GESTIÓN EN SALUD PÚBLICA: 2018 Health Datapalooza

miércoles, 25 de abril de 2018

2018 Health Datapalooza

HEALTH OR HEALTHCARE PROBLEM/CHALLENGE BEING TACKLED:

Clinical decision support (CDS) aims to bring the right information, to the right audience, using the right technology and right format, at the right time during work flow to improve quality of care (often known as the Five Rights of CDS.) Translating evidence-based care recommendations, such as recommendation statements within clinical practice guidelines, into computer-interpretable CDS for local implementation is time consuming and expensive. The process involves many steps, including translating guideline recommendations into structured logic, identifying appropriate codes from standardized terminologies, and understanding the nuances of system- or site-specific technical implementations. Many health care organizations do not have the resources or expertise to undertake the entire process. Further, the translation often requires making assumptions about and filling gaps in practice guidelines to enable computerized facilitation. These assumptions, which often reflect local clinical and technical preferences, are valuable pieces of knowledge that are often missing in efforts to make CDS systematic, replicable, and effective across organizations and sites.

INNOVATION/ADVANCE ACHIEVED:

CDS Connect is prototype national infrastructure for developing and sharing computerized CDS, including interoperable CDS “artifacts” and knowledge about their use. It is part of a larger AHRQ effort to advance research evidence into practice through CDS and to make CDS more shareable, standards-based, and publicly-available (http://cds.ahrq.gov). CDS Connect represents two main innovations: a repository of CDS artifacts for public use and a CDS authoring tool. The repository thus far contains CDS developed within the project and focused on cholesterol management. Cholesterol management was the initial clinical domain as a proof-of-concept for the project’s first year. The project developed CDS based on the U.S. Preventive Services Task Force (USPSTF) recommendations for the use of statins in the primary prevention of cardiovascular disease (CVD) in adults. We piloted the CDS through a Health Center-Controlled Network that supports the electronic health record (EHR) and informatics applications (such as CDS) for community health centers. Importantly, details about the process of transformation from the guidelines to the CDS as experienced by the clinicians and patients, including the technical and non-technical considerations and assumptions, are available on CDS Connect.

The authoring tool is designed to help non-software engineers build the structured logic necessary for CDS and to create the CDS in a standards-based format. This facilitates consistency in specifications, transparency when sharing, and interoperability. The project’s second year focuses on pain management and encompasses further development of the repository, implementation of newly created CDS, and demonstration of the infrastructure and tooling. Plans are underway to make the authoring tool open source.

TYPES OF DATA USED:

The types of data that drive the CDS are directly related to the underlying guideline recommendations or evidence-based practice. In the case of cholesterol management and the USPSTF guideline, several types of data are important when clinicians and patients consider starting a statin medication to prevent CVD, including demographic data (e.g., age) and clinical data related to CVD risk factors (e.g., presence or absence of diabetes). Medication data are important because of choices related to dose intensity when a statin is recommended and prescribed. Data about patient preferences and values are also important. We constructed and implemented the CDS as an interactive screen in the EHR that enables and documents shared decision-making between the clinician and patient.

In terms of the technical standards that are used to represent the data and to specify how the CDS should function, we are using two Health Level 7 (HL7) standards, the Clinical Quality Language (CQL) v1.2 and Fast Healthcare Interoperability Resources (FHIR) DSTU2. We use CQL for the logic specification and FHIR for the underlying data model and framework for how the CDS should be operationalized. We also use data from the National Library of Medicine’s Value Set Authority Center, namely, code sets for common clinical concepts that aid interoperability and consistency between implementations.

FINDINGS/LESSONS LEARNED:

Several findings have emerged over the past year. First, we have learned that there is tremendous interest in a common platform to share CDS and knowledge about CDS in a way that reduces the burden of having to author and to develop CDS from scratch. Even larger organizations with strong informatics capabilities have an interest in sharing and reusing CDS artifacts developed elsewhere as building blocks for their own CDS and as a way of learning from what others have already done. Other Federal agencies are interested in CDS Connect as a potential home for their CDS and as a platform for dissemination. To date, work groups supporting CDS Connect have numbered over 30 organizations from both the public and private sectors. Second, we have learned the importance of close collaboration between technical and clinical teams to effectively implement the CDS. Most clinical information systems deliver CDS as a service, but their capacity to incorporate external CDS depends on how that service is delivered technically and how the CDS is viewed clinically. Third, we have learned the importance of external factors, including governance, marketplace, and legal factors, that can impact the readiness and trustworthiness of content in a public CDS repository.

NEXT STEPS/MARKET STRATEGY:

CDS Connect is funded by AHRQ through a contract with the MITRE Corporation, which operates the CMS Alliance to Modernize Healthcare (CAMH) Federally-Funded Research and Development Center. This second year of the project will focus on enhancing the repository, further building the CDS authoring tool, and developing CDS in the clinical domain of pain management/opioid use. We continue to conduct outreach as we seek potential contributors to the repository, implementers, and partners to use the CDS authoring tool. We continue to present at national meetings, such as AMIA, HIMSS, and HL7, and are collaborating with the Patient-Centered CDS Learning Network (https://pccds-ln.org/) in their effort to advance patient-centered CDS. Longer term strategies may include public-private partnerships.