Accorded Tech-Enabled Services

Today we’re talking about how you used Accorded at your previous company. Could you give a quick overview of the company and what your role was?

I was the COO at a provider of specialty care services for children. We provided a range of therapeutic services within our chosen specialty for families through our brick-and-mortar centers across multiple states. We were founded in the past few years, so we’re a pretty new company. While I was there, we were also trying to move the industry forward in terms of value-based care, so we had active partnerships in progress with multiple insurance companies to structure value-based reimbursement in our specialty. The vast majority of our services were done in a fee-for-service context, but we were working on contracting value-based contracts as well. That’s the context within which we worked with Accorded (previously known as Cerebrae).

What services did Accorded provide and how did you work with them?

I might not be able to relay the full scope of what they do overall, but in our context they were a relatively new actuarial consultant. They essentially have access to a claims database and offer analytical services based on that database.

We were working to negotiate contracts with payers for a value-based reimbursement structure. So, we contracted with Accorded to do some analytics related to those pitches. We wanted to evaluate our own services against the industry standard. Working with Accorded allowed us to have nationwide multi-payer data points around the cost of service, which helped us put together notional structures of how we might do bundled payments that incorporated different types of services that we provided.

What did the process of working with them look like?

Our Chief Business Officer was the one pitching to these insurance companies to set up partnerships, so she worked most closely with them. We would determine the analysis we wanted. For example, we might say that we want to understand the typical spend, by state, for a certain set of CPT codes, over a given period of time, in specified age groups. We would set the strategy of what different contract options or bundle structures might look like, and then we would go to Accorded. We would specify exactly what we were looking to understand. They would then come back with an analysis, and there would often be a little bit of back and forth to address questions or gaps in data.

How long did you work with Accorded for over the course of this whole process?

I think it was a year. We signed a contract where we paid monthly, and committed to a year.

Was this focused on a single VBC contract or used for multiple contracts?

It was for multiple contracts: we would use the same data for multiple pitches. In some cases, it differed. For example, there was one payer that we were pitching to in one state that we did not have any presence in, and they did some custom analytics for us for that specific RFP.

Who within your organization was engaging with Accorded?

It was our Chief Business Officer, who had a lot of expertise in structuring value-based contracts. I had some of that expertise as well. So we were very, very prescriptive with them about exactly what we wanted.

So to summarize: you were getting a sense of what costs were, based on CPT codes, across states. Was there also modeling of potential impact you could have on the populations you were serving?

We were trying to understand, historically, how much payers had spent on the population that we serve. These are large amounts – for a child, we are talking $60,000 to $100,000 a year in spending. So, we were trying to understand what payers have historically paid over a length of stay. To build on top of that, we were using some key internal metrics about our length of stay and our clinical outcomes. We were comparing those to what limited information you can get in claims data to show that we would reduce the cost of care over the entire length of treatment for any given client.

Were you leveraging them to help with the actual contracting or contract design?

No, we did not work with them for that. And given the existing expertise on our team in structuring value-based contracts and how prescriptive we needed to be, I don’t think they would have been very helpful for us in the actual contract negotiations. We were working with them for data access.

Did you guys evaluate anyone else alongside Accorded when you were going into it?

We didn’t. We had had experience working with different actuarial vendors – my colleague had worked with Mercer prior to this, for example. We heard about Accorded from somebody who knew the folks who had started it, and their pricing was just wildly better than anybody else out there that any of us had heard of. So it was really easy for us to at least just test it out.

When you say the pricing was different, how big of a difference are we talking?

We paid under ten thousand dollars a month, which was a very good price, compared to other claims databases. We didn’t formally evaluate any others in a formal search process, but it was a better price.

Can you talk a little bit about the quality of the work and analysis that you got from Accorded? Did you get what you were hoping to get? How did it compare to what you were looking for going in?

The quality was great. We were very prescriptive – we would give them a really specific analysis that we wanted. If we went to the level of CPT code and diagnosis code and all of the actual data components, they were very good. They were also very responsive and fast. That made them very easy to work with as a vendor.

With this kind of service, is it mostly about the quality of the underlying data? Is it the quality of their analytics work on top of it? What do you look for in a vendor like this?

The underlying data that they have access to is the most important. I believe they used IBM MarketScan data. In my experience, people either have Optum or they have something else and you never get everything all together. So you’re getting some percentage of all claims out there, either commercial or government, and you’re never getting all of it. Accorded’s data is not perfect, but it’s a solid data set, and we understand the limitations of it. The analytics was a secondary aspect for us. We would much rather have just had access to the data ourselves, but it is very expensive to get that. So their analytics on top of it was helpful to us as well.

Got it. The analytics sound like more of a nice-to-have for you. If I can clarify one thing, when you’re looking at the data that a platform like this has access to, what you’re trying to optimize for is how complete or close to complete that dataset is, so that it is representative of the entire universe of claims that exist. Is that right?

That’s right. And to give an example, if you have a dataset on a state that has United, Aetna, and some of these other big national players, but is missing a local player like the Blue Cross of that state, you’re going to have some significant gaps in the data that skew your numbers. So part of the interview process of talking to any of these vendors is to understand how complete their dataset is. You ask questions about the percent of claims the vendor has in their dataset for states that you care about. Sometimes, they’ll have got 80% for that state, and sometimes it will be lower, around 15%.

With Accorded, did you feel like you had sufficient visibility into what claims they had access to, such that you could be prescriptive to the level that you needed to be?

As good as any vendor will give you. Legally, they’re not allowed to tell you which payers’ data they have. But you can potentially infer, in some cases.

Besides price, the completeness of the data, and to some degree the analytic capabilities, is there anything else that you would evaluate as you’re looking at vendors in this space?

The underlying data is the most important one. Every company is going to have different reasons they want this data, which drastically changes the type of analytics that you may need or want on top of it. We were a services company with minimal tech investment, but something that other companies would also be interested in is how you get the data. We would just exchange Excel sheets, for example, but I’m sure some other companies would want something a little more technically forward-looking. I think Accorded were a little bit stronger on that than some of the big companies in this space.

Big companies, like Milliman?

Yes. And you can always ask for sample data. I would never be shy about saying, “Hey, we want a small analysis, as part of this vetting process.” I would always pick a location where we actually knew the numbers and wanted to see how well the data matched what we knew were the right numbers.

How was the sales process with Accorded?

I wasn’t close to it. I think we had a personal connection to them, and their pricing worked for our needs, so it was a no-brainer.

What was the quality of their support?

We mostly just had a point of contact who our Chief Business Officer met with, regularly. Support in a traditional sense wasn’t really relevant for us, but they were responsive, easy to work with, and understood what we asked of them. Overall, I’d say they were good.

Were there any major limitations that you found with their data?

It was as good as with any other vendor. Yes, there were limitations. For example, they did an analysis where we looked at the data and knew it was wrong for one of the states because we were in that state. We thought it was probably getting skewed by their data set, missing certain important payers. Those gaps existed, but you’re going to find that with any other vendor in this space. In my experience, it is just something to be prepared for – know about the limitations.

Was there anything else that didn’t work well with that one-year relationship that you had with them? Anything that could have gone better or that you would have liked to see work out differently than it actually did?

Not really, to be honest. They were great value for the money, and we were very happy. We had the relevant expertise on our end to know where the limitations were and to know what questions to ask when we saw weird things in the data. So, overall, I think they were a perfect vendor for our use case.

You mentioned a few times that you had to be fairly prescriptive in terms of what you were asking them to provide you. Could you say a little bit more about where that came from? Is that just a kind of good rule of thumb for working with this type of vendor in general?

I think some of the analysts you’re going to work with at Milliman or Mercer have been in the data more and know it more. So, you have the opportunity to be less prescriptive.

Also, asking Accorded for the cost of care for our specialty was a big, complicated thing. They could interpret it in a bunch of different ways, and there was no standard for how to do that. Part of what we were trying to do as a company was to be innovative in that space. They were not an innovative thought partner at that level. When I say we were prescriptive, we would look up all the CPT codes that we wanted them to include for the services being provided. That was easy for us to do because we provided those services, so we knew what CPT codes we bill. But they’re obviously not going to know that as well as we do. We would also set the parameters on population characteristics.

Were there any other ways in which your specialty focus changed or influenced the relationship with Accorded or their scope of work in any way?

I think it just made our requests super niche, compared to a primary care group that wanted to do some Value-Based Care analytics. They’ve renamed themselves Accorded, which shows they’re trying to be more of a consultant on value-based care. That’s where they, as a vendor, can look to differentiate – having expertise in more niche markets within healthcare, not just overall cost of care.

And in an ideal world, then would that expertise actually translate into more thought partnership than maybe you had experienced in this one instance?

Yes. In my ideal world, I would love for it to be the type of conversation that I would have with our Chief Business Officer. We had both worked for organizations operating in a value-based construct, so we both had that practical knowledge. If Accorded had somebody on their team who had actually operated in those contexts rather than just being the data analyst, I think that would have been nice. But I wouldn’t have wanted to pay for that.

How, if at all, did the limited presence of tech in your operations influence what you were looking for? Is there anything else that a virtual-first provider or on the other end, a more traditional provider might want to keep in mind, when working with a vendor like Accorded?

The tech-forward provider should ask this vendor questions about how they might integrate with any existing technology that they have. So, if they want this vendor to, say, get into their data and do analytics that compares their data with the database that the vendor has, they should figure out how that might work.

A more traditional healthcare services company that’s not a startup or that has no focus on technology should go into this with an understanding of the limitations of data in general. I think a lot of folks who haven’t been in the weeds of the data don’t always see that the data is imperfect and it doesn’t always reflect reality as they may know it. They need to go in being prepared to actually look at the numbers and vet that it’s giving them what they want.

When Accorded did some analysis and you found their analysis to not align with what you knew to be true, what was the resolution process like?

We recognized the data limitation. We saw what their database said, but we knew the reality was different, so we kept that in mind. It was a bit more of internal level-setting. There were some other cases when they gave us data with funky results and we’d ask them to dig deeper into those. They were very receptive. When we had questions about what they gave us, they would pretty quickly turn around whatever updated analysis we wanted. But we were the ones hypothesizing what the explanation for any differences could be.

If you didn’t have that expertise in house, what would you have done or what would you recommend?

In that situation, having one of those larger consultants would have helped. And you want to either look for advisors or for somebody within this vendor organization who has the actual operating experience, rather than just access to a dataset. I would look for somebody who has experience writing these contracts at a nuanced level, not just negotiating them at a high level; somebody who can speak to the specific details of diagnosis codes and CPT codes. So, I would recommend either having an advisor like that, or pushing the vendor to assign someone with that expertise to your project.

What would you have expected them to catch, as a result of that experience or expertise?

For example, if we had state-by-state reimbursement rates for specific codes, it might have been nice if they were able to point out how unusually big a spread there might have been and prompt us to pay attention to that. It would have been nice to hear if they suspected it was because their dataset might have been incomplete in a certain way. But again, I have never seen that from any vendor in this space. I don’t think that’s a negative reflection on them.

That makes sense. To close, what did you like most about working with Accorded?

Great value for the money; so much cheaper than other vendors. They also had a relatively new team that was responsive, fast, and actively worked with us.

What did you dislike most about working with them, if anything?

Having to be very prescriptive about what we wanted, although I wouldn’t say that was a big negative.

Looking back, I’m curious if you felt like Accorded was the right vendor to go with.

Definitely the right vendor for us. They got us exactly what we needed for negotiating contracts, which was the outcome that we wanted.