Health systems need data discipline for effective AI

Robert Slepin, chief digital officer and senior vice president at SE Health

Artificial intelligence may dominate health system boardroom conversations, but Robert Slepin believes many healthcare organizations are still overlooking the prerequisite that will determine whether those investments ultimately succeed.

Before health systems can trust AI, they must first learn to trust their own data.

Slepin, chief digital officer at SE Health, one of the largest home health and community care provider organizations in Canada, argues that healthcare leaders have long managed one enterprise asset with exceptional discipline while treating another with far less rigor.

Finance leaders can typically explain the organization's cash position, its uses and its return on investment. Few CIOs, he contends, can offer an equally confident assessment of enterprise data.

"From my experience, healthcare providers do not govern and manage data assets as well as they oversee and administer financial resources such as cash," Slepin said.

That comparison frames what he sees as one of healthcare IT's most pressing leadership challenges. As organizations modernize their technology stacks and pursue AI initiatives, data governance and data architecture have become strategic business capabilities rather than back-office IT functions.

Treating data as an enterprise asset

Slepin, also a CIO advisor in the emeritus program of EHR giant Epic, views data as an asset capable of generating measurable organizational value. Properly governed, it can be transformed into information, insight, knowledge, predictions and action that improve access, affordability and health outcomes.

Without that foundation, organizations struggle to produce consistent reporting, reliable analytics and dependable decision support, he said.

He also emphasizes that governance is about more than regulatory compliance. Confidentiality protects privacy. Data integrity helps clinicians avoid harmful decisions. Availability reduces delays, duplication and unnecessary care.

Architecture complements governance by providing the blueprint for organizing sprawling clinical, financial, operational and research information across increasingly complex environments, he added.

Building an AI-ready foundation

SE Health has elevated enterprise data governance and architecture to one of the pillars of its digital transformation strategy while modernizing legacy applications and moving toward new business intelligence and AI platforms.

Slepin said users do not always agree on the authoritative source for key data elements, while decentralized analytics, spreadsheets and multiple reporting tools create inconsistent outputs. Carrying those problems into new platforms simply institutionalizes yesterday's weaknesses.

He also cautions that poorly governed data increases the likelihood of AI systems producing biased, unsafe or confidently false results.

Rather than attempting enterprise-wide perfection, SE Health established a Data Governance Council focused on practical objectives. Initial work includes clarifying HR and finance data ownership, resolving contested definitions, and building an evidence-based action plan before major system go-lives.

"Get going. Start from wherever you are. It's not a one-and-done project. Rather, it's a journey to establish data governance and architecture, and it's an ongoing process to maintain and evolve it over time," Slepin said.

He recommends beginning with a clearly defined business problem that executives already recognize as urgent. Connecting governance work to measurable organizational priorities helps secure sponsorship, builds momentum and demonstrates value early.

For CIOs, the lesson extends well beyond AI. Organizations that know what data they have, who owns it, how it is defined and how it flows across the enterprise will be positioned to make better operational, financial and clinical decisions, he said.

In Slepin's view, governing data with the same discipline applied to cash is no longer an IT aspiration. It is becoming an executive imperative.

Follow Bill's health IT coverage on LinkedIn: Bill SiwickiEmail him: [email protected]Healthcare IT News is a HIMSS Media publication.

WATCH NOW: AI design lets health systems solve organizational problems