My name is Raja Sekhar and I’m a computer science postgraduate from the Indian Institute of Sciences. I lead the technology team at Karkinos and anything that comes under the remit of technology, whether directly or indirectly, comes to me.
I’ve been working in the technology industry for the last 25 years, primarily with financial technology companies. I also had my own firm for around a decade in property tech.
Then I became part of the founding team here at Karkinos, which is primarily focused on oncology: cancer and all aspects of cancer from early detection, diagnosis, care delivery and research.
How did you move into healthcare?
Healthcare is still new to me on a relative scale: I came to healthcare with a very clean slate. I don’t have any baggage or experience, so most of the things that we try to solve at Karkinos I try to look at from a first principles basis, and apply the knowledge that I gained in other industries.
I bumped into openEHR, as a part of my self-induction into healthcare: I was learning about healthcare tech through reading, attending various conferences, participating in hackathons, speaking to industry experts and gathering some understanding – and that’s how I came into this madness of standards in healthcare tech and avalanche of acronyms!
Why did openEHR appeal to you?
openEHR made a lot of sense for me – the maturity of the specification; the way it actually separates the data and application layer, the quality of the documentation; in my view its brilliantly articulated specification (without getting too deep into the technology or implementation!).
That is the brilliance of openEHR compared to the other standards that I had come across: it doesn’t talk about a protocol or JSON or an XML. It is fundamentally a construct, a specification and that’s the way it is supposed to be.
Other standards that I have come across are relatively shallow, if that’s the right word. I feel they’re trying to focus on one particular problem and to solve it instantly. But most of these standards evolved around a particular use case which they were not able to solve, and then drown when trying to solve another problem in depth.
That’s where I found openEHR to be more suitable for the kind of work that we were embarking upon, but I was aware that I was going against the norm when I compared myself within the healthcare tech industry.
How do you use openEHR at Karkinos?
Karkinos is essentially a technology platform that integrates a multitude of data elements for the cancer-related ecosystem, which in itself is fragmented in nature, because cancer is a long-term disease. It is not an emergency or an episodic disease; it has a very long tenure. The patient has to go through a multi-year journey involving multiple care providers, laboratories or oncologists and so on. They may have to go for a mammogram or a genomic2 testing, or for biopsy or a PET CT and on each of those occasions the data that is generated for the patient is absolutely fragmented, it is in different places, in different silos, where they don’t have access to the data when they want it.
From a technology perspective, what we are essentially trying to do at Karkinos is to bring together all these fragmented care points onto a single platform so that whoever is dealing with the patient – whether it’s a clinician, a nutritionist, a surgical oncologist, a molecular tumour board, even the patient and their family – has a consistent view of the patient.
It is bringing together clinical, imaging, pathology and genomics data into a single place and in a longitudinal way. Whenever a patient comes for further treatment we try to gather the incremental data from the patient and then stitch together their journey in a seamless way, and also leverage the data to see if we can generate any insights from a research perspective.
Cancer treatment is protocol-driven and very tightly controlled. There are guidelines for any type of cancer, depending on the stage and the type of cancer and the diagnosis that the patient has. The discretion of the clinician is still there, but it is very limited and the next steps are generally pretty codified.
But the biggest challenge with cancer is early diagnosis. If you are able to diagnose, then the treatment is pretty much, for lack of a better word, automated.
That is what we are essentially trying to do: get the data in particular structured way and store it, and by doing so recommend the next line of treatment in an automated way to suggest to the clinician on the ground that these are the next steps you need to take for this particular patient, based on this data.
The clinical decision support system is very close to what we want to implement with associated task planning. For example, if somebody has to undergo chemotherapy for six sessions across six weeks, the nursing staff basically should know that this particular patient has already undergone two rounds of chemotherapy. He or she is coming for the third round, so before their appointment they need to do the usual precautions and they know the steps that they have to follow. They can advise the patient proactively.
We also want to employ mechanisms that help the patient to report their status – patient reported outcome metrics. So from an end-to-end perspective we are playing across the entire spectrum of the cancer ecosystem.
At each phase, a different element of technology comes in a different shape or form, or the experience itself is different. So the platform has been designed in such a way that the experience layer is separated from the actual application logic and the data.
And in that data layer, we decided to use openEHR, essentially to persist the various elements of data.
In a nutshell, what we offer in our data platform is an open data standards-based oncology-specific clinical data repository.
Is that where you get involved, Chaya?
Yes, I along with the rest of the team are involved in the clinical modelling, data ingestion and extraction for the EHR data. There are many features of openEHR that make it relevant for Karkinos to persist data using openEHR.
The fundamental principles and concepts of reference models which include archetype, template and data structure models makes it easier to ingest data and to analyse and extract insights.
At Karkinos, the data that is ingested spans from initial risk assessment to screening and the clinical data, lab reports, digital images and the genomics data which represent a wider range of clinical datasets. The concepts and components of the openEHR system helps to build these complex datasets easily. openEHR’s flexible,Interoperable and reusable data model enables us to capture this wide range of clinical information in a standardised format.
At Karkinos, the data that needs to be extracted varies across the different stakeholders. It could be cohort data to identify the patterns and trends in patient populations, or be able to create a comprehensive and longitudinal view of the patient along the clinical journey. This is possible in a standardised way using the AQL. As this supports the use of standard terminologies such as SNOMED CT it ensures the consistency and interoperability of the data. This makes data extraction comprehensive across various sources of data such as community setting across screening sessions, clinical encounter, laboratory reports etc. where all the queries are based on the SNOMED-CT terminology
So openEHR with the flexible data model, standardised AQL along with standard terminologies such as SNOMED CT allows for the creation of powerful data analytics across all the fragmented data collected during the patients journey with Karkinos.
Is there anything you’d like to see from the openEHR community?
Most of our communication is through Discourse – openEHR’s internet forum where our doubts are clarified. People out there are very responsive.
It would be good if we had a sandbox where we can try things out. It would definitely help us to evaluate our work before we jump into clinical modelling, just to check if it’s already been done. Also, an active community around any open source openEHR server would be highly desirable and improve adoption of openEHR.
What’s next for Karkinos?
We are an end-to-end player in cancer care. In India today, 70% of cancer cases diagnosed are in stage three or stage four, so we are trying to shift that to see if we can diagnose at stage one or even earlier. But we already have patients who are at that certain stage and we also need to take them into consideration. By India’s numbers – because of the size of the population – it may be small percentage-wise, but it’s still very large absolute numbers.
So we’re trying to increase the accessibility of cancer care as much as possible, because in a country like this, accessibility is a major challenge and cases are getting detected late.
To improve accessibility, building new cancer hospitals is not going to take us anywhere. It’s just impractical to start building cancer hospitals in every locality. So what we try to do is partner with existing care providers, like a local hospital, primary care health centre, nursing home or even a dentist clinic. We approach each of these individual players who already have a certain catchment area and we educate them about a particular type of cancer and also enable them to treat those, with help of the technology platform.
We bring the technology into these particular primary care hospitals or non-cancer hospitals, and we implement oncology modules in all these places. And in this way, we increase the capacity.
At present, we have around 70 partners across the country in different regions. And we have a large team in our ‘command centre’ which navigates the patient through the next steps, so when somebody is ratified as high risk or somebody undergoes chemotherapy, they can use the platform to understand what the next steps are for the patient.
We call it a DCCN – a Distributed Cancer Care Network – and as Raj said, one particular element of the network can give you a view of the patient related to diagnostics, another care point might tell you about chemotherapy, and all these nodes which would be otherwise isolated and fragmented are connected on the platform. So that is the power of the platform: it’s completely cloud-based: a cloud native sort of architecture that we have come up with.
One more thing: what does Karkinos mean?
That’s a very interesting question! Well we wanted to create some intrigue. The root ‘Karkin’ in Greek/Armenian refers to a giant crab, a symbol for cancer. The word Karkinos was used by Hippocrates to indicate cancer. In addition, Carcinoma in Latin stands for Cancer and Karka in Sanskrit symbolises the zodiac sign Cancer. And that’s why we’re Karkinos.
Leave a Reply