HLN's Report on Patient Centered CDS Learning Network Conference

Noam H. Arzt, Ph.D.On October 3, 2017 Daryl Chertcoff and I from HLN Consulting attended the Patient Centered Clinical Decision Support Learning Network (PCCDS-LN) 2017 Annual Conference in Crystal City, VA. HLN was a Bronze-level Sponsor of this event, which brought together clinicians, informaticists, and health policy advocates to discuss current trends in clinical decision support leveraging information from Patient-Centered Outcomes Research findings and patient-specific information.

The Learning Network is an initiative funded by the Agency for Healthcare Research and Quality (AHRQ) whose goal is to encourage collaboration and development of tools and resources to support clinical decision support (CDS) as it relates to individual patients and their caregivers. This work is an offshoot of the Patient-Centered Outcomes Research Initiative (PCORI) which focuses on the relationship between healthcare providers in patients in making better, more informed, healthcare decisions. A sister project to the Learning Network is the AHRQ-funded CDS Connect Repository which has begun to house shareable CDS artifacts which represent evidence-based standards of care.

The conference featured national experts in CDS and patient-centered research in the morning, including a plenary session with two amazing speakers (patient advocate Dana Lewis who spoke about her invention out of necessity of an artificial pancreas, and Dr. Ken Kawamoto from the University of Utah, one of the leading CDS researchers and implementers in the country), and a panel discussion, In the afternoon a participative set of breakout sessions focused on developing an idealized design of the patient-centered CDS process through four different perspectives: patient engagement, measurement and outcomes, user needs and technologies, and continuous learning. Participants worked in small groups corresponding to these four perspectives to first describe the current state of patient-centered CDS, then an ideal future state, and finally steps that might be taken to enable movement from the current to the future state. At various points each group presented their work to all attendees, and at the end key action items were drawn from the small group activity.

Must of this conference focused on patient-centered CDS. But it seems that this term absorbs a number of somewhat different concepts and is used by different people to mean different things. “Patient centered” to some people means “patient-facing” – that is, it refers to an application (in this case a CDS-enabled application) that is used or accessed by a patient directly. To others, patient-centered is synonymous with “patient-empowered” which implies support for more patient control of his or her care and records. Many uses of CDS support do not directly involve patients in real time while still impacting patients and their care; this is true of CDS capabilities in most electronic health records (EHR) which are used by clinicians but not by patients.

For HLN’s work in clinical decision support, this conference presented a number of useful themes and raised a number of important issues. While there was a lot of focus at this conference on user interface, it is important for us to maintain our focus on the back-end services we are developing, like the ICE Open Source Immunization Forecaster. ICE is a fully configurable vaccine forecasting/decision support web service tool. Making patient-specific vaccine decisions based on US Government American Council on Immunization Practices (ACIP) recommendations is a very complex task. It requires translating ACIP recommendations - the detailed rules that examine a patient's immunization history and determine if all vaccinations are valid, and what vaccinations might be missing - into computable language for clinical decision support. ICE evaluates a patient's immunization history and generates the appropriate immunization recommendations for the patient. These recommendations are communicated to the clinician at the point of care, as well as the patient, health department or other users. We usually leave the display of this information up to the system that is sending an evaluation request and receiving the response; we are just concerned with providing the right data back from the service. 

But the display of this information is ultimately what is important, and HLN’s services can make that display easier or more difficult based on choices we make in our technology and deployment. It is equally important that we continue to develop a rich set of application programming interfaces (APIs) to our services, including our traditional SOAP-based Web Services, supplemented by various RESTful services (including FHIR, SMART, and CDS Hooks). In this way our work encourages and enables a wider variety of application development strategies. And all of this work should support and use accepted standards embraced by the healthcare industry. HLN’s own web client to our ICE CDS engine is a Java application that uses our current web services model to submit CDS requests and receive the response. HLN expects to create a new SMART application to replace it using CDS Hooks to interact with the ICE CDS service. This will serve as a proof of concept application for a FHIR-enabled CDS service.

HLN believes that patient-centered means patient-impacting. It should not matter if an application is used by patients or clinicians so long as it supports CDS related to a specific patient and his/her medical situation. HLN’s immunization-related products, like our electronic case reporting products supporting mandated reporting of certain clinical conditions to state and local public health agencies, are patient-specific and impact their treatment directly. These open source products are standards-based and fulfill the objectives of the Learning Network. HLN looks forward to continuing involvement in Learning Network activities and in making our knowledge artifacts - which are already publicly posted for our immunization work - available to the CDS Connect Repository.