HIMSS Overhauls Digital Health Scoring System to Reckon With AI's Reach

For the people who run hospitals, and increasingly for the patients who rely on them, a simple question has become harder to answer: is this health system actually ready for AI, or does it just look ready on paper? That question sits behind a quiet but consequential update to one of healthcare's most widely used digital maturity scorecards.

HIMSS has previewed the second version of its Digital Health Indicator, a tool that hospitals and national health systems around the world use to gauge how far along they are in building digital capacity. The update, unveiled at HIMSS26 APAC, expands the model from four dimensions to five, with artificial intelligence woven into every one of them rather than confined to a single category.

Think of the DHI as something like a health inspection checklist, but for a hospital's digital nervous system instead of its kitchen. Since it launched in 2020, following initial development in 2018, it has scored organizations on a scale of 0 to 400 across governance and workforce, predictive analytics, interoperability, and person-enabled health. Health systems have used it as a strategic roadmap and, just as importantly, as a way to benchmark themselves against peers worldwide.

That four-part structure served its purpose. But Dr Anne Snowdon, chief scientific officer of HIMSS, says the ground has shifted enough that the old framework can no longer capture what is actually happening inside health systems. "It served us very well, but now it needs to be updated and modernised," she said during the DHI 2.0 preview.

Splitting Out the AI Workforce

The most notable change is the decision to carve "AI-enabled workforce" out of the existing governance and workforce dimension and give it a standalone place in the model. It is a small structural move with a big underlying message: AI is not just another tool bolted onto existing workflows. It is changing how clinical teams operate, how facilities connect with one another, and how leaders make decisions.

"We have split out the AI-driven workforce because AI now is really transforming work environments, the connectivity of environments across settings, governance and leadership, and how we lead digital transformation," Dr Snowdon explained. Given "the complexity, variability, and, to be quite frank, uncertainties around AI," she added, it earned the right to stand on its own.

That new dimension will look at whether clinical staff actually trust the AI tools placed in front of them, and whether those tools are usable in the messy reality of daily practice. A brilliant algorithm that nobody trusts enough to use is, in practical terms, no better than no algorithm at all. That is the kind of gap the updated model is designed to surface.

The governance and leadership dimension, now freed from workforce concerns, will sharpen its focus on AI strategy, leadership accountability, lifecycle monitoring, and risk management, with more weight given to cybersecurity and system resilience. Meanwhile the analytics dimension will examine how AI gets folded into clinical settings and whether it actually advances an organization's broader transformation goals and overall health system performance.

Underneath all of this sits a data infrastructure and interoperability layer that HIMSS expects to emphasize more heavily than before. The updated model is expected to look closely at how trusted data moves across facilities, how accessible and intact that data remains as it travels, and how organizations protect privacy and security along the way.

That emphasis is not arbitrary. Dr Snowdon said HIMSS' recent work with health systems shows that data and digital infrastructure lag noticeably behind governance as organizations scale up their AI use. Hospitals have gotten reasonably good at building interoperability within their own walls. The harder, still-unsolved challenge is letting data flow across facilities and follow a patient through their entire care journey, from primary care to specialist to hospital and back again.

There is also a subtler shift happening in analytics itself. AI is making this work increasingly automated, but automation brings its own headaches around accuracy over time. "It's not like a regression model where you know exactly what your confidence intervals are," Dr Snowdon said. "AI is constantly learning depending on the data it's exposed to, so you've got to constantly ensure that data is accurate." A traditional statistical model behaves predictably once it's built. An AI system keeps evolving as it absorbs new data, which means the checks that worked yesterday might not catch tomorrow's problem.

Notably, HIMSS is not turning DHI 2.0 into a mandate for how organizations should build infrastructure or roll out AI. Dr Snowdon was clear that the framework assesses how those capabilities get used to advance an organization's own priorities and performance goals, not whether an organization has adopted AI at all. "If you're not advancing AI, it doesn't mean you won't have a high score on this model," she said, a reassurance aimed at health systems wary of feeling pressured into technology they aren't ready for.

Dr Snowdon noted that health systems across the Asia-Pacific region have been especially effective users of the original DHI framework, calling their approach to digital transformation strategy "exceptionally advanced and impressive." That regional track record makes the APAC preview a fitting venue for the update, and it signals that the revised model will likely see close scrutiny from health systems already using it as a planning tool.

Why It Matters

A day before the DHI 2.0 preview, HIMSS also announced its first partnerships to build an AI Outcomes Framework, described as the world's first longitudinal, multi-institutional effort to measure AI's real impact, value, and return on investment in healthcare. Dr Snowdon said that parallel project will help document and measure outcomes, giving health systems concrete evidence to guide how they deploy AI rather than adopting it on faith.

Together, the two initiatives point toward a broader reckoning in digital health policy. Scoring frameworks and outcome measures might sound like bureaucratic housekeeping, but they shape which technologies get funded, which get scaled, and which get quietly shelved. For patients, that translates into whether the AI tools touching their care have actually been vetted for trustworthiness, or simply for novelty. Getting that distinction right, at scale, may matter more than any single algorithm.