Ready AI: Radiology data has been in digital format for a decade | Photo credit: Álvarez
From health trackers and physical conditioning in smart watches to the growing number of health applications on mobile phones, their health data changed hands several times.
Some of them can be processed to give their doctor a complete image of their health. And, sometimes, it is used to develop technological solutions, which, together with artificial intelligence, can be used to plug in health gaps in the country, improve health results and patient care, discover efficient medications or map patterns.
And yet, access to health data, a delicate problem, remains a challenge, say health solutions providers, which seek standardization, among other enabling measures.
Dineh Koka, main founding officer of Onard Assist, a cancer analysis platform that helps pathologists in the diagnosis of cancer, underlines the importance of “standardized” data to allow new health companies to develop solutions based on AI.
Reflecting on nuances in the collection of medical care data, the data says that the data generated from biometry have a bone good in the last 20 years due to the adoption of laboratory information systems. The challenge, thought, is not always “structured” or “uniform”, namely, there are variations in units and labels (such as differentiation measurements for sugar levels). “This requires normalization and standardization,” he says.
Image dates
The image comes with its own complications, since it implies the use of cameras to collect facial details, including oral and eye (background or retina). “Capture these data using simple or specialized cameras are consulted quite easy. The challenge lies in using these data for analysis or AI. Factors such as the type of camera used and lighting can import,” he says.
Another child or medical images is radiology, which involves X -rays, computerized tomography or ultrasound. “Radiology has an advantage since the data has been digital for 10-15 years, with radiologists that report on work stations instead of a film. This digitalization and the good quality of the images, along with a decent amount of Ai-Ai alterithms.
The third important data source are the tissue pieces involved in the pathology. “There is detailed information on tissue samples examined under a microscope after adding a reagent. Traditionally, this leg of hashes in analogical way, existing as slides or glass reports written on paper or as text in laboratory systems. This poses data data data data data data data data years, driven by the decrease in costs, “he says.
Reliable data
Akkiraju Bhattiprolu, Director of Technology at Vigocare, a patient monitoring platform promoted by AI, says that data sets are not easily notable in India, unlike countries such as the United States. In addition, there are groups on the source and reliability of the data. When developing solutions, models trained in data from other regions should be tested in the data of Indian patients to guarantee troops and meet the regulatory requirements.
But there are also advantages, says a starter founder, who points out the relative ease and profitability of generating new data through associations with hospitals, following ethical and legal processes. “Startups can collaborate with hospitals, sacrifices services in exchange for access to data, which helps overcome the limitations of pre -existing data sets. However, the initial lack of standardized and easily aviable data remains a key challenge,” he says.
Telangana made a directory, offering new companies access to medical care data to train AI models to help solve real world problems. “There has been great interest after announcing our decision to give access to new medical care companies. Many of them have approached us,” says Jayesh Ranjan, secretary (TI and Industries).
Confidential data
Pointing to personal identification information (PII), including the patient’s name, the telephone number, the professional and the work details, which is highly sensitive, Akkiraju says that the combens are required to adhere to regulations such as European restrictions and predation predation. “They impose guidelines on PII handling, including geographical restrictions on data storage. Data related to vital (temperature, heart rate, respiratory rate and ECG) can be used after it is anonymized, which means that it is to astararate that it is storing storing storretarate storretarate storretarato Estorretarate Estorretarate Storreta Estorretarate Estorretarate.
“Current efforts to create large and cured data sets are mainly financed by subsidies and led by public institutions and hospitals. I think that new companies should also make government subsidies build data sets, which could make the resources of the aerial resource of injury in a beee axes athes to the Therm forces in the resources. Space.” Koka
More like this
Posted on April 20, 2025
