5 reasons health system innovation struggles to reach patients

Most health systems already know how to talk about innovation. They run pilots, build venture partnerships, and launch AI initiatives meant to improve clinical care. At the May CHIME Innovation Summit in Jacksonville, leaders focused less on launching innovation and more on why so little of it sticks.

Over two days of fishbowl panel discussions, where audience members rotated into open chairs to join the panelists, summit participants kept returning to the same tension. The hardest part of innovation is not developing technology, but absorbing change, tolerating experimentation, and aligning people around new ways of working. Five themes surfaced repeatedly throughout the summit.

Health systems buy innovation instead of developing it

Several conversations drew a sharp distinction between buying innovation and building the ability to innovate from within. Health systems often purchase polished enterprise platforms developed elsewhere, leaving clinicians and staff to absorb the disruption after rollout. The technology itself may function well, but organizations often struggle to adapt it to existing workflows or prepare staff for the operational change around it.

The idea of “co-creation” generated a lot of energy during the summit, with supporters heralding the practice of clinical and operational teams working directly with vendors while products are still under development. Panelists pointed to weekly workflow conversations, iterative redesign, and feedback loops meant to reduce friction before broader rollout. Co-creation also reframes innovation teams as operational partners helping shape tools around frontline realities instead of evaluating finished products after the fact. This process is slower and more resource-intensive early on, but leaders argued it makes clinicians far more likely to embrace new platforms.

AI works best redesigning workflows, not replacing clinicians

The summit’s AI discussions were noticeably more restrained than the broader healthcare market conversation. Rather than framing AI as a replacement for clinicians, summit participants pointed to its ability to take low-value work off clinicians’ plates and free up more time for the relational, judgment-driven parts of care. Routine prescription refill workflows came up as one place organizations are already testing automation.

Conversations around AI automation quickly shifted toward governance, oversight, liability, monitoring, and how organizations should respond when an AI system drifts from expected behavior or introduces subtle clinical errors into a workflow. One point surfaced repeatedly: automating a flawed workflow with AI does not solve its underlying problems.

Several leaders described AI adoption as fundamentally different from adding another software platform because it often requires redesigning the workflow itself. Skipping this redesign step can break workflows already functioning well or worsen the flaws in those that are already broken.

Collaboration problems start with incentives, not technology

Panelists rarely framed interoperability and cross-functional silos as purely technical problems. Healthcare organizations, the argument went, often ask teams to collaborate structurally while rewarding them individually. Departments and business units are frequently measured against different financial and operational targets, leaving little practical incentive to share ownership of broader system problems.

Leaders pointed to the long-standing disconnect between providers, payers, operational leaders, and clinical teams, all working under different definitions of success. In that environment, collaboration depends heavily on individual relationships and executive intervention instead of organizational design.

One recurring argument held that silos are usually a symptom, not the core issue. Panelists described the deeper challenge as alignment: building systems where the easiest operational decision is also the right one for patients, clinicians, and the organization. The discussions also framed healthcare work in broader terms than job descriptions, with organizations functioning best when operational, technical, and clinical staff see themselves as serving the same patient-centered mission rather than “protecting their own territory.”

“Failing fast” is encouraged in theory and punished in practice

The phrase “fail fast” came up repeatedly, though many conversations exposed how difficult that mindset remains to sustain inside large health systems. Summit participants described organizations that publicly encourage experimentation while internally creating strong incentives to avoid visible failure. In practice, that pushes teams toward safer, incremental projects instead of riskier redesign. 

Several exchanges contrasted the timelines of startups and health systems. Startups operate under compressed timelines and make decisions quickly, while health systems often move through extended legal, compliance, procurement, and governance reviews that slow experimentation and delay projects. Some of those safeguards are necessary where privacy, security, and patient safety carry real consequences, but several leaders argued that healthcare organizations make innovation harder than it needs to be with overly burdensome review cycles.

Panelists stressed that defining the problem first, identifying success metrics early, and piloting initiatives before scaling them can help organizations cut through unnecessary bureaucracy. Without that structure, several participants argued, organizations struggle to distinguish meaningful improvement from temporary enthusiasm around a new tool.

The system assumes patient priorities instead of asking

Some of the summit’s most reflective discussions focused on the assumptions health systems make about what patients actually want. Summit participants described environments that default toward institutional priorities, escalation of services, or operational efficiency without fully understanding whether those choices align with patient priorities.

Innovation that benefits the organization is not automatically innovation that improves the patient experience. The discussions ultimately suggested that innovation fails less from a lack of ideas than from organizations struggling to align technology, incentives, workflows, and patient priorities around the same definition of success.