CXOToday has engaged in an exclusive interview with Mr. Ganesh Prabhu, Vice President – Sales, Indium Software
1.What are some of the data challenges the healthcare sector is facing in India?
Modern healthcare setups have broadened their data sources to deliver comprehensive care. This has resulted in enormous amounts of unstructured/semi-structured data. Also, extracting and analysing this in the healthcare ecosystem is challenging and requires advanced systems and models to help derive all the vital information needed to improve health outcomes. Several healthcare facilities need more resources to leverage critical information to enhance healthcare services.
Some of the key challenges include:
- Data aggregation from disparate clinical and health insurance systems
- Health insurance and Clinical data convergence to enable quality measures
- Healthcare data translation across various formats & structures (for e.g. EDI, HL7, FHIR, CCD)
- More importantly, with the Ayushman Bharath initiative that the Union government has embarked upon, there is an urgent need for “Health Information Exchange” hubs to transact the data between the government and private care providers to deliver comprehensive care.
In this context, AI has the potential to bring in automation, reduce manual intervention and deliver real-time insights from the data.
2.Can you elaborate more about the adoption of NLP in Healthcare in India?
Becoming digital in healthcare does not mean just setting up a computer network but also integrating technologies such as AI & ML to analyse data & provide meaningful insights.
As a subbranch of AI, NLP is gaining good momentum across the healthcare sector in India. NLP-based solutions have immense capabilities to process human inputs and reduce workloads of physicians and administrative tasks. Patients can gain information on specific health conditions and plan their doctor visits if symptoms seem severe. The increasing use of smartphones and websites for booking appointments and even conducting online consultations provides tremendous opportunities for NLP integration.
Several leading hospitals in India are leveraging multiple Electronic Management Records (EMR) systems to capture a huge amount of patient and clinical data. Most of these data are unstructured and semi structured and need AI and ML techniques to interpret data faster & accurately to provide patient insights. As healthcare sector transitions to value-based care (focusing more on the qualitative outcome rather than the traditional models), one of the critical success factors is changing the Patient experience. NLP systems can analyze these unstructured data and derive meaningful medical insights that helps 1) Physicians makes better clinical decision 2) Record Linkage and Genome analytics can help devising personalized care
Especially post Covid, the new normal of Tele-Health, Virtual care giving mechanisms, Patient outreach based on risk stratification gains huge traction and the NLP supports immensely to handle the different data sets.
Leveraging our proprietary AI/ML based NLP solutions teX.ai, Indium has helped a large third-party provider of data by extracting reports from EMR dumps with upto 95% accuracy and redacting of PII / PHI data from medical records with almost 100% accuracy. We have also helped a documentation services provider with the automated conversion of tele-health interactions between physicians and patients into a claims-submissible medical record format with 90% accuracy.
At the heart of digital evolution, AI / NLP based solutions in the healthcare landscape are poised to make a significant difference in the future.
3.What are the key trends you’re foreseeing for 2023 in this sector regarding technology?
The pandemic has forced healthcare organizations to rethink the way healthcare is delivered. They are now readily investing in technology-driven models such as Tele-Health, Virtual Care, and Remote-Monitoring. Healthcare organizations are also starting to adopt new-age technologies such as Cloud, IoT, AI / ML, AR, VR, Blockchain, etc., at scale to improve customer experiences and drive operational efficiencies. With technology advancements, health care experts are now focusing on taking the point of care to the patient, especially for primary care.
- Interoperability between systems is now at the top of the priority list. Globally we see governments, regulatory & standards bodies working with healthcare players to implement regulations & standards to enable interoperability by leveraging advancements in data standards & technology
- AI/ML Chatbots: Conversational AIML BOT solutions to monitor remote patients post discharge and record medical conditions periodically
- Image Analytics: Physicians needs sophisticated tools that can analyse various Images (X-Ray/CT/Digital Scans) and provide pointed issue while treating chronicle care patients with multiple problems and Oncology patients
- Wearable Medical Devices : As Internet of Things (IoT) has gained huge momentum and most of the smart watches offers SMART healthcare kit that can track the BP rate, SPO2, Heart rate, Pulse and possibility if patient can get tremors and so on, Healthcare tech firms leverage IOT solution to drive preventive care management
- Personalized care management: Traditional caregiving models are slowly getting replaced by more personalized care leveraging Precision Medication, Digital Twins kind of trends and technology. This implies constant remote monitoring of the patient and have required intervention mechanisms and reducing the probability of getting into emergency rooms.
4.What does Indium software do for healthcare providers in India?
Indian healthcare is complex due to its rising population, rapid urbanization related changes in people’s lifestyles and lack of basic healthcare infrastructure, especially in the rural areas. Some key stats
- In India, the physician-to-population ratio is 0.8, compared to the WHO’s recommendation of a minimum of 1:1000 and a maximum of 4:1000.
- There are 5 hospital beds for every 10,000 people.
- The ratio of ICU beds – 6.88/ 100,000 people
- In both the above cases, rural accounts for only 31% of beds while hosting 65% of our population
- Number of Registered Physicians in India is ~13.5L & Number of Specialists ~3.5L
- Again 80% of the physician concentration is in Urban areas
Consider an emergency scenario in a remote area where the first point of contact is the Primary Healthcare Center. Based on the condition, the patient is often referred to a district hospital for secondary care or a tertiary care hospital in the city, resulting in the loss of precious time & in some cases, life. This situation can be better handled with a service such as Tele-ICU, with anytime anywhere access to specialists specializing in critical care. Especially with the advent of 5G services, Tele-ICU is slated to play a major role in improving healthcare.
Indium has an experienced team of domain & technology experts who have worked in the healthcare sector. During the pandemic, Indium engaged with a healthcare technology company which provides platforms & products for care providers such as clinics and large hospitals. The client wanted to leverage Indium’s healthcare experience & digital engineering expertise to develop a virtual care platform to address the needs of ICU patients, as physicians could not move beyond their assigned zones within the hospital due to pandemic restrictions.
The product developed has the below features:
- The bedside physician registers the patient as soon as he is enrolled into the ICU, including adding specialist physicians required for the case
- The conversation can be through multiple digital channels such as chat, audio or video through a phone, tablet or web interface
- The app is integrated into the clinical ecosystem via HL7 integration, through which the physicians get near real-time access to patients’ critical parameters
- The app is also integrated with the medical records systems allowing the physicians to access patients’ care history and past episodes of treatment
- Post the consultation, the physicians can update the patient records. Also, they can prescribe medications/diagnosis and procedure details into the electronic medical record systems of the hospital
- The app also has built-in voice-to-text functionality that reduces administrative overheads and allow Physician to focus more on engaging the patient
Virtual care is a milestone in the healthcare industry as it widens the reach of care in rural areas, where there is a lack of specialized critical care access.
In addition, Indium also helped to develop the Voice to Text capabilities for Telehealth consultations. This app is based on Indium’s AI/ML-based NLP accelerator teX.ai. This app helps convert virtual medical consultation transcriptions from voice to a structured medical record covering patient demographics, vital signs, diagnoses, medical history, lab orders, prescriptions, etc.
5.Can you share few examples of how Indium software approaches NLP?
Indium’s proprietary, home-grown AI engine “Tex.AI” comes with prebuilt solution for PHI/PII redaction, concise data extraction, data classification and clustering from unstructured data and clinical notes, audio to text conversion and derive clinical means based on the doctor-patient discussions, segmenting various drug categories based on the dosage/strength. Tex.AI is proven, time tested solution accelerator that provides 10x faster time to market than conventional solutions, reduce TCO by 5x and provides 90% accuracy based on the programs delivered.
Some of our key success stories are provided below:
- Support large third-party provider system by extracting reports from EMR dumps with up to 95% accuracy and Redaction of PII / PHI data from medical records with almost 100% accuracy
- Successfully converted Physician – Patient audio files into text and developed meaningful clinical interpreted notes for clean claim submissions
- Reduced Physician/Nurses administrative time by automatically updating the Electronic Medical Record systems through speech to text conversion process
- Developed “Smell digitization” module ( outside Tex.AI) to ingest live stream data from an Olfactory sensor to detect the abnormalities in the air. Can be extended for detecting onset of cancerous disease at an early stage as a few organizations are exploring disease identification through the sensory capabilities of pet animals
- Classify the Patient disease categories, Medication details into broader prediction models to drive preventive care management
- PII/PHI redaction of Clinical, patient & Clinical trail data for information exchange between various research based healthcare organizations