As data science capabilities take a more central role in advancing innovation in life sciences, access to high-quality datasets across functions is crucial to remaining competitive and improving patient outcomes. Whether an enterprise is dealing with proprietary data, de-identified EHR system records, supply chain management data, or any other relevant dataset, it can be challenging to understand how to make such data accessible across vendors and partners or gain greater insight from partner data.

Historically, many companies had policies against sharing data or required convoluted security and approval processes to make data accessible across units or externally. One must navigate contracting, competitive concerns, information security reviews, go-to-market considerations, and more. Add heightened regulatory and reputational risk, lengthy approval processes, unique technical requirements, and legacy data systems to this process, and you have the ingredients for a difficult, if not impossible, challenge to effective data collaboration.

As data assets are increasingly digitized and employees have developed their data science skill sets, technologies have emerged to give covered entities handling personal health information (PHI) or other sensitive datasets peace of mind during data analysis. One such technology: the data clean room.

How Data Clean Rooms Work

The purpose of a data clean room is to allow employees from different internal teams or external organizations to analyze datasets without providing access to the raw data itself. Other clean room operators will have different technical approaches for how this works, but all reputable vendors will have:

• User controls dictating who can have access to which datasets
• Task controls dictating what operations can be performed on the data
• Standard security protocols dictating where data analysis occurs and where the outputs of analysis (typically aggregated outputs) will be written
• Low-code tools enabling non-technical users to access insights and orchestrate data usage across various business units

Several providers, including Habu (learn more here), also inject sophisticated privacy-enhancing technologies to reduce the risk of re-identifying individuals in a given dataset. Once all of these controls are in place, analysts are able to run data science tasks across provisioned datasets, without direct access to the data itself.

While this configuration has many applications across industries for data collaboration, it represents a fundamental change in available opportunities for life sciences companies to scale data operations, invest in co-development partnerships, and speed time-to-market for treatments. Let’s explore just some of the use cases across the drug development lifecycle where data clean rooms can unlock insights.

Collaboration Throughout the Drug Development Lifecycle

Clean rooms have the potential to reduce costs and time-to-market across the drug development life cycle by enabling greater access to unique datasets, data science talent, and partner data used to inform strategic decision making. Some of the ways in which clean rooms are leveraged for data collaboration include:

1. Preclinical drug discovery and development
2. Clinical trial recruitment
3. Post-hoc analyses
4. Co-commercialization planning

Preclinical Drug Discovery and Development

During preclinical research, it is common for scientists to leverage machine learning models to predict various molecular responses to treatment in order to de-risk the development process. However, these scientists are often limited to the data available to them within their own company. With the use of data clean rooms, scientists can form consortia with partners and other like-minded research organizations to “pool” data for use in developing more generalizable models with greater predictive power. MELLODDY is one such publicly coordinated effort demonstrating how these partnerships can work in practice.

Once a promising treatment is identified, consultants or in-house data scientists can leverage public and proprietary data to run demand forecasting models and incorporate manufacturer data to qualify the appropriate suppliers for development and distribution.

Clinical Trial Recruitment

One of the most challenging aspects of bringing a drug to market is finding the right patients to enroll as trial participants. Given the fragmented nature of healthcare delivery, recruitment has historically relied on a network of clinics recommending patients as and when they come in for treatment and are made aware of the trial. The patient enrolled:screened ratio remains relatively low, leading to inefficiencies in the recruitment process.

With the use of clean rooms, pharmaceutical companies and contract research organizations (CROs) can leverage their existing clinician networks to “pre-screen” patients based on past claims data, EHR records, and other medical data (e.g., imaging) as necessary to filter out patients who do not meet enrollment criteria and predict those who are likely to be a good fit for the trial. Analysts can also benchmark sites and identify optimal sites for trial care delivery. Clean rooms allow for this analysis to occur without sharing model IP or raw patient data. Once patients are identified, their physicians can initiate the outreach to confirm the screening and enroll them in the trial.

Post-hoc Analyses

Data doesn’t stop being a factor once a trial is completed! Post-hoc analyses also often benefit from use of real world data (RWD) or social determinants of health (SDOH) data used to analyze whether patients continue to adhere to treatment requirements post follow-up, identify potential long-term effects, monitor for off-label use cases, and analyze other post-trial risk factors. Not only can clean rooms facilitate these analyses, they can also connect enterprises with data partners who have unique data assets otherwise unavailable for analysis.

For example, Habu has relationships with several of the top grocery retailers in the United States who provide transaction data illustrating household food consumption patterns. This data can be used to analyze how dietary patterns might affect the post-trial condition of previously enrolled patients. Other data partners include facilitating easy access to public health databases, claims data, media consumption, and other such data sources for use in post-hoc analyses.

Co-commercialization Planning

It’s increasingly common to co-develop and co-commercialize treatments. Once those treatments have been approved, partners need to evaluate how they will go to market and plan the distribution rollout. This can be challenging to manage among partners because no one wants to directly share their historical sales data. Making both parties’ data available via a clean room means each party can benefit from the others’ data without revealing information they’d rather keep private. Pooling the data allows for more strategic go-to-market planning by territory and better media mix modeling for the expected rollout of the treatment.

Final Thoughts

Data collaboration in life sciences may seem daunting, even impossible, at times. Yet, the potential opportunity it unlocks throughout the drug development life cycle is vast. Data clean rooms aim to provide scalable, repeatable data collaboration infrastructure across teams and external organizations, without compromising on privacy or usability. Their technical guarantees mean you can overcome barriers to collaboration previously thought intractable, helping you get one step closer to greater innovation and better patient outcomes.

Habu data clean rooms allow healthcare and life sciences companies to confidently collaborate with patient-level data while protecting privacy and sensitive information. You can learn more here, and speak with a Habu representative today.