How Semantix Turns Healthcare into True Open Health

15 years ago, when he was born, Semantix was a traditional consultancy within the universe of big data, analytics, artificial intelligence (AI), focused on the sale and implementation of third-party solutions.

At the time, few people ventured to talk about it. And far fewer people understood what it was about.

The game changed from 2018 onwards, when the Semantix founders and team realized that they were running expensive, complex and multi-vendor projects along the chain.

Jorge Carvalho, 44, Head of Health at Semantix, recalls:

“There was an opportunity to create our own end-to-end, low code, big data product with more flexibility, portability, free of lock-ins. [travas que impedem a troca do fornecedor] and cheaper, because it would be a national solution.”

There were two basic ways to do this: start developing a platform in-house and make strategic acquisitions to accelerate the journey. And so it was done: in February 2021, the company bought LinkApi, renamed SDP Data Integration.

Today, Semantix has the SDP Data Platform, a proprietary platform for building big data projects. It is a modular system, SaaS model, which aims to provide end-to-end, from ingestion, integration, engineering, science (algorithm construction), visualization and monetization of data from various industry segments. Including the health sector.

As a company of deep tech (technology that is below the line of sight of consumers and users and does not provide solutions with an interface with the final consumer and/or patient), Semantix appears little, but in its portfolio there are 700 clients from various segments, in 20 countries.

In August 2022, the company made its IPO on Nasdaq – the first Latin American deep tech listed there –, at a valuation of approximately US$ 1 billion.

The move capitalized the company to continue its organic and inorganic expansion.

The health sector was already part of Semantix’s strategic plan and Jorge’s arrival at the company raised the vertical bar at the company. In addition, the service to some of the biggest players in the market – hospital groups, pharmaceutical companies and health and dentistry operators – received another boost with the joint venture with the Hospital Care Group, materialized in the Tradimus (2021), and with the acquisition of healthetch Zetta Health Analytics, shortly after the IPO.

The SDP Health big data platform helps hospitals, clinics and laboratories become more profitable with more efficient operational, financial and population health management processes. The tools available include ready-made APIs for the healthcare sector, analytical dashboards, algorithms, curated data sets, among others.

In a relaxed chat with FUTURE HEALTH, Jorge Carvalho talks about the use of big data in the healthcare market and presents some of the best customer cases served by Semantix. Read below and understand how to move from paper-and-pen prescription to an artificial intelligence algorithm:

FUTURE HEALTH: How is Semantix’s Health vertical constituted: how many people do you lead? Do they have autonomy to make customized projects or does everything go on the platform as a tool for everyone?

JORGE CARVALHO: I’m going to take a step back and explain the concept of verticals. Their objective is to bring us closer to the industry, to allow us to understand the pains – from a data point of view, integration, data platform – and build products or solutions that can address these pains.

These products and solutions can be customized for the customer who wants something specific – the marry de Hospital Care is a good example of this, because they had a specific need and wanted a customized solution.

But we also offer ready-made solutions. Zetta brings a lot of this ready-made knowledge. For example, someone wants analytics for a health plan – here it is, I already have all the algorithms, indicators and dashboards ready and, in a few days, the client will have this analytics to start making a decision.

Today, we have around 60 people in the healthcare vertical. When we talk about SDP for health and products we want to put on the market, we have a roadmap for short, medium and long term health.

FH: Zetta came with the expertise in AI and analytics. Did Semantix already have something for AI? What did Zetta add when it was acquired?

JC: Yes, Semantix already had several artificial intelligence, big data and analytics projects for healthcare customers. We have already carried out relevant projects for dental and health operators.

The Hospital Care Group is one of the most comprehensive cases for the size of the group – there are 35 assets and more than 200 million lines of data, every day. We elaborate the entire process of big data, building dashboards and algorithms.

When talking about artificial intelligence, there is a whole maturity of the company in relation to data to enable the construction of AI models. We like to say this because 100% of the companies I visit say they want to do an AI algorithm and have no idea how to get started.

So, we have to explain that there are levels of maturity: level zero is non-digitization; level one is being able to bring that data into a single platform; level two is data engineering; and number three is data science. It has a ladder.

At Hospital Care, we have an algorithm that predicts the average length of stay in the emergency room, in all hospitals in the network. The first two layers we addressed in this project were financial efficiency and operational efficiency. Now care efficiency begins, when we actually enter the world of medicine, of care indicators.

For this, Zetta provides a team of specialists in health. They brought in epidemiologists, physicians, scientists and data engineers who have worked in the healthcare industry for many years. It’s a team that understands these data sets and how to do this kind of tying.

The other thing they contribute are ready-made products. For example, they brought a Comand Center hospital care product, which we will include in the Hospital Care project, already with several algorithms ready – sepsis risk detection, etc.

Another example, for the portfolio of health plans: we have five models for the predictability of a person entering a line of specialty care – based on exams performed and indicators collected over time.

We now have ready-made algorithms, but that doesn’t mean we can’t create others from scratch; we have two models.

There are companies like Hospital Care that want a blank slate: a data project in which they contribute their knowledge together with us. And we have ready-made products for a hospital that wants an assistance Command Center with validated indicators and ready-made algorithms that already read the hospitals’ databases.

FH: What difficulties does Semantix usually encounter in big data and artificial intelligence projects in the healthcare sector?

JC: From a digitalization point of view, the healthcare sector is a few steps behind other segments such as retail and the financial market. But all major healthcare groups today have robust digital transformation projects, and this process involves digitizing processes.

The first step is always to know what data is available. When we feel that the company’s maturity level is not high, we make a assessment of data. What is that?

It is to understand which systems are used; what the data is and how it is organized; how is the level of maturity of people, teams in relation to data. There’s no point in providing a bunch of indicators or data for people to make a decision if they’re not prepared for it.

In some customers, for example, we do the assessment of data and discovered systems that many employees don’t even know they exist. We pull that data, sanitize it, engineer it, and organize it so it’s best used.

This is not so clear for companies. Many of them think that the implanted system will generate the data. It’s like that oil analogy… As long as it’s just oil, it’s worth very little. You need to refine it so that it has multiple uses and more value.

There are many healthcare companies with large volumes of data, but they are disorganized data, a lot of historical data that is not ready for use.

Most healthcare groups are at a stage where a data lake has already been created: the data is organized but the question remains: “Now, what do I do with it? How do I create value?”

A good example of this is our project with Hospital Care, which wanted to predict what the volume of people in the emergency room would be like. So, is the day’s temperature data important? Are we going to deal with this data to see if it changes anything? Perhaps, on colder days, the tendency is for the emergency room to receive more people.

This is just one date set. Imagine taking several other sets of data, putting them there and enriching this information? A more robust algorithm is created from the point of view of intelligence, forecasting, looking ahead. But you can’t look forward if you don’t understand where you are. That’s the big point.

FH: What does the seal “We’re open health ready” on the Semantix website, especially on the SDP Health landing page, mean?

JC: It is, in a way, a provocation. We say that we are willing to integrate with any system, to be the platform that will work along the lines that the Ministry of Health has set for the National Health Data Network (RNDS).

We have already worked on the FHIR (Fast Healthcare Interoperability Resources) protocol, which is being used for interoperability, because we want to participate in this new wave – the big data transfer to improve operational and care processes between healthcare actors. That’s why we use this seal.

In my understanding, the establishment of Open Health, which is so sought after in the health sector, is only possible if the user manages to have a more integrated care journey through data. For this, it is necessary to have these “free” data – within all security and secrecy rules – to be used by health institutions.

This is the way to have a more fluid, more digital care and experience, with platforms that integrate with each other and even with the public system.

Leave a Comment