The promises of a learning health system are endless. We’ve heard about them for years and recent progress should give us all hope that NOW is the time! Intelligent automation could reduce clinician burden. Clinical decision support systems may lead providers and patients to the best treatment option. Major federal initiatives are focused on putting the patient at the center of the health care system. Real-time surveillance systems could track epidemiological phenomena. Predictive models may well identify individuals at high-risk of adverse events.
During a recent AcademyHealth 2018 Annual Research Meeting (ARM) session titled, The Learning Health Care System: Promises vs Realities, panelists discussed converging forces pushing the development of large integrated healthcare organizations as learning healthcare systems, these include: digitalization, quality improvement efforts (e.g., value-based purchasing), evolving sciences, and demands from payers and providers.
Three things are evident after attending this session. First, this is the right time to focus on building a learning health system. This includes large integrated organizations (e.g., Mayo Clinic, Kaiser, Intermountain) with an infrastructure and culture capable of supporting continuous learning, as well as the less organized constellation of health care providers (e.g., small or solo practices). The latter currently face a strong gravitational pull from new payment models and other government and industry signals to organize (e.g., Accountable Care Organizations).
Second, while the ideal learning health system will include data on non-clinical factors that can dramatically shape health outcomes (e.g., socioeconomic, genomics, home testing, biometric monitoring), current efforts are led by large institutions primarily focused on the medical care component (i.e., clinical) of the learning health system.
Third, given the current state, to build the learning health system, we must first harmonize the technical specifications of clinical data, and facilitate its movement across entities. To achieve this, we note several critical factors:
-
- Beyond administrative and claims data: Historically, health data analyses have focused on claims and administrative data. This is not enough. The adoption of electronic health records (EHR) now permit incorporating medical data (e.g., chart information) into a learning health system. However, this data is helpful only when structured in a manner that the learning health system can use, which will require standardizing EHRs. Efforts are already underway to bring together stakeholders in the health community, from clinicians and policymakers to patients and informaticists, to establish a single target for health data standardization. Such a standard can lead to high quality, computable patient information that is interoperable across EHRs. But more work needs to be done. There are many ongoing initiatives to be leveraged – including the “One Human…One Record” initiative of the Standard Health Record Collaborative.
- Facilitating the flow of data: In a system where patients move among insurance plans (e.g., employer-sponsored insurance to Medicare) and obtain care from providers in different institutions, facilitating the flow of data is critical. This is only possible if data is readily accessible and movable. Initiatives are underway to build health information exchange networks – as well as information exchange networks across sectors - but this will take time. As a supplement to these efforts, patients could be vectors that move data from one place to another. Patients can log data about their own health experience and share it with their providers as they go from specialist to specialist. As noted in a recent blog post about the 2018 Health Datapalooza, health system stakeholders can implement recommendations proposed in an October 17, 2017, article in the Journal of the American Medical Association to enable individuals to control their digital health record.
- Data collected once and used multiple times – by multiple entities: There are costs associated with data collection, and duplicative efforts can increase clinician burden. Thus, an efficient learning health system must allow data to be collected once and used multiple times. For example, one institution might use blood pressure data to predict cardiovascular risk, and another to study the relationship between blood pressure and glaucoma. However, a study that combined the data from both entities (or additional ones) to execute a cardiovascular-risk/glaucoma study might be much stronger. This is more likely if the data can be aggregated without additional clinician burden. To fully unlock the power of a learning health system, different entities must be able to use the data; this requires interoperability. AcademyHealth recently released an environmental scan of interoperability guidance across sectors finding that while the health care sector contributed the most guidance, several gaps exist in detailed guidance around critical components of interoperability.
The pace of technological change is ever accelerating. There is a clear vision for a data-driven learning health system, and it is achievable now. Patients are crucial, however one can also imagine a future where public health and biomedical research stakeholders are key members of the learning health system – the sky is literally the limit. Incredible advancements have been made and while there is work ahead, important initiatives like the Office of the National Coordinator’s efforts to promote application programming interfaces (APIs), and CMS’s Promoting Interoperability program (encouraging the adoption of EHR systems that include APIs) could allow us to get there quicker than we previously imagined.
The authors' affiliation with The MITRE Corporation is provided for identification purposes only, and is not intended to convey or imply MITRE's concurrence with, or support for, the positions, opinions or viewpoints expressed by the authors.
The opinions expressed in this blog post are the authors' own and do not necessarily reflect the view of AcademyHealth.
Organizational Affiliates are a critical link in AcademyHealth’s ability to effectively advocate for the field, and support the future field of health services researchers. Organizational Affiliates gain visibility among AcademyHealth membership, enjoy unique networking opportunities, and benefit from event discounts. Click here to learn more.