Petteri, you are responsible for clients in the medical and healthcare domain at Silo.AI. Could you elaborate on some AI use cases where you’ve seen a lot of value being created in these fields?
I’d say that we have two types of projects that we’ve seen tremendous value being created through the usage of artificial intelligence (AI).
The first one is applying computer vision algorithms to medical imaging, such as X-ray or CT images to spot anomalies such as bone fractures or cancer cells. With sufficient training data, customized algorithms and effective user interfaces, AI can be a great asset to doctors who typically spend significant amounts of time scrutinizing the imaging and planning operations. If that time can be reduced by AI preprocessing and making suggestions, the doctor will have more time available to actually focus on treating the patient and we’ll improve the hospital efficiency.
Secondly, the medical field has moved from acknowledging and treating medical conditions to forecasting the situation based on machine learning algorithms that process the patient data. We’ve worked with patient vital signs analysis and anomalies prediction, and there’s a huge potential for improving the quality of life of the patients as well as saving time and resources with predictive treatment. The similar theme of AI-driven predictive maintenance has been introduced to industrial settings already some time ago, but it is now finally being adopted by the medical and healthcare industry.
Your background is in human-centered strategy and user-experience. What are some of the best practices about how you have seen AI supporting healthcare workers while keeping the patient in focus?
Yes, in the end technology is all about the people.
As I mentioned before, currently AI works best as a support for healthcare professionals by offering preprocessed suggestions and enabling them to spend less time in analysis and more time with the patients. That is crucial not only to enhance the patient experience but to make caregivers’ work more human-centric with more interactions with patients. Less time with bureaucracy and planning, more time with actual care.
By increasing the efficiency of the healthcare services with AI it is also possible to treat more patients with the same amount of human resources, which means shorter wait time to treatment and better clinical decisions. AI assisted diagnosis also helps in ensuring the quality of diagnoses for doctors working in very stressful, fast paced situations or with considerable fatigue. Ensuring an even level of decision quality is important both for the patient experience and for the overall efficiency of the healthcare system.
This trend of extending the capabilities of humans working in healthcare is not only seen in hospitals, but also in other care facilities. By creating smart, AI-enabled services we’re able to support the care of elderly or those with special needs with e.g. better surveillance and more accessible predictive services.
Healthcare is a specific field in terms of conditions and regulatory aspects that affect the possibilities. What are your best practices of working with these challenges when developing smart health applications?
Regulation has both good and bad consequences. On one hand the GDPR and other patient data safeguarding mechanisms make sure that the data is not misused, but on the other hand these make it more time-consuming to create AI-enabled services that require a lot of data to work properly.
One big challenge in applying AI in the medical domain is lack of training data. This means that the medical facilities that hold the patient data are in a key position to develop next generation AI products – the medical device or service manufacturers need to find ways to collaborate to get the data they need for their AI solutions. In this industry data is certainly the new oil, as no AI service can be built with inadequate amounts of data.
What comes to other medical industry regulation, as consultants we at Silo.AI usually work within clients’ R&D programs. The client typically has the regulatory processes already in place, and we naturally comply with them. We also have partners that supply the necessary frameworks for those projects that require regulatory processes.
These are great insights, Petteri. To conclude, let us know what would be the next steps to know more about applying AI into a medical or healthcare business.
If you want to know more about AI implementations in a medical or healthcare setting, I’m hosting a webinar series on Successful AI projects in smart health together with our medical and healthcare clients. The idea is to run a series of webinars to present different kinds of AI implementations that we’ve been working with.
You can enroll in the entire webinar series on smart health here.
If you have questions regarding Silo.AI’s medical and healthcare projects and know-how, you can always reach out to me – I’ll be happy to tell more!