The patient-centered SCAlable National Network for Effectiveness Research (pSCANNER) will contribute to PCORI's National Patient-Centered Clinical Research Network (NCRN) initiative as a clinical data research network (CDRN). pSCANNER will be a stakeholder-governed federated network that will utilize a distributed architecture to integrate data from three existing networks covering over 21 million patients:
- the University of California Research eXchange (UC- ReX) network, with data from University of California Davis, Irvine, Los Angeles, San Francisco, and San Diego;
- VA Informatics and Computing Infrastructure (VINCI), with data from Veteran Health Administration's (VHA) 151 inpatient and 827 outpatient facilities supplemented with data on veterans and active-duty service members from 231 military treatment facilities worldwide; and
- the SCAlable National Network for Effectiveness Research (SCANNER), with data from UC San Diego, the VA Tennessee Valley Healthcare System, and three ambulatory care systems in the Greater Los Angeles area supplemented with claims and health information exchange (HIE) data.
All of the participating institutions have an extensive history of integrating translational research into clinical operations and organizational priorities for patient-centered care.
pSCANNER has strong foundational data capabilities to serve PCORI's CDRN. All systems have fully operational electronic health record (EHR) systems that are capable of facilitating randomized interventions. The existing data infrastructures have all been used for exploratory analysis in preparation for research; some are also in use for observational and randomized intervention studies. All participating systems have information and processes to recontact patients. The participants' clinical data warehouses routinely collect clinical data for multiyear longitudinal patient records. Security, privacy, and confidentiality of identified patient data in clinical data warehouses are maintained using the same standards as those required of EHR systems in HIPAA-compliant environments. The participating organizations have the proper infrastructure for handling biospecimens and currently host several biorepositories. Streamlining of the Institutional Review Board (IRB) process occurs through a "trust-and-rely" system among the UC sites. VHA has streamlined procedures for access to data once IRB's approve a protocol. Finally, the patient population from pSCANNER institutions is highly diverse in terms of insurance coverage, socioeconomic status, demographics, and health conditions. We will focus on three conditions:
- congestive heart failure,
- Kawasaki disease, and
Patients, patient advocates, domain experts for these conditions, health services researchers, clinicians, and administrators from across the country have agreed to participate in pSCANNER's governance. In addition, we will use innovative user-friendly, online software to conduct a rigorously designed Delphi consensus process that engages 360 patients, clinicians, and researchers in the prioritization of research questions. All systems involved have agreed to allow pSCANNER users, as well as other CDRNs, to access their data using a privacy-preserving distributed computation model and research portal that was successfully piloted in SCANNER. This portal includes a study registry and per-site approval of specific analytic tools and protocols. Federal, state, and institutional policies are encoded in the registry to be easily configurable depending on where the institution is located and which policies it utilizes. Patient and institutional privacy can be preserved with the use of new privacy technology algorithms. Analysis tools include methods that allow model fitting, causal inference, and hypothesis testing without the need for patient-level data to leave the institutions. This workflow allows for local control and increases efficiency by avoiding more complex IRB and data-use agreements, though the network can handle data transfers where permissible.
We have developed two modes for network operation:
- synchronous mode, where pre-approved types of computation on the data are immediately allowed to run at local sites; and
- asynchronous mode, where a human inspects the results of all computations before approving their delivery to the requesting hub.
Our distributed system will allow the construction and evaluation of multivariate models that can be used for statistical process control, data safety monitoring in clinical trials, adjustment for confounders, propensity score matching, risk prediction, and other methods employed in PCOR. For studies that require pooling of data, the hub will store the data in a HIPAA-compliant private cloud. Programming interfaces to the PCORI NCRN will be developed to ensure pSCANNER's interoperability. pSCANNER will adopt the Observational Model for Outcomes Partnership (OMOP) specification as its data model. Mappings from clinical data warehouse data to OMOP are already in place in SCANNER sites and will be developed as part of this project for the other sites.