“Live as if you were to die tomorrow. Learn as if you were to live forever.” — Mahatma Gandhi
We often give lip service to the idea of a learning health system, but we must all work to build it effectively and efficiently if we do not want our healthcare “system” to collapse under its own weight.
Over the past decade we have made great strides in the digitization of medicine. The challenge before us today is to extract insight and improvement from all those bits and bytes, to extract knowledge from all that data that is usable and that adds value to patient treatment, outcome and experience for all stakeholders in the healthcare system.
If we do not proceed to improve both the efficacy and efficiency of healthcare, then we will have failed. As the proportion of gross domestic product expended on healthcare approaches 20 percent, and as life expectancy in the U.S. continues to drop, we need to make a radical shift, one worthy of the Mahatma’s admiration, in order to move healthcare and the health of individuals and populations in the right direction.
Now is the time for health IT leaders to internalize these characteristics of learning health systems and advocate to all of their constituencies so that we can make great strides in improving health and healthcare and reducing costs. As in many domains, the future is already here – it’s just not very evenly distributed. Some healthcare organizations have become learning health systems – but most have not, and we have a lot of work to do. This work is foundational to being able to deliver on the promise of value-based care – which can’t really happen without being able to understand and communicate and replicate healthcare that works for each patient in all contexts.
The Agency for Healthcare Research and Quality describes learning health systems as those that:
Many health systems have most of the tools necessary to become learning health systems, but not all are coordinating all of these assets, and not all have the institutional support and resources necessary to deploy all of these tools as effectively as possible.
It is critical that health system leaders understand that these tools are not simply nice-to-have window dressing and that they mobilize all available resources in order to bring them all on line in a coordinated fashion.
The latest generation of health IT tools are going to be able to supercharge the approaches taken by health systems seeking to become learning health systems.
Not only can we gather evidence in order to apply it to guide care, we can analyze that data using algorithms – and human input – in order to add value to the raw data before using it to guide care in real time. One thing we have learned over the past 10 years is that we need both technology and the human touch – neither one standing alone is truly sufficient – to guide care in as optimal a manner as possible.
Not only can we use health IT to share evidence with clinicians to improve decision-making, we can use it to collect data from patients in order to personalize decision-making. We know that it can take upwards of a decade for innovation to make it from lab bench to clinical practice. That process must be short-circuited. In addition, sensors and smartphone applications can collect valuable data from patients, and when that data is cross-referenced with data already in the clinical record (preferably an integrated longitudinal health record that crosses institutional boundaries) it can be analyzed and used in a productive manner without unduly burdening patients or providers.
Not only can we include patients as members of the care team, we can amplify their voices by adopting a learning mindset, by truly hearing them and by integrating the patient perspective more holistically into the healthcare paradigm. Until this happens on a regular basis the healthcare system has one hand tied behind its back, and is not fully serving its core constituency.
Not only can we capture and analyze data to improve care, we can identify what data is useful to collect and analyze, we can work to improve the signal-to-noise ratio in order to address the data overload facing clinicians, and we can work to develop accurate algorithms for data analysis free of human biases without losing the value of the experienced clinician’s instincts.
Not only can we implement processes for continuous learning and improvement, we can demonstrate commitment to these processes, and the ultimate value they deliver, by dedicating the resources necessary to implement all of the other elements of a learning health system. Without wholehearted commitment by leadership and by institutions, the enterprise cannot succeed. The bits and bytes are the basic building blocks – but without a well-funded infrastructure of human and institutional commitment, and resources that have the unwavering support of system leadership, the center cannot hold, and technological tools alone will be powerless to move the health system forward.
While building the learning health system we must vigilantly protect patients’ rights, including patient data privacy and security. We see new evidence of erosion of data privacy and security on a weekly basis, and we must work to ensure that patient data is adequately protected.
We need to build a network of connections between the islands of healthcare resources. We need the vision and the clarity to see the far shore, to visualize the pattern of connections that need to be built, and to marshal the resources to prepare for the tsunami that is already at our doorstep.
We need to do it today. And we need to make sure that it is built to last.
The views and opinions expressed in this blog or by commenters are those of the author and do not necessarily reflect the official policy or position of HIMSS or its affiliates.