Medical Breakthroughs

Cohere’s 10,000 HEDIS Gaps: A Workflow AI Investment Catalyst

Listen to this article · 6 min listen

We all hear the big talk about AI in healthcare, clinical breakthroughs, slashed costs. But if you’re a quality or revenue cycle leader at a health plan, the real, right-now impact isn’t so dramatic. It’s happening in the small but significant shifts inside your existing workflows. The point isn’t just about automating a few tasks. It’s about whether that automation completely changes how the work gets done, freeing up your people for more important things and, hopefully, improving patient outcomes.

The Headline: 10,000 HEDIS Gaps Closed, An Extraction Achievement

A Cohere Health case study just dropped a big number: one plan apparently closed 10,000 HEDIS gaps just by reading utilization management notes. They claim this hit a rate of 256 gaps per 1,000 members with zero manual chart abstraction. That’s the headline. Looking at their wider results with the Cohere Capture engine, they say clients are closing one in four high-value gaps, boosting Stars results by 15% per measure, and getting a 5:1 ROI, with an average of 107 gaps closed per 1,000 members with records. Frankly, those numbers are hard to ignore. For any leader who’s been stuck in the HEDIS reporting grind, the thought of automatically pulling that much proof from documents you already have is a huge relief. This is the sales pitch for vertical AI in healthcare, using specialized tools to fix processes that have always been manual, slow, and full of human error. The whole thing works because a disease-specific AI can actually read and understand the messy, unstructured data inside a UM note. This is where a specialized vertical AI tool earns its keep. A general-purpose AI platform would likely get tripped up by the specific clinical terms and data points HEDIS requires, but a focused solution is built for it. Getting that part right is everything for finding and closing documentation gaps, which has always been a nightmare for quality reporting.

Extraction at Scale: What Changes, and What Doesn’t (Yet)

So, the Cohere Health study shows you can extract data at a massive scale. But let’s be real: pulling quality evidence from notes the plan already has is a workflow change first, a technology win second. The efficiency you get from not having your team do manual abstraction is obvious. In theory, your quality people, who used to spend their days buried in records, can now focus on things like member outreach or digging into quality trends. But there’s a big piece missing from these vendor-supplied results. What happened to the quality team’s hours? What about headcount? Without knowing the staffing impact, you can’t really judge how much the workflow actually changed. Is the team just doing the old job faster, or are their roles completely different now? This isn’t a small detail. When you’re trying to calculate the ROI on a specialized AI tool, you have to know what happened to your people. The tech automates a painful task, sure, but the real prize, moving your team to higher-value work, is a total unknown in this story.

The Workflow Change and Prevention Bar: Learning from Omada Health

If you want to see what a deep workflow change and a real outcome look like, look at a company like Omada Health. For them, the recorded integration challenge is “Change Management” and the outcome is “Adverse Event Prevention.” That combination sets a pretty high bar. Omada’s work in adverse event prevention, often using AI for behavioral health, directly connects their tech to better patient health. They are actively stopping bad things from happening, which is a world away from just checking a compliance box. The “Change Management” label also tells you that you can’t just buy their platform and flip a switch. You have to fundamentally change how your organization works to get the value. That means retraining people, changing processes, and building a culture around data-driven intervention. So, how does Cohere Health’s claim of 10,000 closed HEDIS gaps stack up against Omada’s “Adverse Event Prevention”? Closing doc gaps is important for reporting and compliance, but it’s not the same as preventing a bad health outcome. We can guess that better docs lead to fewer adverse events, but that’s a leap that still needs solid proof.

The Unrecorded: Staffing Figures and Outcome Evidence

So what’s the bottom line on the Cohere Health case study? With the information we have, it’s a story about extraction at an impressive scale. It proves that vertical AI can pull data from messy clinical notes automatically. But for any quality or revenue cycle leader, the next move is to start asking for the rest of the story. Without verified numbers on staffing and actual patient outcomes, these vendor claims are only half-told. What happened to the quality team’s work hours? Did they reassign people to more proactive care? And the million-dollar question: did closing all those HEDIS gaps actually lead to healthier patients, the way Omada Health claims with its “Adverse Event Prevention”? There’s a big difference between extraction and outcomes, and it’s a gap we have to watch. Closing a documentation gap is good for your reporting and gets you paid, but that doesn’t make it a prevention result. The idea that better documentation causes better prevention feels right, but it’s just a hypothesis until there’s solid, outcome-based data to back it up, something the guidelines for evaluating AI would demand. The real worth of these specialized, disease-specific AI platforms comes from proving they can improve patient health and operational efficiency, with hard data on both workflow changes and patient benefits. That’s the standard for any AI implementation. best practices for AI implementation in health plans

Frequently Asked Questions

Does the 10,000 HEDIS gaps closed figure indicate a workflow improvement or solely automation?

The article suggests it is primarily an achievement in automation, demonstrating extraction at an impressive scale. While it eliminates manual abstraction, the article states it’s challenging to fully assess the extent of workflow change without staffing figures or recorded time savings.

What is the impact on quality team hours or headcount from this automation?

The vendor-presented case results are silent on the impact on the quality team’s hours or headcount. Without these staffing figures, it’s difficult to determine if the team’s role has fundamentally evolved or if they are merely performing the same work faster.

Does closing 10,000 HEDIS gaps translate to improved patient outcomes or adverse event prevention?

The article states that closing documentation gaps is necessary for quality reporting, but it is not, in itself, a recorded prevention result. The connection between improved documentation accuracy and reduced adverse events requires further evidence to be definitively established.

What additional information is needed to determine the true operational leverage and return on investment?

To understand the true operational leverage and return on investment, health plan leaders need more comprehensive data, specifically staffing figures and independently verified outcome evidence. This would clarify if human capital is being reallocated to higher-value activities.

Share
Was this article helpful?

Editorial Team

The Editorial Team at Vertical AI Health Leaders is a group of writers and editors who cover health, workplace wellness, and everyday care. They research each story and publish under a shared desk byline — experts in their fields, working as one team.