Shaping a future-ready healthcare ecosystem with AI and analytics

2 Sep 2022
Shaping a future-ready healthcare ecosystem with AI and analytics

In a recent presentation at the Indian Institute of Foreign Trade, Indegene’s Vice President of Analytics, Jitesh Sah, delved deeper into the digital adoption and innovation journey of global healthcare organizations. This blog covers the main highlights of his presentation.

A few years behind the technology curve

Healthcare lags other industries when it comes to digital adoption 1 . Multiple factors contribute to this (illustrated in Figure 1 and 2).

 

A few years behind the technology curve

 

Figure 1: The slow pace of digital adoption in healthcare

 

 

Reasons behind healthcare's slow pace of digital adoption

 

Figure 2: Reasons behind healthcare’s slow pace of digital adoption

 

While the pace of digital adoption has been slow over the years, things are picking up. Healthcare organizations are ramping up their AI investments 2 (Figure 3), and a lot of this is driven by the demand for modern healthcare systems and online engagements by customers.

 

Healthcare organizations are ramping up their AI investments

 

Figure 3: Healthcare organizations are ramping up their AI investments

 

How healthcare investments are directed

Investments in healthcare are moving beyond telehealth platforms to areas such as virtual patient monitoring platforms, digital therapeutics, surgical robots, AI-based drug discovery, patient identification, recruitment, and monitoring of clinical trials - all of which play an important role around:

  • Strengthening clinical decision making
  • Reducing healthcare expenditures
  • Reducing administrative costs
  • Driving customer retention
  • Enhancing marketing and branding
  • Creating efficiencies in filing and billing processes
  • Delivering financial support through robotic investment recommendations, mobile payments, credit services etc.

However, at the very core of this transformation lies a foundational piece - data. Data is the bedrock of intelligence upon which healthcare business processes and analytical layers are built. Without a solid data foundation, AI machines will struggle to deliver intelligent and trustworthy insights and recommendations.

How healthcare is leveraging data

The volume of healthcare data trickling in every day is rising exponentially. A single patient generates nearly 80 megabytes of data each year in imaging and electronic medical records 3 . This massive trove of data has obvious clinical, financial, and operational value as it allows organizations to unearth golden insights hidden deep beneath the data surface.

"To let all this data paint a larger picture, healthcare organizations are leveraging modern data platforms, such as a data lake house, to connect large volumes of patient or consumer data from disparate sources, unlocking a comprehensive, 360° view of their key customers," Jitesh said, adding that these platforms are fuelled by powerful data discovery, management, quality assurance, governance, and security processes. This helps organizations strengthen their data pipeline orchestrations, Data Operations (DataOps), and Machine Learning Operations (MLOps) at scale.

Healthcare organizations need a holistic data management platform that spans their entire enterprise and serves a number of critical functions, including ingestion of all data streams and sources such as electronic health records, standardization of data through industry code-based classification, standardization of frameworks to send patient data to upstream and downstream workflows, and more.

"By serving these critical functions, healthcare data platforms help you maintain the consistency in data management processes. It allows you to effectively take control of your entire patient and consumer data landscape and gives you access to actionable insights in real-time – all through a single gateway." Jitesh said.

 

Key elements of a robust healthcare data platform

 

Figure 4: Key elements of a robust healthcare data platform

 

Real-life applications of data-driven AI

  • Machine learning (ML) and advanced analytics

ML and advanced analytics, combined with data, are helping organizations run targeted clinical trials, transform drug discovery, and control treatment outcomes effectively.

"Take chronic diseases as an example. They do not follow a linear progression. They may get worse, better, worse again, better again, and so on. Understanding what influences the progression, predicting when the progression is going to occur, and deciding what interventions can be made to impede the progression are essential to determining effective treatment strategies. The use of data, ML, and advanced analytics models here are crucial as they assist organizations through the faster diagnosis of diseases, early detection of adverse events, prevention of medical errors, and more," Jitesh said.

Here are a few more examples of how AI/ML and analytics are driving positive healthcare outcomes:

AI/ML uses cases in healthcare

 

Figure 5: AI/ML uses cases in healthcare

 

  • Natural Language Processing (NLP)

NLP helps structure valuable information from unstructured biomedical literature. Deep learning models such as Generative Pre-Training, Bidirectional Encoder Representations from Transformers, Multitask Unified Model, and Pathways Language Model can help create large knowledge graphs (with ontology mapping) from a massive corpus of textual data. The knowledge graphs can then support several downstream tasks, including medical content processing for patient safety incident reporting, regulatory compliance, commercial content creation, and more.

In addition, NLP is also being used to extract information on patient treatment patterns to identify drug switching or discontinuation. This is done by extracting numerical data such as lab values, dosage information, and patient-specific details such as disease history, demographics, social factors, and lifestyle. NLP is also used in Automatic Speech Recognition (ASR) medical systems where virtual and in-person patient/clinician interactions are captured and injected directly into medical charts, enabling real-time medical transcription.

  • Virtual Care 2.0

AI/ML has shown promise in driving innovation for a Virtual-First Future (also known as Virtual Care 2.0), where patients start their care journey with digital interactions. Advance ML models deliver clinically-relevant and accurate predictions of patient progression or deterioration, enabling healthcare organizations to more easily triage the patient’s care needs, determine the best care setting from in-person to virtual, and drive follow-up activities.

 

Key components of the Virtual 2.0 framework

 

Figure 6: Key components of the Virtual 2.0 framework

 

  • Metaverse

Immersive and interactive platforms like the metaverse are poised to make interactions more fun and engaging. Here are a few examples of how metaverse is set to transform the healthcare industry  4 .

A few examples highlighting the promise of metaverse in healthcare

 

Figure 7: A few examples highlighting the promise of metaverse in healthcare

 

The way ahead

The impact of digital transformation and technology in healthcare is clearly being realized in many areas. By investing in the right kind of talent, skillsets, digital processes, and best-inclass technologies, organizations can accelerate their transformation journey at scale and help shape the future of healthcare.

Reference

  1. Driving impact at scale from automation and AI, February 2019 | McKinsey
  2. Global-survey-The-state-of-AI-in-2020, November 2020 |McKinsey
  3. Better Patient Outcomes Through Mining of Biomedical Big Data | Frontiers
  4. Metaverse: The Next Iteration of Online | Indegene