Already 20 years have passed since the groundbreaking report from the Institute of Medicine (IOM) was published in 2000 indicating that over 98,000 preventable deaths occurred yearly in U.S. hospitals. In 2004, then President George W. Bush inaugurated the government’s campaign to make EHRs universal within a decade. These key milestones helped to serve as catalysts for healthcare providers to adopt Electronic Health Record (EHR) technology and involved significant financial and operational investment.
As a result, the healthcare industry as a whole has made great strides in patient safety, drastically reducing medication errors while automating previously manual processes and documentation. Large data sets have been created and made available for analysis and reporting while improved IT system integration now links different aspects of the care continuum including the ED, inpatient and ambulatory settings.
Although there is little room to argue against the value created by implementing EHR technology, there are still many lessons to learn from the past two decades of experience. The overall user experience still needs to be improved. Clinicians, especially, have often found adjusting to the new level of documentation now required burdensome and somewhat mind-numbing due to the number of clicks and data elements involved. This “death by clicks” has not only reduced provider’s direct patient interaction, but has also fueled uncertainty that the investment in documentation will yield sufficient reward downstream, in terms of insightful analytics and improved clinical outcomes. Valuable health data still remains trapped - not only outside health system boundaries of incompatible systems and complex regulatory hurdles, but also siloed within the four walls of organizations where departments and programs are still creating redundant datasets housed into a single EHR instance.
Population Health Management: A Much Different Foundation
Even more important than ease of use and adoption of EHRs is a fundamental architectural issue that has challenged the EHR vendor space over the last decade. With the broader shift of the healthcare industry towards value-based care, the foundational design differences continue to be exposed between the focus on quality and outcomes when compared to our traditional fee-for-service model based on transactions and episodes. EHR data models and workflows were originally designed to maximize billing transactions. As such, regulatory compliance and encounters for efficient billing were the main architectural design drivers in clinical documentation, not longitudinal care for individuals and patient populations.
Population health management requires active engagement among a multitude of stakeholders across a community, all sharing data that supports care delivery processes regardless of care setting. EHRs have had some difficulty evolving into a whole-person solution that focuses on managing the patient over a broader span of time. Some of the persistent gaps exhibited by EHRs in effectively enabling Population Health Management include:
1. Interoperability and Comprehensive Data Sets
Population health management must be anchored by a full view of the whole person, including a robust picture of a patient’s complex health history. PHM requires interoperability beyond clinical systems. IT systems and databases must be able to store newer types of data such as patient-generated health data, social determinants of health, environmental, and genetic data in order to truly support chronic care management and care coordination workflows. EHRs must also connect individual team members across organizational boundaries who are documenting and need to access this information. There are many other entities across the continuum that have relevant patient data: post-acute care (PAC) providers such as nursing homes, rehabilitation facilities, home care agencies; reference labs and imaging centers; retail pharmacies; behavioral health specialists. Many of these providers have their own EHRs, but in many cases, their systems are incompatible with those of hospitals and physicians or limited in their ability to integrate workflows and / or data. While EHRs have begun to build the data structures to support this expanded view, there is still quite a bit of work to do around compiling and normalizing data around claims, ADT, practice management and scheduling, post-acute care, behavioral health, and other ambulatory or community-based sources of information. The EHR is focused primarily on the inpatient, departmental, and some outpatient settings; thus, it remains but a piece of the overall patient’s full data profile.
2. Predictive and Prescriptive Analytics
“Big data”, “predictive analytics”, “data universes, cubes” and “artificial intelligence” - they have all had their turn as terms in the healthcare data spotlight. And yet, the massive output of terabytes and petabytes of data from EHRs continues to outpace the production of useful predictive and actionable insights gleaned from it. While improvements have been made in aggregating and normalizing data into centralized repositories, that collection goes only so far if it can’t help facilitate next steps in terms of activities or behavior change. Raw data needs to be transformed so that it can guide and improve clinical care. What to do, in what sequence, for whom, and by whom are all important questions that have significant operational and technical implications. While new advancements and innovation in fields such as data science are increasing, we remain a long way off from bridging the gap between data collection and rich insights that result in robust prescriptive action steps which improve care quality and outcomes while reducing the total cost of care.
3. Proactive, Personalized Care Management
Care management functionality is expressed differently across EHR vendors. Care plans themselves are also more clinically focused within EHR systems. Multiple types of care plans with duplicative data often exist within a single EHR, whether they are geared towards social determinants of health or certain specialties. For example, oncology care plans combine treatment protocol data with other, more traditional care plan data. In certain instances, care plan workflows and data are often lacking structure and only exist as static data residing in a .pdf document.
Additionally, care plans (from a workflow perspective) are often developed once as a single snapshot in time, making them difficult to update and manage on an ongoing basis. These care plans may then be buried somewhere in the patient record, making them difficult to view or access by and across an interdisciplinary care team. In order to be truly effective, care management must be an ongoing process of actively managing the whole person across time rather than in fragmented care episodes. This care should be personalized with a discrete set of detailed problems, goals, and interventions assigned to specific resources who are most qualified to address clinical, behavioral and social service needs.
An IT platform needs to not only enable this functionality, but also meet a set of robust data requirements and facilitate real-time access and interoperability so that all members of a care team—including patients—are able to work together across care settings and organizational boundaries.
EHR Strategy for Population Health Management: Know Thyself
Providers across the country have spent considerable financial and organizational resources to implement enterprise-wide EHRs and are looking to maximize these IT assets. All of the major EHR vendors have been investing heavily over the past several years in their offerings for value-based care and population health management. These include partnerships that major vendors such as Epic, Cerner and Allscripts have secured in key areas like predictive modeling, or through internal development of specific offerings such as Epic’s Healthy Planet suite or Cerner’s HealtheIntent platform. In the meantime, the majority of healthcare organizations have begun to dip their toes into value-based care with different risk-based contracting initiatives involving varying levels of capitation or savings / losses.
There are many different flavors of risk-based arrangements that have significant operational and IT implications for enabling population health management initiatives. Understanding one’s own unique “portfolio” of value-based arrangements and the specific details of each arrangement will help elicit key business and functional needs. These needs can then be mapped against one’s own EHR and then the organization’s broader IT ecosystem before seeking external vendor alternatives.
Insight into key process, workflow, and data requirements will be important in analyzing and designing new solutions or in optimizing current EHR deployments. Beginning with some of the key principles outlined above regarding interoperability, comprehensive data sets, strong analytics, and proactive care management, providers should attempt to align their PHM and care management needs with what their current EHR vendors are offering.
Is Good Enough Just Perfect?
While there is no shortage of new PHM, analytics, and care management vendors flooding the marketplace, EHR vendors can often accomplish a number of functions that third-party standalone applications are also trying to address. Many of the new products being introduced are offering point solutions so the inevitable hurdles regarding access and workflow / data integration will ultimately surface. Therefore, enterprise EHR vendors are at times better positioned to capitalize on their large technical and data footprint. While there can be a tendency to seek after the next shiny IT vendor solution, it is always important to weigh key criteria such as integration versus functionality when making important IT portfolio decisions. While EHRs are not perfect, they are equipped to handle many of an organization’s PHM and care management needs.
Still, the slower pace of development of EHR vendors remains a significant issue as a large installed customer base creates some inflexibility and impedes agility just naturally as a by-product of size and scale. In some cases, there must be a willingness to accept some limitations in order to reap the benefits of a more integrated IT environment and consolidated patient profile along with a more seamless end-user experience. In other instances, the benefit of acquiring another niche solution from a more responsive and agile vendor may well outweigh any integration challenges. These are major strategic portfolio decisions that will need to be carefully considered and uniquely facilitated by each organization as our overall healthcare system continues to evolve and transform over the next 5 to 10 years towards value and outcomes.
