Health Tech Ecosystem: How Data And Analytics Are A Crucial Part Of It?

A Robust Health Data analytics engine is therefore extremely essential to derive actionable insights from large swathes of data.


Tavishi Dogra
Written by: Tavishi DograPublished at: Jan 25, 2020Updated at: Jan 25, 2020
Health Tech Ecosystem: How Data And Analytics Are A Crucial Part Of It?

One of the biggest issues that healthcare suffers from across geographies is that of data asymmetry. This is caused due to the usage of various data interfaces and health records across various providers, payers and individually at patient levels. Specifically, in India, modernisation and digitisation drive across doctors specifically in Tier 2-3 cities is pretty nascent and as such the health information is disparate and available in a non-standardised format. This problem poses a challenge when it comes to research, medication, the proclivity of an individual or customer demographics in their receptivity to a particular ailment, medication.

Further, when we look at this problem from the provider perspective, there is a 2 pronged need to digitise and analyse data:

Also Read: Remove Gel Nails At Home Without Damaging The Nails With These Simple Home Remedies


At customer onboarding Stage

The insurers have to underwrite a prospect basis financial and health status. This allows insurers to decide on allowing a customer at standard/non-standard premium rates or in a worst-case scenario decline/ accept the policy with exclusions. This is to ensure that they do not underwrite substandard lives.

At Claims stage


Insurers need to understand the claims patterns and accordingly frame. The propensity to Claims model to attune their claims experience. Similarly, at the provider level, revenue cycle management, follow up on treatment, etc, is heavily dependent on patients EHR and digitised records. Therefore, there is an inherent need to digitise, standardise & analyse data for understanding correlations across the payer, provider and life sciences worlds.

Robust Health Data Analytics Engine


  • Data stored across healthcare systems is multivariate (multiple parameters), multi-transactional (more than 1 interactions) and multidirectional (Payers to providers, providers to payers, payers to TPAs, etc.). This complexity gets multiplied when we analyse data over years, to study a specific data cohort basis a trigger, event and passage of time. A Robust Health Data analytics engine is therefore extremely essential to make sense over large swathes of data.
  • An increasing number of companies are using this data to understand correlations using multivariate analytical models. This needs to back by a robust backend database, disposition-based frontend which gives the end-user the flexibility to adjust the cohort basis a specific data dimension or a time series. A robust engine is also able to track rules and create alerts or a sequential workflow, basis the particular trigger in the data series. Health Data analytics engine is supposed to study past data and predict future outcomes from seemingly unobvious data correlations.

Some of the used cases in the industry for analysis

  • Risk score for underwriting
  • Claims risk score or Propensity for claims
  • Predictive claims analytical model

What caused the shift and emphasis on Data/ AI & ML/ Analytics


  • Usage of Artificial Intelligence & Machine learning to augment analytics and provide more contextual decisions and output to the end-user is now a global phenomenon. No tech expo or meeting is now complete without the usage of AI/ML-powered discussions. Traditionally codes and analytics would be basis binary values and trigger-based rules of If and Then. That in itself is controlled programming, however, given the explosion of data, usage of language, varied methodologies of documentation, there are only so many rules that can be configured, basis user experience and comprehension.
  • Artificial Intelligence is controlled programming which goes beyond the realm of If and Then rules and Machine learning self-trains it rules algorithm basis understanding of a large dataset and calibrates to understand patterns which are not manually discernible. Simply add fuzzy logic to the mix, and you have a program that understands non-binary values. This adds a human instinctive layer to the program.
  • A classic use case, the Google Assistance discover page shows you more and more of the same type of article basis what you read and keeps on weaning away articles which you dismiss. Similarly, Alexa listens to your audio commands, transcribes them and provide you with the nearest configured solution. However, it has improved more and a period using AI.
  • By analysis and predicting the possible risk exposure of an individual, the system is then able to decide how to make the health recommendations customized to the needs of an individual. Furthermore, basis health status, an individual can be given specific recommendations to manage lifestyle by managing diet or exercise regimen or reduction/increase of specific foods. Curated content specifically for the individual can be made possible basis this multivariate analysis.
  • This means every individual within a specific cohort may receive more reference-based contextual treatment. This movement is largely spurned by the need to steer away from the One Size fit all to a customer bespoke program specifically suited to an individual's need.

(With inputs from Mr Adrit Raha, CEO, Vivant)

Read more articles on Miscellaneous