We had published our prediction solution termed System Response based Triggers and Outcomes Prediction (SyRTOP) earlier in January 2020 titled ReSurfX::vysen Yields Remarkably Accurate and Actionable Insights Using System Response to Triggers – A Drug Response Study. Recently we published more results based on validation of both the solution and productized functionality at scale together with broader implications and outlook of our functionalities roll-out for enterprise scale and vertical as well as horizontal applications in business chain in the healthcare sector titled ReSurfX in 2021 - Best-in-class Outcome Predictors, Innovation Catalysts and ROI Multipliers.
We took the opportunity to also clarify that the set of functionalities including the analytics modules, total system response based prediction modules used here forms an application configuration that is SyRTOP, as we found that the phrasing in the previous post on SyRTOP gave the opportunity for just the predictor module to be misconstrued as SyRTOP.
The at-scale validation of predictions from these functional modules leveraging AHT provided proof of immense value added by each functional module as well as the ability to predict far downstream events from even a single system status reporter. For example, SyRTOP predicted multiple instances of drug combination effects that had been recognized by the US Food and Drug Administration (FDA) as well as the American Heart Association (AHA) based on post-market surveillance (PMS). Those results also demonstrated the presence of over 30% errors in processed and compiled information used for making decisions by ML/AI solutions (naturally, by humans as well). Over the large scale of validation we also found that these characteristics of ReSurfX solutions leveraging AHT are the norm rather than exception. Given the multiple use configurations and deliverables from the product we accomplished these in approaches that are data-source, data-formats and data-storage infrastructure agnostic - simplistically relatable to the data-lake approach that is getting commoditized. As to development of functional modules AHT is data-source agnostic.
On commenting about those results our CEO, Suresh Gopalan, summarized the impact as follows.
- These results even from the small portion of results we have released through this use case proves every value proposition of ReSurfX to significantly improve outcomes, innovation and ROI of enterprises. The results address accuracy, novelty and robustness of both the analytics components and the prediction components, as well as advance prediction of outcome directions.
- In addition to creatively leveraging the powerful mothership technology, these results (i) indicate that our powerful data designs for validation is also major factor in success of using AI (data-based decisions in general and through ML); (ii) highlight the incredible value that can be realized from the new insights ReSurfX solutions can generate by way of elucidating outcomes considered 'Real World Evidence' (RWE) from what is termed 'Real World Data'. Such RWE are usually difficult to achieve even through targeted validation, whereas ReSurfX solutions obtained them using much simpler data.
- We further made customer-centric decision to make offshoot deliverables to provide accurate Knowledge Repositories (KRs) to enable our customers to build custom (and proprietary) repositories to improve their workflows that use other solutions, in formats that confer cost-effectiveness in reducing some repeat steps involved in generating predictions from AI solutions. These are incredibly valuable to enterprises not only in using our powerful and generalized adapt-for-need-modules made leveraging AHT, but also all other part of their workflows in an integrative manner.
We continue to add newer functionalities with capabilities to glean value metrics evident to our customers at all stages.