HVH Precision Analytics: Data-driven Insights for Early Diagnosis
pharmatechoutlook

HVH Precision Analytics: Data-driven Insights for Early Diagnosis

CIO VendorSteve Costalas, CEO Oct 5th, 2016 marked an exciting day in the healthcare space with a new player entering the market.

HVH Precision Analytics (HVH) was built out of a unique partnership between Havas Health (part of Havas Group), a global communications company focused on health and wellness communications, and Perspecta (formerly Vencore), a proven provider of information solutions in the defense-grade analytics space for the department of defense and intelligence communities. The first-of-its-kind entity would leverage sophisticated analytics and predictive modeling with deep therapeutic area expertise to provide unique insight into the way healthcare companies identify and connect with patients and providers at any point in the product lifecycle.

The basis for the HVH analytics platform was developedover 40 years ago for the most sophisticated intelligence community agencies. These programs needed to assimilate vast amounts of disparate data— photographs, correspondent and non-curated numeric data – at a time when no one knew how to do that. They then had to organize and make sense of it from an analytical perspective, so that the end users could take action based on the insights it provided. Today, with the use of advanced machine learning, this process has gotten faster and the outputs much more robust. The underlying technology is data-agnostic, so it is not limited to insurance claims data. For instance, it can also take in electronic health records (EHR), lab tests, genomics, consumer, social media, and environmental data. “The databases we use have well over 200 million de-identified US patient lives. Some go back 12 years, others five or six. We can ingest these vast amounts of data, curate or collate them, and use machine learning (ML) to uncover very specific insights that can be tailored to each customer,” says Steve Costalas, CEO of HVH Precision Analytics.

Solutions

HVH has three healthcare solutions tailored to the pharmaceutical industry.

Together we can create a world where people with diseases don’t have to find therapies… Therapies will find them


Pharma. The first is a commercial product offering across the entire pharmaceutical drug development lifecycle. The company provides healthcare provider targeting and market sizing—what it calls Switching Analysis. “If a product is coming onto the market, but is third in line, we do a detailed analysis to show when the first two products are completed and identify the types of patients so that you know where the whitespace is to target in the market,” asserts Costalas.

Clinical Trials. The next offering is a clinical trial practice, where HVH works to assess and improve protocols and maximize trial enrollment. It assists contract research organizations (CROs) and sponsors regarding evaluating their protocols, inclusion/exclusion criteria, and location of the trial sites—based on geographic analysis of where the diagnosed patients are and where the undiagnosed patients are likely to be. HVH is often brought in to help trials that are already underway with established protocols, to look at it from healthcare provider’s perspective, to find additional patients and drive enrollment.

Payer. “The third leg of the stool is a practice on the payer side, where we can take algorithms developed on our platform and install them behind payer firewalls so that we can run them on payer systems and on their data, “continues Costalas. The algorithms can be used to find fraud and abuse, and get additional medical information to patients and their providers that need it proactively.

Approaching Patients

The precision analytics company goes through a three-step approach when working with clients—Patient Definition, Feature Extraction, and Prediction. HVH is data-agnostic has access to multiple large de-identified datasets in-house. This ensures that they can work on virtually any disease state without biasing the solution.
In the Patient Definition process, the firm creates an extensive set of criteria (ICD codes, procedures, prescriptions, providers etc.) to define the types of patients the client wants to study. For Feature Extraction, HVH designs algorithms based on the journey of the patients identified by the patient definition, and then applies those algorithms back to the original datasets to try to find people who have similar histories, but are not yet diagnosed. Every person in the dataset is ranked, relative to the model that has been created, and then further review is conducted to ensure that the patient indeed has the disease. For each project that HVH works on, it looks at the disease indication from an integrated analytical and medical perspective.

HVH has worked with a longstanding client on projects for ALS, which typically leads to death within three years. The main objective was to identify ALS patients earlier in their disease progression because ALS treatments currently on the market are more effective if started earlier. It has been generally accepted that ALS patients are symptomatic 12-18 months prior to diagnosis. In this study, HVH demonstrated that some patients are symptomatic 5 years prior to diagnosis. This provides the opportunity to diagnose patients and start therapies sooner.

HVH jointly presented these findings with its client and a well-respected Harvard ALS expert at multiple medical conferences in 2017 and 2018. There is a manuscript discussing these findings in review with significant national neuroscience journal. The idea of accelerating diagnosis in ALS – or any disease - is groundbreaking in the medical community. This work lays the foundation to improve the quality and length of life for those suffering from life-threatening disease for patients around the world.

HVH has a strong positive outlook going forward and many exciting things in their pipeline. The company is continually looking at different data sources in and outside the U.S., and while the data landscape today is radically different from what it was 18 months ago, continuing to identify new opportunities to bring in data and integrate that into the analytics will be key. In the last year, HVH has grown in terms of employee strength and now has social media and environmental analytic capabilities that can integrate with its other healthcare offerings and for non-healthcare clients. HVH is also building its Health Economics and Outcomes Research (HEOR) and government healthcare practices in areas that complement its existing work.