See an essay at Linked In by the prolific Alex Bendersky on whether it's realistic to expect good results when we dump fine innovatios into a chaotic "ecosystem" of healthcare stakeholders.
Essay here.
His take-off point is a new article by Reed, Svedberg, Nygren, in J Med Internet Res, called:
Enhancing the Innovation Ecosystem: Overcoming Challenges to Introducing Information-Driven Technologies in Health Care.
Find the article open access here.
Abstract by authors:
As health care demands rise and resources remain constrained, optimizing health care systems has become critical. Information-driven technologies, such as data analytics and artificial intelligence (AI), offer significant potential to inform and enhance health care delivery at various levels. However, a persistent gap exists between the promise of these technologies and their implementation in routine practice. In this paper, we propose that fragmentation of the innovation ecosystem is behind the failure of new information-driven technologies to be taken up into practice and that these goals can be achieved by increasing the cohesion of the ecosystem.
Drawing on our experiences and published literature, we explore five challenges that underlie current ecosystem fragmentation: (1) technology developers often focus narrowly on perfecting the technical specifications of products without sufficiently considering the broader ecosystem in which these innovations will operate; (2) lessons from academic studies on technology implementation are underused, and existing knowledge is not being built upon; (3) the perspectives of healthcare professionals and organizations are frequently overlooked, resulting in misalignment between technology developments and health care needs; (4) ecosystem members lack incentives to collaborate, leading to strong individual efforts but collective ecosystem failure; and (5) investment in enhancing cohesion between ecosystem members is insufficient, with limited recognition of the time and effort required to build effective collaborations.
To address these challenges, we propose a series of recommendations: adopting a wide-lens perspective on the ecosystem; developing a shared-value proposition; fostering ecosystem leadership; and promoting local ownership of ecosystem investigation and enhancement.
We conclude by proposing practical steps for ecosystem members to self-assess, diagnose, and improve collaboration and knowledge sharing. The recommendations presented in this paper are intended to be broadly applicable across various types of innovation and improvement efforts in diverse ecosystems.
## #
AI CORNER
###
Reed et al. frame their article as a Viewpoint rather than an empirical study, and that choice is deliberate: their core claim is conceptual rather than evidentiary. They argue that the persistent failure of information-driven technologies—especially analytics and AI—to take root in routine health care is not primarily a problem of immature algorithms or inadequate data, but of innovation ecosystem fragmentation. Drawing explicitly on Ron Adner’s “wide-lens” theory of innovation, they contend that health care innovators focus too narrowly on their own products and pipelines while neglecting the dense web of co-innovators, adoption-chain partners, professional norms, incentives, and institutional constraints required for success. The paper’s central move is to recast implementation failure as a collective coordination problem: developers, researchers, clinicians, regulators, and funders may each be “doing their jobs well,” yet the system as a whole fails because no actor is structurally responsible for aligning these interdependencies.
The authors organize this argument around five recurring failure modes. Developers over-optimize technical performance while under-investing in usability, workflow fit, and downstream practice change; implementation science produces rich theoretical insights that are poorly translated into actionable guidance; clinicians’ perspectives are invoked too late and too superficially; incentives reward siloed excellence rather than cross-boundary accountability; and almost no one is funded or protected to do the slow, relational work of ecosystem coordination. The result is a familiar pattern in digital health: impressive proofs of concept that stall at scale, blamed alternately on “resistant clinicians,” “immature evidence,” or “slow systems,” when the deeper issue is misaligned expectations about who must do what for value to materialize. Reed et al.’s strength is showing how these are not isolated frictions but mutually reinforcing structural features of the innovation ecosystem itself.
What is most promising in this article is its synthesis rather than its novelty. Reed et al. integrate ecosystem strategy with established health-care literatures on implementation, sociotechnical systems, and improvement science, translating them into a coherent diagnostic narrative. Their critique of the linear “innovation pipeline” and the misleading metaphor of the “last mile” is particularly persuasive. By reframing innovation as an iterative cycle of system improvement rather than a handoff from inventor to user, they explain why health-care AI repeatedly disappoints even when accuracy metrics look strong. Equally valuable is their insistence that coinvention—changes to workflows, roles, authority, and accountability—is where most value is actually created, yet is systematically under-resourced and orphaned across organizations.
The most squishy or problematic elements lie in the proposed remedies. Calls for “ecosystem leadership,” “shared value propositions,” and “protected spaces” for boundary-spanning work are normatively attractive but operationally vague. Who, in practice, is empowered to act as ecosystem leader in pluralistic health systems characterized by fragmented governance, heterogeneous payers, and competing commercial incentives? Health care lacks an Amazon-like focal firm capable of underwriting coordination costs, and the article does not fully confront that structural reality. Similarly, the emphasis on participatory diagnosis and local ownership is well grounded in complexity theory, but risks becoming an unfalsifiable prescription: when adoption fails, one can always argue that ecosystem work was insufficiently deep, inclusive, or iterative.
There is also an unresolved tension between descriptive realism and prescriptive optimism. The authors convincingly explain why current incentive structures discourage ecosystem work, yet largely assume that senior leaders can or will realign those incentives once the problem is named. For readers attuned to regulatory, reimbursement, and capital-market constraints, this may feel underpowered. The paper diagnoses the pathology with clarity but offers interventions that depend more on cultural change than on enforceable mechanisms.
In sum, Reed et al. provide a strong conceptual lens for understanding why information-driven health technologies so often fail to deliver on their promise. Their framework is most useful as a diagnostic map—a way to surface hidden dependencies and misplaced assumptions—rather than as a ready-to-deploy playbook. The article’s lasting value will likely lie in helping innovators ask better questions earlier, even if its solutions remain aspirational rather than fully specified.
###
Sidebar:
Ron Adner's Wide Lens
Ron Adner is a leading strategy scholar best known for developing innovation ecosystem theory—the idea that an innovation’s success depends on coordinated progress across multiple interdependent actors, not just the focal firm. He is the author of The Wide Lens: A New Strategy for Innovation (2012), which popularized this framework, and Winning the Right Game (2021), which extends it into a practical guide for strategic choice; his widely cited academic article “Ecosystem as Structure” formalized these ideas for management research.
Adner is a professor at the Tuck School of Business at Dartmouth College, where he has been on the faculty for many years
