Incorrect therapies are a major problem in health care. According to the WHO, one in ten patients receives the wrong treatment, resulting in unnecessary suffering and often death.
is an EU-awarded AI decision support tool that helps doctors find the right treatment for their patients. The software, which already covers endocrinology and cardiology treatments, matches patient data with evidence and treatment guidelines, leaving the final decision to the fully informed and AI-assisted doctor. We spoke with Helene Schönewolf, CEO of the young Berlin-based startup, about her innovative solution, the collaboration with Siemens Healthineers and the linking of artificial and human intelligence.Hello Mrs Schönewolf. Let's jump right into the topic: How did the idea of RAMPmedical come about?
I have been working in the health care sector for a long time – I grew up in it as the child of a family of doctors, and I am also a member of the medical profession myself. Again and again I have come across the fact that medical errors happen relatively everyday. Not that we don't have incredibly well-trained medical professionals. Medicine is simply infinitely complex and research is now advancing faster than anyone can grasp it – at least not in a way that prevents mistakes from happening. At some point, an unbearable thought occurred to me: Someone – it could be you – is given the wrong drug for a serious illness, even though the right drug would be available, and that's only because knowledge about such drug has become so unmanageable. Someone in my family has also died from it. We expect doctors to know everything, but few doctors have the luxury of spending hours researching a patient’s condition. Moreover, our treatment guidelines and potential reimbursements are practical. With my co-founder Jacques Ehret, who at the time was working on his PhD in AI and diabetes, we had to find a solution to this.
According to the WHO, one in ten patients receives the wrong treatment – with severe consequences. How can your AI solution decisively help and distinguish between proper and incorrect treatments?
We have translated into software the considerations a doctor makes to find the right therapy. The doctor records the details of the patient, must understand correlations, include and interpret the current state of studies, follow the hundreds of pages of treatment guidelines and weigh the individual therapy options. We have solved all these steps using everything from simple algorithms to state-of-the-art AI. Mistakes are avoided and doctors are told why certain therapies would be wrong. Optimal and slightly suboptimal therapies are presorted for the doctor and presented with comparative parameters.
After all, in some cases – as without AI support – these are recommendations that make the difference between life and death. What about the legal issues here?
In fact, an overlooked contraindication, such as macrolide antibiotics for some heart conditions, can lead to death. Yet, our software works as a reminder. Doctors once learned that, with coronary heart disease, certain antibiotics, namely macrolide antibiotics, should only be given when nothing else is available. Suppose our software has an error and misses this point and the doctor prescribes this drug, then the doctor missed it too. The final decision remains with the doctor. That's how it's designed – and that's how it should stay.
How is your team structured and what expertise does it bring to the table?
Fortunately, I have a terrific, friendly and extremely expert team. Jacques Ehret and I founded the company six years ago. Jacques is a computational chemist and computer-aided drug designer. He has already received two EU grants for his AI research – one on diabetes and one on nanoparticles.
For ten years prior to founding the company, I worked at various levels in health care – in the hospital, in pharmaceutical marketing, MedTech business development and sales, including internationally. At university, I became acquainted with everything one needs to know in the industry, from accounting, marketing and economics, to programming, law and chemistry. At some point, I took the examinations of the Chamber of Industry and Commerce in order to officially belong to the professional medical community. I like to be able to converse as an equal in any area.
In addition, we work in the company with additional programmers, and closely with doctors as consultants. In addition, a few experts have agreed to be our mentors, and last but not least, our committed shareholders.
You have recently started working together with Siemens Healthineers. How did this come about and, so far, where has your solution been used in general?
The cooperation came about through Flying Health. If I understood correctly, Siemens Healthineers was looking for decision support software. Walter Schmid from Siemens Healthineers understood our product right away; after all, he himself had been involved in similar internal projects. This is good for us – because the complexity of our software is not always easy for everyone to understand.
Our solution has been integrated technically with that of Siemens Healthineers at a deep level. This means that customers who want to use the Siemens solution can also use our software, put simply, at the touch of a button. This is extremely important, because hospitals have very limited capacity for intregating software.
What are the opportunities and risks – especially in relation to a wide variety of health insurance and hospital systems?
We can, as we did with Siemens Healthineers, at the push of a button provide greater therapy safety to all patients whose data is in the systems. However, not everyone is reacting as quickly as Siemens Healthineers. Many still have a distant problem focus. In corporations, individual employees often have different priorities such as realising a smaller project that the employee understands and can implement quickly. We are a software that is complex and needs to be deeply integrated, because it needs to understand the whole patient. In return, it has tremendous intangible added value, but also financial added value. Specifically, incorrect therapies expensive for hospitals. And this opportunity to also save costs via better therapy decisions is slowly arriving, and is now what is leading in this system to hospitals thereafter being able to afford the expense. And what hospitals can afford and desire is critical to hospital system providers.
We don't have a great deal to do with health insurers, because they are not in the digital health app space – so they don't have an app for patients. Anything else is too costly for a startup like us. As a patient, I would like to see health insurance cover our costs. I think insurance should pay for what makes or keeps me healthy. Yet, maybe that will change. We're definitely keeping an eye on it.
After all, we are talking about the processing of highly sensitive patient medical data. What about the data security of your solution – the key phrase being "data leak"?
If you place your data in our solution, you need not worry about someone getting it who shouldn't have it. Of course, something can always happen, such as someone breaking into the hospital, accidentally receiving all the access data and reading your data. Yet, someone can also break into your home or into a doctor's office with paper files.
With our solution, data is stored on German Telekom servers in compliance with the German Data Protection Act (DSGVO) and is divided among various databases, completely anonymised, and partially encrypted. All data – if you want to do anything with it – can never be encrypted. And we would like to check your data to make sure everything is in order. I also have my personal data and that of my family in our software. Believe me, I keep the data as safe as possible.
Of course, the success of a project often stands or falls with a grant or award. This has worked out very well for you so far...
You couldn't have put it better. The IBB staff went to great lengths to support our project and I am infinitely grateful! Our software has previously worked in small medical sub-areas, because building out each area is extremely costly. Each area must be worked out over months, then professionally validated and certified. There are always adjustments. User interfaces need to be created so that our software actually fits into medical processes. All of this is extremely expensive up front, before any revenue can actually be generated. We are creating an enormous IP, a software that no one can replicate so quickly – because this requires funding. It is rare that a private, large investor will accept early experimental phases – or that a high level of revenue will take a long time due to long development phases. Yet, software with a unique selling proposition and market maturity is extremely profitable.
How do things look in terms of the Berlin ecosystem's ability to innovate – with a particular focus on the space of AI Health?
In economic terms, Berlin needs innovation. Innovation is the only way for economic growth. I know some great startups that have emerged in Berlin. Berlin is such a culturally diverse and open place – where one can find creative solutions. I hope it stays that way.