Pathnostics is a precision diagnostic solutions company that specializes in infectious diseases. We perform a variety of tests including novel genetic and phenotypic assays for pathogen identification and antibiotic susceptibility, anatomic pathology and cytology, Bladder FISH, and STI testing. Our flagship technology is Guidance®, which uniquely and rapidly measures pooled antibiotic susceptibility, with which we developed the market leading Guidance® UTI assay.
Guidance® UTI provides personalized results in approximately 24-36 hours from receipt at lab, while culture can take 3-5 days, reducing the need for empiric therapy and supporting antibiotic stewardship initiatives. Guidance® UTI does this by combining the sensitivity of PCR testing with proprietary Pooled Antibiotic Susceptibility Testing™ (P-AST™). This allows for more informed treatment decisions for the management of complicated, recurrent or persistent UTIs (urinary tract infections) and elevated-risk patients to better address polymicrobial infections.
Our company is already a leader in the urinary tract infection space, but we want to continue to stay ahead of the curve. At the core of this strategy to refine our market-leading products is our data. This can be a challenge with the large amount of unstructured data that is produced by clinical laboratory testing.
Due to IT and infrastructure development focusing on patient reporting and immediate laboratory data needs, the data is usually not structured for large scale analysis and is distributed across myriad data sources and formats. There are old laboratory SQL databases that were 10 years old, irregular CSV files from LIS and RCM vendors, Excel spreadsheets comprising R&D, Clinical Trials studies and Operations data. Despite newer SQL databases being developed by the IT department, most of our systems are built on this legacy data, which is used for day-to-day laboratory operations.
In order to build a data lake that works flexibly with all these discordant and unstructured data sources, we chose Databricks. The Data and AI Department reviewed several vendors and found that most alternatives worked better if you had structured data in data warehouses, which was less ideal for Pathnostics’ use case.
Databricks Data Intelligence Platform solves these problems that Pathnostics was facing. It’s extremely flexible, plays well with unstructured data, and helps us create structure from these disparate sources, while being comparatively more affordable than other alternatives.
We use Databricks to provide daily data updates to the Medical Affairs, Sales, Operations, and Research teams. We are currently analyzing millions of samples with complex multidimensional data over a period of three years and in the future, plan on not only increasing volume, but expanding to include samples as far back as 8+ years.
We leverage this information to improve our predictive models and clinical trials and provide client-facing teams with bespoke reports and data, to help providers advance their patients’ care. We have built our Data Lakehouse infrastructure on the Medallion Architecture using the Unity Catalog and we power all our offerings in Power BI as well as several custom reports throughout the company through this system.
We are building an AI offering for our clients that would predict antibiotic susceptibility based on patient demographic data and PCR results using the one-of-a-kind training dataset Pathnostics has access to, thanks to our P-AST product. The model is already patent pending and is a novel algorithm for the field of UTIs. This will allow providers to give patients actionable antibiotics with very short turnaround times as they wait for the phenotypic component of the assay to be completed. This is in the late stages of development and is planned to be released soon for our clients.
We create custom reports for Operations such as turnaround time breakdowns and quality checks for reports sent to clients. We also offer a product pairing we call the Guidance Practice Review and the Guidance Practice Summaries. Our Medical Science Liaisons and Sales teams, respectively, give practices and clinicians an overview of their data in a highly curated format: a custom-formatted PDF made using some Python methods involving LibreOffice and some advanced cluster setups that Databricks facilitates.
Thanks to Lakeflow Connect, we were able to integrate Salesforce data directly into our Databricks platform, so that we could pull Sales territory, region, and technology partner information from Salesforce. This was transformative in terms of reliability, speed and accuracy, as prior methods using the API were difficult to maintain. It is now available in GA in Databricks, making it production ready, and saving us time integrating and empowering our Sales team’s ability to better serve our customers with their data.
We have created a flagship product in Power BI we call the Guidance Analysis Report which gives our department and the Research and Medical teams incredible ability to deeply scrutinize and analyze our data. This report is used by multiple teams within the company, and it is critical for putting out fires before they start, diving into the most obscure QC cases, and discovering insights for clinical research and publications.
Most recently, we have launched a Sales Dashboard that is fully integrated with our LIS system and Salesforce, complete with advanced row-level security to provide every member of the Sales team complete oversight of their territories, regions and provider specific data, as well as their forecasted data and monthly/yearly goals, and leads in their territories.
To learn more about Pathnostics and our Guidance UTI product for complicated and recurrent UTIs, you can visit our website at Pathnostics.com