Introduction

With more than 10,000 employees worldwide, this global pharmaceutical company takes a patient-first approach to therapeutic development, focusing on neurology and oncology. In addition to running clinical trials and producing four global brands, its researchers are exploring how to measure biomarkers for Alzheimer’s treatments and leverage DNA sequences to fight cancer more effectively.

The organization operates more than 40 subsidiaries, and has R&D, clinical research, and production facilities around the world. To improve patient outcomes, scale cost efficiencies, and fuel drug development and innovation, it needs to efficiently share data across regions and lines of business, comply with industry regulations, and leverage predictive analytics and machine learning on sensitive health data, without compromising patient privacy.

Challenge

The company’s ability to deliver the best service possible to its customers through data was inhibited by three primary issues: an outdated data storage solution, inaccessible and siloed data, and enterprise-wide reliance on sensitive data use.

Siloed data use among business groups and data owners proved to be the most pressing issue. The oncology and neurology teams stored and used distinct types of data, ranging from patient health statistics to medical and X-ray imagery, which required compliance with a long list of compliance laws and regulations. To manage this in a global organization, teams established different controls for different projects, leading to complex data silos and access management processes that were manual, time intensive, and risky.

Exacerbating the issue was the compliance team’s lack of visibility and authority across the organization. For example, data containing personally identifiable information (PII) and protected health information (PHI) was frequently downloaded to individual workstations and shared via email, and third-party tools were installed without proper vetting for security compliance. These risky practices were not a product of the data team’s ignorance, but rather a result of a lack of methodology. To truly scale operations and deliver meaningful solutions while maintaining patients’ trust, the organization needed a more streamlined, comprehensive data strategy.

Solution

To start the process of building a new data strategy, the data team worked backwards. Considering general company goals and collaborating with data scientists and analysts across departments, they determined that a cloud-based, big data repository with a global catalog for medical images and textual data would best suit their short- and long-term goals. Working in tandem with a unified data science and BI/analytics platform, their approach needed to be capable of:

• Enabling atomic storage
• Producing a detailed data audit trail
• Querying data at a point in time
• Securely sharing data across multiple business units
• Analyzing, reporting, and dashboarding data use
• Supporting compliance with a range of privacy regulations

The team began its data transformation with data security and governance in mind. From the start, they knew that they wanted to leverage a data lake that would act as a repository for structured and unstructured data. Originally, they chose AWS as a cost effective, cloud-based storage solution with a wide variety of advanced compute and analytics capabilities. While AWS’s capabilities provided a strong foundation, they found that a more diverse data analytics solution was still needed to meet their goals.

To extend support for SQL, R, and Scala – and gain access to advanced auditing capabilities – the data platform team also adopted Databricks. Databricks’ Delta Lake provides an open format storage solution with widespread support for analytics languages, as well as dashboarding and auditing that is compatible with other leading reporting tools. Databricks’ flexibility and advanced reporting capabilities allowed the company to consolidate many of its central data analytics activities into a single, unified platform.

After establishing storage and analytics platforms, the team turned its attention to finding a method for securing them. Immuta ultimately stood out among competitors for its holistic approach to data security, which includes data discovery, access control, and data monitoring, as well as its ease of integration with AWS and Databricks. With Immuta, they were able to apply fine-grained access controls authored in plain language, dynamic data masking, and privacy enhancing technologies (PETs) without additional overhead, manual attention, or lengthy approval processes. The team also discovered that Immuta’s auditing and reporting were more powerful than their initially desired capabilities.

Result

Data lakes offer a high degree of flexibility and analytics power, but may become inaccessible or diminished in value if mismanaged. But by integrating Immuta into its tech stack, the company was able to avoid this outcome and instead leverage the full potential of its data lake.

Immuta’s automated data discovery and classification, data sharing challenges and data silos remain a thing of the past. Now, teams are better able to collaborate with sensitive data, without having to worry about data localization laws or industry regulations laws like HIPAA. Having a purpose-built toolset at their fingertips allows data users across all departments to work more quickly and efficiently toward developing the next pharmaceutical breakthrough.

With faster time to data access, more efficient security and compliance efforts, reduced data storage and management costs, and increased collaboration across teams, the organization’s data warehouse, security, and analytics transformation empowers improved patient treatments and outcomes.