Twenty years ago, Bala Balasubramanian was flying around the world for a big-pharma sponsor trying to gather seven basic pieces of information about the company's own products. Generic name. Trade name. Dosage form. Strength. Country. Status. Approval date. It couldn't be done at scale. He told leadership so at the time — and since then, he's spent nearly two decades trying to build the thing that would make it possible. Bala is the subteam lead for the DIA RIM Reference Model — now in Version 2.0 — and one of the industry's clearest voices on why regulatory operations still treats data as an afterthought, and what it will finally take to change that. Matt Neal sits down with him to trace the arc from a single unusable Excel workbook to a real conceptual data model, why the discipline problem has never been a systems problem, and what changes when data starts being treated the way finance and clinical already treat it: as core to the job, not adjacent to it.
V. "Bala" Balasubramanian, PhD, MBA, is a strategic advisor in healthcare and life sciences and the subteam lead for the DIA RIM Reference Model. He has been associated with DIA for more than 14 years and also leads the DIA AI Consortium Regulatory Frameworks and Terminology workstream.
Bala spent his earlier career inside big pharma, then led the Healthcare and Life Sciences Industry Solutions Group at Orion Innovation as Senior Vice President before moving into independent advisory work. Across that arc — sponsor, vendor, standards leader — he has been consistently focused on the same problem: shifting regulatory affairs from a document-centric practice to a data-driven one. He's a co-author on the DIA RIM white paper (V2.0) and the RIM Reference Model V2.0 conceptual data model that came out of it.
The seven pieces of information. Bala opens with the story that has stayed with him for twenty years: circa 2004–2005, working for a global sponsor that couldn't reliably answer seven basic questions about its own products across markets. Each affiliate tracked things its own way. Free-text fields everywhere. "This is humanly impossible" — his message to leadership at the time. The problem was never system availability. It was data discipline.
How the RIM Reference Model actually got built. The origin story: a DIA working group formed around 2015, but it stalled on the definitional question — what is RIM? — before it could even talk about data elements. Bala joined the effort in 2018 as a second workstream launched alongside the white paper. Version 1.0 (released 2022–23) was a single, comprehensive Excel workbook that became unusable through filter-fatigue. The team pivoted to a proper conceptual data model — an entity-relationship diagram, objects and attributes catalogued on individual tabs, definitions and controlled vocabularies aligned where possible with IDMP and FHIR. That's Version 2.0.
Not a standard — a common terminology. Bala is careful to define what the reference model is and isn't. It's not a mandated standard. It's a platform-agnostic common terminology and taxonomy that industry can pick up as a starter kit for a new RIM implementation, as a backbone for a system migration, or as a comparison layer in M&A. Two vendors have told the team their own offerings are subsets of it. Sponsors have used it to build their own regulatory product hubs.
Data citizenship — treat it like money. Bala's most repeated line in the episode: in 40 years of banking in America, he's never received a wrong bank statement. Financial institutions reconcile trades every night because the discipline is embedded. Clinical does the same because trial data has to be pristine. Regulatory has been the outlier — filing today, entering the record next week — because the field has "gotten away with" treating data as secondary. That has to end.
Why performance objectives matter. Bala's structural fix: put data quality in the performance objectives of regulatory professionals. Not aspirational, not project-scoped — actually tied to review and incentive. It's uncomfortable, but he's clear that without accountability at the individual level, discipline never sticks.
Matt's baton metaphor — and the 10-to-12-year horizon. Matt's contribution to the discipline problem: a product lives 10 to 12 years minimum. Ownership passes person to person across that arc. Institutional memory has to survive the handoffs, and that requires each individual steward to keep the record current in real time — not just for themselves, but for whoever holds the baton next.
Configuring systems to fit the lifecycle. A recurring failure mode Bala flags: old RIM systems required 33 attributes on a single screen before you could save the record, without regard for what was actually knowable at that point in the product lifecycle. Progressive disclosure — capture what's available now, add fields as the record matures — plus cascading controlled lists, are table-stakes today. Matt notes a very recent feature enhancement in one major RIM platform that finally lets records start in a draft state even when downstream fields are required. In 2026.
The transformation-program cycle. Bala names a pattern anyone who has lived through a data-remediation program will recognize: three-year initiatives, big teams, big budgets, evangelists appointed. The program ends, reorgs hit, champions move on, resources shrink, data drifts, and the next remediation project starts. Fix: embed evangelism in the operational team — not the program team — so change management becomes implicit and continuous, not episodic.
When health authorities set the new bar. Matt's observation: inspectors have caught up. They now walk in expecting real-time answers from the RIM system and the TMF system, and they'll give an hour instead of days. That external pressure is doing what internal accountability couldn't.
The circle of affinity. Bala's frame for regulatory's place in the org: safety, PV, manufacturing, and labeling all depend on regulatory data being right. Regulatory sits at the center of that circle whether it wants to or not. Owning that role — and being visible in it — is the way ops earns its seat at the strategic table.
Prediction: data-driven submissions arrive, split focus is the risk. The rise of data-centric submissions is real. IDMP, structured labeling, structured CMC, structured clinical reports — all pushing industry toward MDM in R&D for the first time. But the volume of change is huge, resources are constrained, and outsourcing to captive centers can go transactional and lose data quality. His call: keep the discipline attached wherever the work goes.
"In the 40 years I've lived in America, I've never gotten a bank statement that's been wrong."
"As long as data is not treated equal to money, we'll have this situation in regulatory."
"It's not a systems issue. It's a discipline issue."
"For some reason, regulatory always has this afterthought mentality."
"You need to have data as one of your performance objectives — and either get incentivized or penalized based on that."
"In some ways, we've come full circle: we went document-centric, and now we're going back to data."
Regulatory operations, regulatory affairs, and data-governance leaders responsible for RIM strategy or implementation; system owners planning a RIM migration or M&A integration; RIM vendors benchmarking product coverage; and anyone still trying to make the case that regulatory data discipline is worth investing in.
Guest & Related Work
DIA RIM Reference Model — Downloads
Standards & Regulatory Frameworks Referenced
Concepts Referenced
Transcript provided by Otter.ai.