this initiative. These care gaps remind Deaconess’s clinicians to provide care or documentation necessary for gap closure, and also guide Evolent’s care management services. Results: Two months after launch, Deaconess reconciled 14,040 care gaps from Evolent using data only available in the EHR. Additionally, 1,047 care gaps were resolved due to patient refusal or clinical inappropriateness, and 246 gaps were closed by services or prescriptions provided during clinical encounters. Conclusions: We implemented an HL7-based interface to exchange and reconcile a large volume of CDS outputs between a health system EHR and a PHSP. Future direction is to standardize the linkage between a CDS rule and its reference QM by universal identifiers and a taxonomy of variations.