Are We Witnessing a Sunset of Health Informatics Defeated by Health Data Science? Hold On — There Are Solutions

Nikolay Lipskiy, MD, PhD, MBA

October 23, 2025

Health Informatics (HI) job postings are disappearing while Health Data Science (HDS) keeps booming. Even roles labeled "informatics" now list tasks centered on analytics, dashboards, and AI pipelines.

In my forthcoming academic paper, I use BERT-based semantic analysis to show how today's definitions of HI and HDS have become nearly indistinguishable — and why that linguistic convergence fuels confusion in hiring and management.

🧩 When Definitions Converge

A classic 1997 description defined informatics as:

"Health informatics, the science that deals with health information, its structure, acquisition, and use." 👉 (Source: J Am Med Inform Assoc, PMCID PMC61347)

At the time, that sounded revolutionary. But today, the same sentence could easily describe Health Data Science. Both speak of data, structure, and use.

Using SBERT (BERT semantic embeddings), I measured the similarity between modern definitions of HI and HDS. Result? ➡️ A high cosine score (> 0.75) meaning the two concepts now occupy almost the same semantic space. No wonder hiring managers see them as interchangeable.

This isn't just subjective observation — it's measurable linguistic convergence.

⚙️ From Systems Governance to Data Analytics

Historically, Health Informatics was the systems-governance discipline behind digital health. It focused on:

• people, processes, and infrastructure,

• interoperability and data quality,

• ethical and lifecycle governance.

Health Data Science, on the other hand, is about modeling, inference, and prediction — turning curated data into statistical evidence and insight.

🔹 HI built the infrastructure. 🔹 HDS mined the data.

Today those boundaries blur. Data scientists are asked to manage interoperability; informaticians are told to "build dashboards." BERT shows the linguistic proof: their word spaces are converging.

📊 What the Data Reveal

My ongoing analysis confirms that the conceptual distance between HI and HDS has narrowed dramatically. In natural-language space, "informatics," "data science," and "analytics" now activate the same clusters — information, algorithms, prediction, visualization.

In plain terms: When hiring managers searching for talent, the words they use for HI and HDS now trigger the same results. The fields have merged linguistically even if they haven't merged functionally.

That convergence reshapes how organizations recruit, educate, and even imagine expertise. In the eyes of many, HI has become "just another analytics function."

⚠️ Why This Matters

1️⃣ Loss of governance expertise. Without dedicated HI professionals, the orchestration of people + process + infrastructure collapses into ad hoc data management.

2️⃣ Overextension of data science. Expecting HDS teams to handle architecture, ethics, and interoperability creates blind spots and risks.

3️⃣ AI fragility. Analytics without governance means models without accountability — and patients without protection.

🧠 Re-centering the Definition

Through our semantic and theoretical analysis, we refined a clearer definition to restore focus:

Health Informatics (HI) is the systems-governance discipline that designs, manages, and evaluates the socio-technical architectures enabling the trustworthy use of health data.

Grounded in scientific inquiry and practiced through interprofessional collaboration, HI ensures that health-information ecosystems — comprising people, processes, and technologies — are interoperable, high-quality, and ethically governed throughout their lifecycle.

A full comparison with HDS and detailed BERT metrics will appear in my upcoming academic publication.

🤖 Why HI Still Matters — Especially Now

As AI becomes integral to healthcare, governance and ethics move to the center. Health Informatics professionals uniquely manage:

• data provenance and metadata,

• interoperability strategy,

• human-AI workflow design,

• ethical and regulatory compliance,

• usability and accountability.

💡 If HDS builds insight, HI builds trust. One explains what data means; the other ensures data means what they should.

🔄 AI Governance and the Rebirth of HI

The rise in AI will revive, not replace, Health Informatics. Tomorrow's informaticians will serve as intermediaries between:

👥 people (patients, clinicians, data scientists), 🖥️ infrastructure (EHRs, IoT, cloud systems), ⚙️ processes (quality, integration, outcome measurement).

They will safeguard trustworthy, AI-driven, patient-centric ecosystems.

🧭 Five Actions to Reclaim and Strengthen HI

1️⃣ Educate leaders — HI ≠ analytics; emphasize governance and systems thinking.

2️⃣ Modernize job roles — highlight interoperability, data quality, AI governance.

3️⃣ Use semantic tools — apply BERT or SBERT to clarify terminology and avoid overlap.

4️⃣ Promote collaboration — build joint HI-HDS governance teams.

5️⃣ Lead AI trustworthiness — position HI at the center of ethical AI design and policy.

🌇 In Closing

What feels like a sunset may only be an eclipse. Health Informatics hasn't been defeated; it's been linguistically overshadowed.

As AI spreads, the need for governance, ethics, and interoperability will bring HI back into the light. The challenge is to reclaim the language, redefine the mission, and rebuild recognition of HI as the discipline that makes health data not just available — but trustworthy, interoperable, and meaningful.

✳️ Author's Note

I am currently preparing a peer-reviewed academic article that details the BERT-based semantic analysis underlying this discussion. The study systematically traces how definitions of Health Informatics and Health Data Science have converged linguistically over the past three decades — and proposes a framework to restore conceptual and professional clarity.

I welcome collaboration, comments, and perspectives from colleagues across healthcare, data science, and informatics. If you're working on similar challenges in AI governance, data interoperability, or workforce evolution, please reach out or share your insights below — let's advance this dialogue together.