In 2009, I was involved in developing artificial intelligence and Machine learning algorithms for speech recognition and computer vision problems in semiconductor industry, for entertainment applications. This made me passionate and curious about the power of artificial Intelligence algorithms and applying these ideas for early detection of diseases in the clinical domain.
With limited knowledge about healthcare domain, I started studying and developing fragmented solutions with computer vision techniques to detect the diabetic retinopathy and also measure the intima-media-adventitia thickness in the carotid arteries from the ultrasound images. But I realized that these fragmented solutions does not have much impact since providers or any specialist doctors suggest the scanning techniques when there is a clear evidence of the diseases.
I wanted to implement these ideas for early detection of the diseases with respect to the patient historical data such as family histories, patient clinical history and socio economic factors and other third party data.
Later, I started practicing this to predict the risk of readmissions from the transactional data such as claims data. As I started putting efforts on these transactional data sets, I felt these data sets are highly inadequate as the data was not majorly helpful because a lot of hidden insights reside in the unstructured data. Clearly, this was the data which should be aggregated and collected from the multiple populations to study the cohort of disease. Unfortunately, owing to multiple factors, there is no such platform to exchange clinical, financial and all other available data to do any meaningful analytics.
So, I started researching on getting lot of data sets in later part of my carieer, but my attempts were not as expected. That was when I decided to solve the problems of collecting cross organizational dataset by considering the regulations and data security aspects like HIPPA and PHI etc. In parallel, I was observing the developments with Block Chain – which is solving data security aspects, decentralization and immutability of the data, permissions for the given participants and eases the complex process to get the data. Based on these aspects I decided to build this platform to deploy the Block Chain network on the cross institutional data exchange for multiple hospitals. I created a pilot project and to develop a Block Chain enabled network to exchange the health information across the health organizations which is more patient centric, trusted and more authenticated information.
The other challenge is once we have cross organizational data sharing, the data will be in unstructured formats like unstructured clinical notes, images, signals and tabular data sets. To solve this we started building the capabilities to convert unstructured clinical notes into structured data using advanced NLP techniques. We also focused on several computer vision algorithms to understand the metrics from the radiological datasets and big data solutions to hold this large amount of data. This helped us to create the entire echo system to securely exchange the data and then deploy the analytical platforms using these advanced technologies.
Several attempts were made to understand the domain and technical challenges to solve this cross organizational data sharing. Ultimately, all my attempts succeeded to develop this pilot project which motivated me to setup a new organization to solve these kind of complex problems. I am happy to be announcing the launch ‘aciana’ as a healthcare company. We leverage 360 degree data set pertaining to the patient to generate actionable insights and also assist the providers to give more meaningful and coordinated value based care. Our platform also helps to give precision and real-time personalized medicines. To know more about us and what we do at aciana, visit our website, follow us on the social media and contact us.