HIStalk Interviews Paul Roma, CEO, Ciox Health

Paul Roma is CEO of Ciox Health of Alpharetta, GA.

Tell me about yourself and the company.

I have been the CEO of Ciox for a little over a year. I came from the professional services world as the global head of analytics for Deloitte & Touche, which constituted 87 countries and all of the analytics work that that global firm does. My background throughout my entire professional services career, outside of running the global analytics business, was healthcare — life sciences, domestic government work, international healthcare work, providers in health insurance.

Consumers complain about the cost of getting copies of their electronic information from their providers and Ciox has sued over the HIPAA limitation on how much providers can charge. What are the current topics around that issue?

Just to be clear, Ciox is not suing over what consumers get charged, so let me reframe that a bit. We are, in particular, very pro-consumer and consumers getting their health records. I want to be 100 percent clear on that. Our lawsuit has nothing to do with the rate in which consumers are charged or whether they’re charged.

Our view is that there is a burden that is put on hospitals and physicians in professional, for-profit situations. The legal profession, the insurance profession, and others are using a consumer angle to create a burden on the doctors to feed them the record at a very low rate. Our lawsuit has to do with that. We believe that the people that are using information for commercial purposes should pay, and that the cost of producing that information in the proper format should not solely rely on the doctor and be a loss item on their balance sheet. They should be reimbursed for it. That is the bulk of our beliefs and our lawsuit.

So a for-profit company using a patient’s medical record for commercial purposes is different than patients getting copies of their own records?

It is, yes. There’s a explicit differentiation between the two.

Say I get a study done. I go to one of my local hospitals and I want to get that record to bring it to my primary care physician. We do literally millions of those, sometimes multiple millions a month, in which we either don’t charge at all, which is the usual case, or we charge very little. We are very pro consumers getting their health information.

We also do tens of millions of doctor requests for information for continuity of care and things like that, for which we don’t charge, of course. If it is for the consumer and for their health, we are very much in favor of that information being dispersed, being liquid, and transacting at a frequency and rate that is conducive to health being improved.

Are you seeing anything on the horizon that would change the way that the ownership and exchange of medical records will work in the US?

Near-term, no. Long-term, in my opinion, it is somewhat inevitable that the benefit of the data flowing in a secure, de-identified, and traceable way and being available for research outstrips all the reasons the walls are built up for us not to share the information. Long-term — whether it’s a change in definition, a change in regulation, or a change in the belief system of how that information moves — I do think we will see change.

What is the interoperability technology marketplace position of the newly announced HealthSource?

HealthSource fits squarely in the enterprise need for clinical information. We service providers, health plans, life insurance companies, and life sciences companies. HealthSource is a cloud-based, HITRUST-certified product that allows for both the interoperability with third parties — because we have hundreds of thousands of digital connections that we build into workflows — and sharing within the enterprise.

At some of our larger clients, we service 100 different use cases that require clinical information. Health insurance examples would be prior authorization, medical management, risk adjustment, and quality. Our HealthSource software integrates to those use cases and provides the information that they need from the medical chart, the EMR, to improve their process with the clinical information instead of relying on, in the health insurance case, claims and other secondary clinical information. We’re using the primary source to improve their use case.

How much technology and labor is involved in providing a complete electronic chart?

It varies. I’ll say two things. One would be that, as both a citizen and someone running a business, I wish it didn’t vary. I wish it was more liquid and that the outcomes were faster. The reality of the situation is that a large integrated health system has, on average, 17 different EMR systems. A vast majority of hospital systems have not even brought their acute systems to a single system, let alone all the specialty, post-acute, ambulatory, and other. Even within one practice area, they haven’t centralized. I would say that’s the norm.

Because of that, to your point on labor, about half of the cost comes from technical integration, formatting, and information and data management. About half the cost is still from manual touches, whether that be on the front end to work with the information or on the back end from a QA perspective.

Our particular business is reliant on, and cautious of, the regulations that are put on it. We are fully compliant to SAMHSA, as an example, which is a federal regulation to redact substance abuse information. There are many other things that we do. We not only get the information and put a longitudinal view together, but we structure the information — both technically as well as from a redaction perspective — so that it is compliant in the situation we’re offering.

One of the major distinctions for us that has cost associated with it is that we are not a generic exchange for clinical information. We are very particular as to what we’re sharing and making sure that it meets the regulations, that the information’s been redacted appropriately, and that the endpoint is receiving the format that it needs. All of those things are unfortunately more costly than just broadcasting information.

We have the possibility of expanded data sets that include genomics data, wearables data, and other data sources that aren’t being widely captured and collected and stored today. How do you plan for that as a company?

Our clients ask us to add a major source almost monthly. Many are the examples that you just gave. For example, genetic information and the translational makeup of information that combines phenotypic and genotypic data together to create a full picture of the person’s health and vitality. That’s been in our system for a long time, so we’re covered off on that. But below that, there are numerous social determinant categories, such as activity-based tracking from wearables and other IoT devices. We have a backlog on a monthly basis for life science companies and health insurance companies that are driving those changes and requests for further integration.

We lean in heavily on the Argonaut system, which is HL7 standards-based FHIR communication. It simplifies those things. The endpoint can communicate with us at that standard and they’re using the CCDA format, which we use. It’s pretty easy. But some of them still require proprietary interfaces. We maintain at this point about 700 different interfaces, so it’s still pretty costly to do all the endpoint integrations.

Are you seeing promising uses of artificial intelligence or machine learning to make sense of that wealth of data that we now have moving around?

This is a whole topic in and of itself. My background is as a data scientist and my formal work is in the technology of artificial intelligence and cognitive computing, so we can go as deep as you want.

Current state is that for us as company, it’s our largest investment — the structuring of data and the intelligent understanding and summarization of that data. Within the HealthSource product, we have a component called Smart Chart that takes all of the unstructured elements — progress notes in the EMR, a pathology report that’s coming out of the prognostic indications or from test results — and structures those and puts them in an analyzable format.

To your point on AI and cognitive technologies, we then come back through in a cognitive match and build a probabilistic model with confidence levels that deciphers the diagnosis codes, the DRG codes, and many of the other prognostic indications and then builds insights from those. Those insights in our generally-available product are generating tons of value.

To get back to the first part of your question, those technologies I just described are already showing literally hundreds of millions of dollars of increased profit for our clients. Hundreds. Not tens, hundreds. That fuels our investment and the industry’s investment. The “man versus machine” shift in terms of capital investment in those things is increasing on a monthly basis. There’s more information that leans in on the limitation of what a human can decipher.

But the information and the correlation of that information is also getting to the point of complication. Even if you or I are reading it, it’s a 1,200-page EMR. You’re deciphering a list of genetic bases from the four billion genetic bases that are written in a progress note that don’t have a paint-by-numbers key next to them. I have an MTHFR gene expression and I happen to know that that’s a methyl pathway issue that could cause drug toxicity. There’s lots of other things. But that’s in the progress note written out, and as the clinician looking at it, there’s four billion of them. How in the world do I decipher that?

I’m using the most acute example, genetic basis, just because of the number. But the complication of this information has exceeded what even the most well-trained doctors can comprehend. That comes back and fuels the investment curve. There’s been so much progress made and it’s starting to pay off.

Do you have any final thoughts?

The US needs a better way of sharing information — with consumers, for seeding research for better therapies, and getting better information to doctors. Ciox is in the middle of helping all three and that’s the mission that we’re on. The HealthSource product is squarely designed to first give better information to doctors, second to facilitate consumers to get that information in a format they can use, and then third to power research and insights at these large organizations — health insurance, life sciences — that are ultimately trying to create better therapies for us. We’re excited to be part of that mission and believe there’s a lot of value in it.