Six Cloud Data-Platform Approaches Making A Difference for Collaborative Healthcare Research

The following is a guest article by Nicholas Merizzi, Principal, Deloitte Consulting LLP, Todd Konersmann, Principal, Deloitte Consulting LLP, and Diana Kearns-Manolatos, Senior Manager, Deloitte’s Center for Integrated Research.

Data is critical for researchers to analyze attributes, understand trends, determine statistical significance, predict future scenarios, and report to regulators. As we see clinical research become more collaborative to bring in real world data to overall improve the quality of healthcare, these cloud data-platform strategies are making the difference.

The story of clinical research and public/private partnerships during the pandemic will be one for the history books and potentially a break-the-glass scenario for R&D, as organizations embrace cloud-enabled data platforms to support collaborative research. A key piece will be the ability bring in real-world patient data from healthcare organizations, and in turn, to use that data ecosystem to improve patient care.

In fact, a Deloitte survey of leaders from 17 pharmaceutical companies found that 94% of respondents believe using real-world evidence (RWE) in R&D will become important or very important to their organizations by 2022, with a clear shift to cloud-based platforms as the preferred method to manage this data. Cloud can mean software-as-a-service, data centers, enterprise platforms, and external distributed ecosystems.

However, many research organizations still have data stored locally, in silos, or stuck within organizational boundaries, limiting the potential of data as a reusable asset. In a sector where human genome data is expected to reach 40 exabytes by 2025 and in which Deloitte’s research has shown shared infrastructure and resources can reduce the research cycle time by 13-18%, embracing cloud data platforms could make all the difference.

In fact, 55 percent of individuals see data modernization as a key component of or reason for cloud migration, and 65 percent of those in the C-suite do.

The following are six cloud-enabled data platform approaches for collaborative research as life sciences, healthcare, academic, and government organizations look to work together.

Six cloud data platform approaches

  1. Enterprise data management – While valuable data is generated in labs and hospitals, it’s important to remember that physicians are experts in research and patient care – not technology infrastructure. Therefore, it is common for valuable data to be saved locally and with data scalability limitations. For example, light sheet fluorescent microscopy generates 3D images using infrared light that while progressive for embryonic research, creates images that are 5TB each. A research team may only be able to save six non-shareable, non-searchable images locally.

A Data Cloud makes research data shareable, customizable, and reusable across the organization for millions of records and converts physical data like x-rays to digital formats in milliseconds. Using cloud ML tools that scan x-rays for personally identifiable data and can blur or alter that data, healthcare organizations can be better equipped to share valuable patient data in an anonymized or obfuscated way. Additionally, with “data sharing neighborhoods,” individuals that might be interested in cross-domain research can draw upon the same data based on access permissions.

  1. External data aggregators, exchanges, and marketplaces – Data Exchanges are another way that organizations can tap into already commoditized clinical trial and healthcare data, including public electronic health records and medical imaging data from pharmaceutical companies, healthcare providers, health insurers, and partners to enable clinical trials, drug research, and more. Some of the resources available are data related to drug switching patterns, fitness benchmarks, life expectancy benchmarks, aggregated claims data, and other resources that could be useful for benchmarking or AI model creation. In 2020, as a result of the pandemic, cloud providers have expanded the depth and breadth of these services complemented by cloud-enabled virtual health solutions.
  1. Open and interoperable data platforms – Solution architecture design is another way that organizations can think through their cloud data platform strategy with collaboration and information sharing in mind. Programs such as the Interoperability Readiness program aim to progress interoperable healthcare data sharing with open frameworks. A key aspect of the initiative is to achieve healthcare transformation goals enabled by more interoperable data models and with application program interfaces (APIs) as the “foundation for interoperability.” When pursing an API-first strategy, think about establishing baseline and incremental controls for security, managing shareable and non-sharable data, and categorizing data for regulatory purposes.
  1. Cloud data warehouses, storage archives, and more – Cloud data warehouses, data lakes, and data centers provide an opportunity for a collection of data to be shared at the enterprise level (or across a network) through database as a service or a collaborative digital ecosystem. Importantly, the cloud encrypts data at rest and can manage high input/output requirements with the ability to scale up and scale down data storage and computing power as needed to bring terabytes or even petabytes of data off of local computers and servers and to address long-term storage and archiving needs. To advance healthcare, MedTech organizations are using IoT devices such as “wearable” smart hospital beds to assess data and monitor patient movement and automatically adjust the bed to comfort patients and reduce bed sores. Additionally, organizations are centralizing voice data in the cloud to analyze patient incidents, identify patterns, and improve quality of care and response time.
  1. Data ecosystems – For those organizations that do want to share data across a network, creating a data ecosystem via a data platform or shared data center accessible to those on the network is also an option. This model allows for a different type of collaboration. A data ecosystem can allow all participants to collaborate across a shared data network. For example, a consortium of private companies developed a cloud-native platform with data storage, ingestion, and 300+ searchable resources to enable secure, scalable, and collaborative research.
  1. Data Science and Machine Learning Platforms – While plenty of organizations approach AI and Machine Learning on premise, a growing number are exploring the potential of AI and the cloud due to newer cloud services that harness cloud’s scalability, model management, pre-trained models, and automation options. One hospital network, for example is using Cloud ML to store, compute, and analyze patient data for advanced diagnostics. At the same time, one of the barriers organizations will need to overcome for global collaboration is language. Pre-built advance services offered by hyperscalers will allow more efficient processing or large-volumes of data through image and pattern recognition and offer better global teaming by translating large-volumes of text to any language. Additionally, healthcare organizations have increasingly started to explore using internal talent marketplace platforms enabled by AI to manage supply and demand across shifting workforce needs.

Healthcare organizations are maturing in their use of cloud, AI, and other advanced technologies to share clinical notes, lab reports, pathology images, radiology scans, wearables data, and more. Deloitte research shows that 84% of physicians expect secure, efficient sharing of patient data integrated into care in the next five to ten years. As public and private healthcare institutions alike look to modernize their data infrastructure with the cloud, these different approaches can help advance the modernization journey without defaulting to a one size fits all approach. Depending on which approach works best, they can gain greater data interoperability and scalability, tap into AI services, and work in a more collaborative way across a shared data-sharing ecosystem.