Search engines, shopping apps, and media streaming services have improved tremendously over the last decade. They are seamlessly integrated into our lives today. Much of this can be attributed to the steep advancement in the underlying intelligent technologies behind the products. These technologies have the ability to not only contextually understand domain data, but also, a person’s preferences, intents, and behavior. However, digital tools in healthcare, specifically in India, are yet to undergo such a transformation. It is ironic that Google can show relevant results on a web-scale for vague queries with typos, whereas doctors are expected to specify the exact name and related details to fetch drugs on their digital prescription pads.
We often hear about the importance of digitization in healthcare. However, a mere ability to store data digitally doesn’t lead to all those benefits. A meaningful way to represent, store and relay healthcare information is needed. Such systems should be able to handle data in a semantically rich manner enabling computers to understand, interpret and allow interoperability across tools being used by different stakeholders. It is only then that the adoption of these systems will not put demands of drastic behavioral change on the user - doctors and patients alike.
The figure below illustrates how a move towards semantically advanced technologies has helped boost productivity (reference).
It is precisely for these reasons Eka uses global standards like FHIR and HL7 at the core of the platform. We also use coding standards such as SNOMED-CT, LOINC and ICD-10 to represent and store healthcare data. In the last decade, the Ministry of Health & Family Welfare (MoHFW), Government of India, has also promoted these standards. MoHFW has worked closely with SNOMED-CT to define reference sets specifically for the Indian context. A list of prominent resources can be found here.
Eka uses these coding systems to represent domain concepts in its medical knowledge graph, which is augmented with practice-based medical data. Our knowledge graph comprises entities such as symptoms, disorders, laboratory tests, procedures, drugs, and their salts. These coded entities and understanding of semantic connections between them is what makes Eka a medically aware software. Below we see some examples of how these decisions benefit doctors and consumers.
SNOMED-CT provides an encoding for hundreds of thousands of medical concepts through a well-defined ontology (reference). This allows us to understand the medical concepts in terms of their underlying attributes and relations with different body parts, their causing agents, morphological structures, etc. For example, we now understand that Pneumonia is a lung consolidation and has an associated morphology of inflammation. The representation below is taken from Shrimp browser for SNOMED-CT.
This semantic understanding of medical terms is what allows us to offer doctors a seamless way to add information related to symptoms and diagnosis while making a prescription. While this saves a lot of time for doctors, for patients, this results in a rich medical history that can be leveraged in future medical encounters. Below is an example of a symptom; Chest pain, we see how relevant information such as the direction to which pain is radiating, what aggravates the pain, ECG finding, etc are pre-populated for a doctor to quickly add the information.
We also utilize this intelligence to contextually recommend probable diagnoses given the symptoms, which helps reduce the time it takes to generate a prescription since it now involves minimal typing. An example is shown below where for symptoms such as Heachache, Fever, and Cough, recommendations are given for other associated symptoms and probable diagnoses such as Flu, Pneumonia, and Dengue.
AI-driven symptom checkers also utilize SNOMED-CT in different ways. Based on the primary symptoms reported by a patient, symptom checkers dynamically inquire about the related symptoms and associated nuances. Such tools not only facilitate patients in providing a detailed log of their medical conditions, but also aids doctors, who can get these symptoms pre-filled in their prescription pad, and hence the doctors can now spend their valuable time on a richer conversation with the patients.
There are numerous other advantages of representing medical concepts and storing records based on these standards. Imagine the ease with which doctors and patients would be able to retrieve their past medical encounters. With Eka’s ability to understand medical concepts, a doctor within a fraction of a second can retrieve all prior prescriptions of a patient for their heart or chest-related problems for instance. This not only would facilitate a better healthcare outcome for a patient but would also improve the doctor-patient relationship.