Technology is evolving faster than we can optimally implement it. The electronic health record (EHR) is a great case study where health systems race to keep up with continuous updates and features. Yet, beneath the surface lies the accumulation of outdated customizations, duplicative logic, and redundant workflows. Are health systems appropriately accounting for the financial and safety impact of this technical drag?
In IT organizations, technical debt refers to the additional development effort required to overcome outdated or inefficient technical implementations from the past. A recent McKinsey article highlights that organizations often pay 10–20% more on projects due to technical debt and it accounts for “40% of IT balance sheets.” Another McKinsey study highlighted that companies can see up to 60% of each IT dollar spent addressing technical debt. They estimate that addressing this debt can free up as much as 50% of engineering time.
Similarly, the IT arm of healthcare organizations continuously build new content (e.g., order sets, alerts, new upgraded features, etc) and customize tools to fit local workflows. While standardized “out of the box” solutions are attractive, they inevitably fall short in accommodating the unique needs of an individual health system. Every hospital, clinic, and clinician operates within a distinct context, shaped by specific patient populations, regulatory environments, and institutional cultures. Therefore, a degree of customization is not just preferred—it's necessary especially when we’re seeing rampant clinician frustration with the EHR as highlighted by the KLAS Arch Collaborative.
Order sets, alerts, and other workflow optimizations are tailored to align with local practices. When done well, these customizations enhance clinician adoption and lead to improved patient outcomes. However, each modification adds a new layer of complexity to the EHR system. Over time, these layers accumulate, gradually slowing down the agility and flexibility of the IT and informatics teams.
Furthermore, medical content has the additional challenge of requiring constant alignment with rapidly evolving medical research, guidelines, and best practice. This could be akin to new technologies that IT companies need to keep up with. If health systems don’t curate and review this alignment regularly, it risks encouraging clinicians to adhere to outdated practices and the delivery of suboptimal care.
No surprisingly, the longer a health system operates its EHR, the more this technical debt piles up. Based on my discussions with some health systems colleagues, the accumulation of this EHR technical debt can actually be one of the (very few) impetuses for considering a switch in EHR vendors. Imagine saying, “my home is too cluttered, I should evaluate if buying a new house is more cost effective than cleaning it up!” Other than a new EHR system, vendors like Epic Systems can support a Refuel initiative that tries to realign health systems back towards the “out of the box” system.
Standing in the way of this accumulation of clutter are IT and small informatics teams tasked with orchestrating the continuous intake and curation of digital workflows in the EHR. They diligently integrate the latest medical guidelines, regulatory requirements, and institutional policies. Yet, they rarely have the bandwidth to evaluate and de-implement existing tools that no longer serve their intended purpose.
De-implementation is an onerous task, often more difficult than implementation. It frequently requires robust data accessibility, stakeholder buy-in, and a deep understanding of the interdependencies within the EHR system. Consequently, outdated or redundant tools remain embedded, creating a drag on IT agility and increasing the risk of safety and quality issues.
Moreover, as value sets, rules, groups, and triggers continue to expand, the EHR system becomes increasingly intricate and prone to malfunctions. Each new addition, while beneficial in isolation, adds to overall complexity. This web of code can hinder performance of the IT team, obscure critical insights, and make the system less responsive to new needs.
Addressing technical debt in EHR systems isn't just about housekeeping; it's about ensuring the system remains flexible and reliable when tasked with delivering intricate patient care. Clinical informatics and IT teams need to allocate adequate time and resources to identify this technical debt and allocate appropriate resources to periodically de-implement workflows that provide no value.
This process is challenging for several reasons:
The technical debt creates a drag and impedes rapid implementation of new features or workflows, as each change must navigate the tangled web of existing customizations. This oftentimes leads to duplicative implementations as it’s easier to create a new EHR element rather than try to adopt existing ones. Furthermore, this introduces reporting gaps and risks to patient safety and care quality as tools may trigger off of incorrect fields or not comprehensively cover a targeted value set.
The parallels between software technical debt and the accumulation of obsolete customizations in EHR systems are striking. Both lead to increased costs and reduced agility. Importantly, there are real risks to reporting, clinical quality, and patient safety when applied to the EHR. However, by acknowledging these challenges and proactively addressing them, organizations can unlock significant efficiencies.
Investing time in de-implementation is not merely a housekeeping exercise; it's a strategic imperative. Just as paying down technical debt in software development can free up engineering time and resources, streamlining the EHR can enhance performance and support better patient outcomes.
A data-driven approach, supported by robust knowledge management systems, empowers small informatics teams to make informed decisions. Here's how:
By systematically identifying and removing redundant components, organizations can reduce costs over the long run, enhance the speed of prospective projects, and protect against clinical risks (and extra clicks!).