Leader Spotlight: Improving data accuracy for healthcare providers, with Hassaan Sohail

Hassaan Sohail is the Senior Director of Product Management at Alma, a mental health practice management platform. He has extensive experience leading healthcare technology initiatives. Hassaan previously served as Group Technical Product Manager at Optum, overseeing Cost Transparency solutions, and successfully launched the DirectAssure provider directory cleanup product at CAQH.

In our conversation, Hassaan provides an overview of the challenges today’s healthcare providers face when it comes to maintaining accurate directories and how he’s developed solutions to streamline these processes — ultimately creating clearer paths for patients to find care.

---

Thinking about the job to be done

With your extensive experience in healthcare technology, how do you approach developing a product strategy for provider-facing tools, especially when many clinicians see directory management as an administrative burden rather than part of their core responsibilities?

It’s really important to know where the need for accurate directories is coming from. It’s been a problem in healthcare for a long time, but really around 2015, some regulations with teeth came in, where payers started caring about it. The crux of that regulation is that inaccurate directory info, especially on payer directories, blocks access to care, especially for Medicare, which those regulations focus on.

Patients have a hard time finding providers when they need care because the directories aren’t accurate. They’re calling providers, scrolling through names, but can’t rely on the info. So grounding yourself in that problem and the drivers behind it is really important.

Now, patients care about getting care, but providers want patients to find them, too. The challenge is framing the problem so providers get the value you’re trying to solve. You’re not trying to add burden or make them create more data.

Providers already spend money on marketing to get patients in the door, so there’s alignment there. Usually, it’s the marketing or referral teams at provider organizations that want to find patients, so the product strategy should work with those teams to make sure they’re providing the data, not clinicians or others.

Given the fragmentation between provider offices, billing departments, and third-party administrators, how do you map stakeholder influence and assess engagement when driving initiatives to improve directory data accuracy?

Payers care about data accuracy, especially when they get audited, but the real problem to solve is — as a patient, can I make an appointment? Can I actually go see a provider? That’s the job to be done. You’ll get a better solution if you focus on that instead of just what payers — who often interpret regulations in a way that works for them and create workstreams that don’t really make sense.

So, if a patient is looking at a directory, what are they actually trying to do? They’re trying to make an appointment. They’re trying to figure out if a provider’s in-network, if they’re accepting new patients. And what I’ve found is that trying to fix this by just pushing for “more accurate data” through that middle layer is often the wrong approach.

If you look at it instead from the perspective of how can I help this patient connect with a provider’s office that can actually get them scheduled, now there are real solutions. Things like direct scheduling are starting to roll out in some systems. You’ve got organizations like Zocdoc that saw this need years ago and built around it. There are others, too.

That lens — thinking about the job to be done — is how you navigate all the fragmentation between provider departments or even across payer orgs. There are so many people trying to do so many things, but if you stay grounded in what the patient is actually trying to accomplish, the solution path becomes clearer.

That said, it’s still hard. A lot of payer directories now show direct scheduling options, but not everywhere. There are real limitations, including different vendors and varying levels of EMR integration. It just takes time.

Understanding the root causes

In your experience, how much of the larger provider data inaccuracy problem is a technical issue versus a human workflow issue? How does that distinction inform your product strategy and roadmap priorities?

At its core, it's a workflow issue. The data that gets used to populate most provider directories wasn’t created to help patients find doctors. It’s usually there to update a roster, get a provider credentialed, or get them in-network. That’s where the problem started. Those workflows were built for very different purposes.

The people doing credentialing or roster updates aren’t thinking about how to help patients find the right provider. There’s usually no job title at a provider organization that says, “Make sure our directory data is accurate.” You might have a network administrator or credentialing specialist, and they’re trying to do a specific job — get the provider enrolled with insurance. They’re not focused on patient acquisition or discoverability. So the workflows themselves are misaligned. And when payers collect data, usually in a spreadsheet, they’re trying to do multiple things with it, which might not be obvious to the people filling it out. I’ve seen spreadsheets that are incredibly complex, and hospital staff are often forced into roles like data analysts just to keep up.

Even when those provider teams are thinking about accuracy, they may not have the right tools to do it well. The whole end-to-end workflow just wasn’t designed with that goal in mind. What’s happened over time is that the task of “keeping directory data accurate” got shoved into a process that was built to solve something else entirely.

Now, yes, there are technical limitations too. There’s talk of interoperability and APIs, and that’s all important. But the core issue is that we’ve been passing data back and forth, mostly via spreadsheets, and a lot happens to that data once it leaves the provider side. People input information, and then it goes through multiple teams, layers, and transformations. Even within payer orgs, it’s often unclear what actually makes it into the directory.

The technical challenges are real, but the bigger issue is the human workflow. If you're trying to solve directory accuracy, whether you’re on the payer side, the provider side, or a third party, you need to figure out who actually cares about this data at each point in the process. Then you have to work with those people to create the right solution for their part of the job.

Because people do care. Payers care, providers care — some teams more than others — and obviously patients care. It’s just about aligning those motivations and making sure the right people are providing the right information for the right reasons.

Given that healthcare provider data flows through many systems, how do you define the source of truth within your directory solutions?

The strategy really depends on where the directory data shows up. If you're a large provider group, you’re generally close enough to the data that your internal systems or someone on your internal team can manage it reasonably well. That said, even some of the biggest provider groups struggle with this.

But in general, providers know who’s working, who’s not, and who’s at capacity. If they’re using an integrated scheduling system, that often becomes the best source of truth. You know the provider is real and available, and you can book an appointment.

For payers, it’s different. They’re the ones regulated on this, and they’re pulling provider data from multiple systems. They’ll get updates from networking teams or rosters, but they’re rarely directly connected to EMRs. So, the most reliable signal for basic questions like “Does this provider exist?” or “Are they accepting patients?” is still the scheduling system.

Ensuring data integrity through maintenance and validation

How do you maintain the integrity of provider data over time?

You need a multi-layered strategy. When I was at CAQH, we saw providers showing up 50 times in the same directory; just basic deduplication wasn’t happening. That’s a technical issue, but it reflects how fragmented things are. A provider might work at multiple locations, and each one gets listed separately. Or they’re part of a large group with 50 offices, and you end up with duplicate entries all over.

There are some foundational things you can do: deduplication, cross-referencing sources, and identifying suspicious records. For example, if a provider suddenly shows up with a new address in another state, that’s a clue they may have moved. You can’t assume it’s correct, but it signals the need for validation.

That’s why having feedback loops is critical. Parts of a provider organization, such as marketing, referrals, and scheduling, care about this information and can help flag or confirm updates. We see this all the time in consumer products. Think about how Amazon will ask, “Hey, is this still your address?” That kind of confirmation logic should apply here, too.

Every piece of directory data can change, including something as basic as a phone number. Or maybe the listing routes you to billing instead of scheduling; that’s a real issue. And half the time, the person giving you the data doesn’t even know it’s going to be published in the directory. These are solvable problems, but only if you recognize that the data flows through disconnected systems and people.

From your perspective, is there a cadence that works best for validating, cross-referencing, and deduplicating provider directory data?

Ideally, top-tier organizations are doing this continuously. If a new address comes in that conflicts with an existing one, that discrepancy should get flagged and resolved, either manually or through a system, right away.

I don’t think there’s a “once a week” or “once a month” answer here. Regulatory requirements now say that directory data needs to be updated within a certain number of days, which is essentially pushing organizations toward near real-time updates.

So from a payer standpoint, you don’t really have the luxury of choosing a cadence like biweekly or monthly. These processes need to be running in the background all the time. If you’re feeding directory data into an interface, you have to be validating and cleaning it constantly.

Is there anything else you want to add in terms of how to surface external changes early and reduce their downstream impact on accuracy?

If you have access to multiple datasets and can cross-reference them, you already have an edge. That’s the most straightforward path.

Back when I was at CAQH, we started experimenting with machine learning to spot suspicious records. If you’re working with different datasets, you can train a model to flag anomalies, like records that don’t look right based on patterns in the data.

It’s been a few years since I worked directly with ML solutions in this space, but the opportunity is definitely there. If your goal is to get the most accurate, up-to-date information and you have a way to verify it, AI tools, especially those that go beyond deterministic logic, can really help. I’m not sure how many organizations are actively using the latest AI for this problem, but there’s a lot of untapped potential.

Standardizing data formats across payers and practices

When working across multiple healthcare providers, what methods have you found most successful for standardizing directory data formats — without overburdening smaller practices with limited resources?

Any solution that works across multiple payers or services tends to get strong adoption. I’ve talked to a lot of practice managers who say they’re uploading these beastly Excel files, each one unique, for every major payer they work with. And when you’re working with 10 or 15 payers on average, that adds up to a lot of administrative overhead, especially for people who don’t consider it their primary job.

This is the root of many data inaccuracy problems. I’ve heard from practice administrators who, just to keep up, reuse one payer’s templated Excel format to send data to a different payer. But those templates are usually very payer-specific. So now you’ve got mismatched formats and unclear interpretations on the receiving end. That confusion flows downstream. People don’t know how the data will be transformed or how it will actually appear in the directory.

So the best thing you can do is drive adoption of a common solution that works across payers. If providers can just do the work once, in one format, they’re more likely to participate, and the data is more likely to be correct.

Building and launching DirectAssure

While you were at CAQH, you successfully launched the provider directory cleanup solution, DirectAssure, from ideation to market release in just nine months. What was that process like, and what key lessons did you take from it?

That was a really exciting and challenging project. Initially, we thought the need in the market was just about regulatory compliance, basically sending a questionnaire to provider groups every 90 days to check if their information was still accurate. That was the common interpretation of the Medicare regulations at the time. But we quickly learned that compliance wasn’t the real problem. The regulation was actually about reducing barriers to care. CMS started doing “secret shopper” audits, calling providers listed in directories and asking: “Is this your office? Are you accepting new patients? Is this phone number correct?” So we had to rethink the problem we were solving.

The first step was validating that problem statement. We partnered closely with payers and ran working groups with stakeholders who owned directory data. CAQH is structured as a payer consortium, so we were in a good position to do that. Once we had alignment on the actual problem — making sure the data patients rely on is truly accurate — we could start shaping the solution.

We also had a huge asset at CAQH, ProView, our credentialing platform. Most providers in the U.S. already use it, and they’re required to regularly update their information there. That gave us a great opportunity to build a feedback loop between provider data attestation and payer directories. Of course, we had a ton of open questions. Would providers actually respond to us? Would they give us accurate data? Would payers be able to use it? Could we deliver something functional on a tight timeline?

So I built a project plan designed to test our riskiest assumptions first. Step one: Would providers actually give us good data? We ran an alpha test. We took 100 records from participating payer directories, asked those providers to verify the info, and got strong validation. The providers corrected what was wrong and confirmed what was right. That gave us confidence that the feedback loop could work.

Next, we had to check if payers could actually use that data. We tested Excel outputs and eventually built out an API to deliver it programmatically. This part was harder because payers have many internal teams, and adoption timelines vary. But it was critical for long-term success.

Lastly, we focused on execution. We had to build the backend, finalize the APIs, and roll everything out. All while making sure it worked at scale.

The biggest learning for me was that if you present data validation to providers in a way that’s meaningful, like “help patients find you,” they’ll engage. They didn’t just verify our data; they also updated outdated info we hadn’t even asked about. On the payer side, the lesson was that even if they buy your solution, implementation can take time. You need to work closely with those teams post-launch to make sure they’re using it correctly and actually seeing results.

Ultimately, just by removing bad records from their directories, we improved accuracy. And that felt like a meaningful impact.

Scaling provider data solutions for small practices and solo providers

In your current role at Alma, where you’re working with many smaller practices and solo providers, how have you adapted your enterprise-focused experience to help them better manage their directory and credentialing data?

One of the biggest areas I've worked on at Alma is around integrating our provider directory with scheduling, because I really believe that’s the ultimate solution.

Our providers want to grow their practices. They want new patients, and helping them get found is a core value prop for Alma. So we’ve built direct API connections with certain payer directories like UnitedHealthcare via Optum, Cigna, and a few others. That means when providers update their info on our platform, it flows through our rosters and gets reflected in payer directories pretty quickly, and even near real-time in some cases.

What’s more powerful is that this data is also tied to scheduling. So if a provider opens up a time slot, that availability can surface directly on the payer directory. That’s a huge win because you’re that much closer to solving the patient’s problem of needing to find a provider and booking an appointment without jumping through hoops.

That said, there are still challenges, especially around credentialing data and what gets published. For example, not all providers want to share the same contact info. Some don’t want their personal number or direct email listed publicly. But payers often require specific fields like phone, address, etc. So balancing provider autonomy with payer requirements is still something we’re working through.

Zooming out, when I first started in this space, it didn’t feel like the most exciting part of healthcare. Provider directory data? Not exactly interesting. But over time, I’ve realized it’s core to how people access care. I recall from our early research, conducted almost a decade ago, that half of the people trying to find a provider turned to Google. And that was striking, because you'd expect your insurance company to have the most up-to-date info. And in some ways, they do, like who's in the network, but the surface-level experience often falls short.

So, it might seem boring. But if something in healthcare seems boring, that usually means it's important. And hard. And probably why no one's fully solved it yet.

What does LogRocket do?

LogRocket's Galileo AI watches user sessions for you and surfaces the technical and usability issues holding back your web and mobile apps. Understand where your users are struggling by trying it for free at LogRocket.com.