When artificial intelligence begins to transform healthcare, a great question is often omitted: do hospitals, healthcare systems and IT suppliers train artificial intelligence on the right data?
Many AI models are largely based on data from the sources of the US and Europe. Therefore, this may cause prejudices that limit treatment options. Valuable observations from other parts of the world are omitted. In fact, tests He showed that biased data sets can contribute to differences in healthcare and overlook effective treatments available outside the USA
John Orosco has extensive experience in artificial intelligence and data sets through his work as a Red Rover Health CEO. The company specializes in simplified EHR integration via a platform that uses sheltered API RestFul interfaces to connect third -party software with EHR systems. RestFul API is a type of internet interface of the internet protocol that allows the consumer to perform data to download data or publish the update of the source system.
All this has been designed to enable healthcare organizations to improve existing EHR with the best systems, improving access to patients in real time and improving clinical work flows.
We talked to Orosco based on the main challenge with artificial intelligence and data, AI achieved full potential in healthcare, training in more diverse and global data, artificial intelligence Connection with genomics and precision medicine and why AI models should consider therapies without a mainstream to provide patients with the best possible treatment.
Q: What is the main challenge with artificial intelligence and data today?
AND. The main problem with AI in healthcare today is not technology itself – it is that we are still at an early stage of its evolution. Enormous language models still ripen at a swift pace and although they already show an amazing promise, it is clear that we only scratched the surface.
Early indicators suggest that AI will remain and that it basically transform the way we undertake automation, decision making and performance in various industries, especially in healthcare.
But as powerful as these models become, their effectiveness depends on access to data. AI can be as good as the information he has to work with. And in healthcare, where data is often crushed in various systems, buried in unstructured notes or blocked behind dated interfaces, this is a real challenge.
Integration is not only helpful – it is necessary. Without access to comprehensive, well -connected data sources, full AI potential is effectively neutralized. It becomes a brilliant tool that only part of the picture can see.
So a real focus should not be just what AI can do in the future, but on what we can do today to prepare for this future. This means breaking down data silos, building smarter infrastructure and providing access to LLM to the highest quality of high quality data.
As the models are still improving – and they will be – this basis will be determined how much true value we should unlock. In low, the models ripen quickly – now it depends on us whether the data is ready for them.
Q: You think that artificial intelligence in healthcare can only achieve full potential when it is trained in the field of more diverse and global data than today. Please develop.
AND. It can reach full potential only if it is trained in various, global data sets. At the moment, most of the data used for LLM training comes from specific regions, mainly the US, while this may seem like a good starting point, taking into account the amount of health care data in the US, it is actually restricted.
AI training only on regional or national data Pieki in cultural, system and clinical prejudices in this region. It gives us a narrow lens, through which AI understands medicine and health, and this fundamentally limits its suitability.
Take, for example, the USA. Our healthcare system tends to favor some of the treatment approaches, such as prescribing drugs or recommending surgery. However, other countries can mostly consist of natural means, alternative therapies or various care paths.
If artificial intelligence is trained only on the basis of US -based data, it naturally reflects and strengthen these treatment patterns, even if other approaches can be equally or more effective in different contexts. This is one of the reasons why many American patients who can afford to care abroad – because they think that there are effective treatment outside the borders of FDA approvals or American clinical norms.
If we really want artificial intelligence to support better health results around the world, we must think outside. This means that training models on a wide range of data from various countries, cultures and care models. It’s not just about volume, it’s about diversity. Various data make artificial intelligence wiser, more adapting and ultimately more fair.
Without this, we risk the construction tools that are technically advanced, but functionally narrow. If we want artificial intelligence to reflect the full spectrum of human health and treatment opportunities, we must give it a fuller picture of the world.
Q: You say that genomics and precision medicine can offer more personalized care. What is AI and a combination of data?
AND. There is a mighty relationship between artificial intelligence, data and the future of personalized care through genomics and precision medicine. Think about the human body as an operating system. Each of us works on our unique source code, which is our DNA.
The mapping of the genome basically decodes this system. He tells us how we are connected to respond to some drugs, how we metabolize drugs and even what conditions we can be predisposed. However, despite this insight, most state-of-the-art medicine still take a trial and incorrect approach.
We rewrite the treatments, and then we say: “Let’s see how you feel in a week.” It is inherently imprecise and often incapable and even risky.
It is there AI can play a transformational role. When AI is trained in the field of genomic data in combination with other clinical data, such as treatment protocols, laboratory results and real evidence, it becomes much more precise in its forecasts and recommendations.
Considering the genomics for the data mix, and can facilitate identify the most effective treatments for each person before starting trials and errors. This can also facilitate avoid solemn side effects by determining the drugs that a person probably poorly metabolizes or does not respond at all.
The future of precision medicine depends on this type of integration. The genomic data themselves are valuable, but their full potential is carried out only when they are combined with wider data sets and analyzed on a scale by AI. When this happens, we are approaching care, it is not only personalized, but proactive, predictive and safer. AI becomes an engine that turns the data into insight, and the genomica becomes the basic layer of really individualized care.
Q: You also suggest that AI models should consider therapies without a mainstream to provide patients with the best possible treatment. What do you mean?
AND. I mean that AI models should expand their view beyond local mainstream treatment protocols, especially when these protocols are determined by the regional management bodies. Too often, AI systems are trained in the range of data sets that only reflect what has been approved or returned in one country, usually based on regulatory parameters or insurance.
Although this may make sense from the point of view of compliance, it limits the potential of artificial intelligence to provide patients with a really comprehensive image of available treatment options. The fact that therapy is not approved by the FDA or is not covered by insurance does not mean that it lacks merit. In fact, it can be widely accepted and effective in another country.
AI not ignored by AI should not be ignored by AI without a mainstream or alternative therapies. Patients deserve what is available – not only in a postal code or an insurance network, but all over the world.
Of course, access and reimbursement are real barriers, and in the United States there are political and regulatory complexity, especially here, but the role of AI should be to inform and expand the conversation, not narrowing.
If the patient sees the recommendation of treatment generated by AI, which includes a promising therapy used on the international arena, he can then discuss it with his doctor and make a conscious decision.
At the end of the day, AI should serve as an objective guide, not restricted by local policies or insurance restrictions. Strengthening patients with a wider view on what is possible can lead to more personalized, thoughtful care. It will not be straightforward to implement this globally integration way of thinking, but the lack of this means that AI will always have the real potential for supporting better health results.
Watch now: the main AI officers require a deep understanding of OPS technology and clinical technology and clinical