By Maithilee Mitra, Director – Healthcare Analytics, Information Technology, SBH Health System
LinkedIn: Maithilee Mitra
LinkedIn: SBH Health System
It started with a cloud infrastructure project that should have been straightforward. An enterprise IT team was tasked with building a cloud environment for researchers and data scientists, call them data users, to analyze findings and contribute to a multi-site data warehouse. The project had executive support, a clear timeline, and dedicated resources. By most enterprise measures, it was a two-to-three-week task. The engineers were eager to work on something closer to the institution’s mission. They delivered. Secure HIPAA guardrails. A free data science image from the cloud marketplace. Multi-factor authentication. A machine sized to the budget, with 5 GiB of storage. On time, on budget.
But when the data users arrived, they found something different. The account was difficult to access. The tools were unfamiliar. Data retention policies didn’t align with compliance requirements. And the pipeline to push data upstream to the multi-site warehouse, the entire point of the effort, was cumbersome. Nobody had failed. And yet, everyone had.
The Gap Nobody Talks About
This scenario plays out across universities, hospitals, payer organizations, and integrated delivery networks. IT builds infrastructure. Data users show up and can’t use it. Frustration accumulates on both sides, and almost nobody asks why. Data users will tell you the problems feel obvious: they wait weeks for access they need today. When data arrives, there’s no documentation, no data dictionary, no context for what the fields mean or how they were gathered. And even when access is fast and documentation exists, the data is often in the wrong format for real analysis. The platform itself is often the wrong fit too. End users can’t bring their own devices, even when they have the budget, because IT won’t support them. These aren’t small inconveniences. They are the difference between a team that generates insight and one that generates tickets.
But IT Isn’t the Villain
Enterprise IT is not failing because of indifference or incompetence. The decisions that frustrate data users are almost always made for legitimate reasons. Risk and compliance requirements are real, health data is among the most regulated in any industry, and IT carries accountability when something goes wrong. Legacy systems constrain feasibility; most health IT environments are built on decades of infrastructure not designed for modern analytics. Competing priorities mean IT is constantly triaging, weighing a data user’s urgent request against clinical system uptime, security patching, and regulatory deadlines. And often, IT simply doesn’t know what data users actually need, not from indifference, but because no one built a process to surface it. IT is managing risk. Data users are trying to generate value. Both are doing exactly what they should. The problem is that no one is translating between them.
A Translation Problem, Not a Technology Problem
The friction between IT and data users is not primarily a technology gap, it’s a communication and process gap. The tools to solve most of these problems already exist. What’s missing is the bridge: someone who understands both worlds well enough to design systems, processes, and partnerships that serve both. That bridge looks different in every organization. It might be a data governance committee that includes data users, not just IT and compliance. A self-service data access model that reduces ticket dependency without sacrificing security. Or something as simple as a shared data dictionary that IT maintains and data users actually trust.
But it starts with both sides acknowledging something uncomfortable: the problem isn’t the other team. It’s the gap between them. And if anyone is pointing fingers, both parties have earned the right to point first. Closing that gap requires each side to understand the other’s priorities, and to ask how both can be upheld while each team still reaches its goals. Three pillars matter here: process, skills, and the willingness to look beyond unit goals.