DATA-SHARING PLATFORM - How to Overcome the Challenge of Differing Digital Maturity Levels in Biopharmaceutical Supply Chains

INTRODUCTION

The past decade has brought new challenges to the single-source manufacturing model with the adoption of more flexible manufacturing networks, utilizing in-region capacity through a combination of outsourcing to contract organizations and the de­ployment of more flexible facilities utilizing single-use technologies.

This evolution toward a network of partnerships between drug sponsors and contract manufacturers has highlighted the challenges of technology transfer, capturing, monitoring, analyz­ing, and reporting manufacturing process and quality data as it relates to data integration between organizations.

Predictably, data sharing becomes more cumbersome as more partners and technologies are involved – with impacts not just on incidental process data, but also on critical quality and approval data that can impact product release and regulatory fil­ings. As a recent BioPhorum report (Vision for Digital Maturity in the Integration Between Biomanufacturers and Partner Organi­zations) indicates, companies share data at different levels of dig­ital maturity depending on their specific business priorities and internal digital maturity.1 Therefore, the efficient and timely de­livery of therapeutics with complex manufacturing needs cannot be achieved by simply unifying the entire pharmaceutical land­scape within the most advanced digital maturity level. A better approach is to implement a flexible data-sharing platform that can accommodate partners with different data management ca­pabilities.

FROM SINGLE-SOURCE TO AN AGILE MANUFACTURING NETWORK

The pharmaceutical industry is transitioning to an agile man­ufacturing network composed of acquisitions, collaborations, and partnerships to support an increasingly complex pipeline from dif­ferent therapeutic modalities and different geographies. This al­lows for an agile supply chain with increased capacity for producing pharmaceutical products using partners’ existing in­frastructure, including CMOs (Contract Manufacturing Organi­zations), that are more rapidly deployable and cost effective than building new infrastructure optimized for a given modality.

Furthermore, sponsors can improve their global presence and rapidly deliver therapeutics to new markets by leveraging these agile networks, especially in new modalities that require a complex manufacturing supply chain, such as cell and gene therapies.

As therapeutic pipeline complexity continues to increase and multi-regional clinical trials are conducted, sponsors are becom­ing increasingly reliant on partnerships with multiple CMOs for manufacturing, inventory management, and quality control.

BIOPHORUM: DIGITAL MATURITY PATTERNS

The new paradigm for pharma collaborations calls for an assessment of data sharing and transfer. To that end, BioPhorum’s IT member companies established a Digital Plant Maturity Model (DPMM) to determine the digital maturity with regard to the busi­ness capabilities and enabling dimensions of any given facility. This was later extrapolated to evaluate the digital maturity for data sharing between sponsors and contract organizations.

The scale consists of five patterns of maturity. Pre-digital communication does not involve integration but rather repre­sents the exchange of physical documents and phone calls or the direct use of a part­ner’s applications. This approach is often preferred by small organizations that lack digital operational systems or for small projects in which digital integration is not cost effective.

The next level is the manual extract, in which data exchange is manually car­ried out via email with manual export/im­port of data using Microsoft Excel, CSV/text files, or Portable Document For­mat (PDF) files.

Automation starts at level 3, also known as auto-extract, in which the con­tract manufacturer automatically transmits standard data sets periodically from its source systems to their partner while there is still a manual process to consume the data on the sponsor side.

Level 4 “auto-ingest” adds another layer to the system by automating the in­gestion of data into the operational sys­tems on the receiving end. Here, data is typically transmitted using machine-friendly data structures, such as XML, and delivered via RESTful API or SQL query access.

Finally, level 5 involves automated data extraction and ingestion in near real-time instead of being timed with batch re­lease or process step progression. Near real-time communication can be beneficial when partners wish to receive instant up­dates about manufacturing milestones or issues to work in tight partnerships. Addi­tionally, level 5 can include interactive services, such as slot scheduling in person­alized therapies outside of sharing process data.

CURRENT TRENDS & BOTTLENECKS IN DATA SHARING

BioPhorum’s Digital Integration of Sponsor and Contract Organizations team surveyed 31 companies consisting of large pharmaceutical sponsors, contractors, and IT support teams. A synopsis of the find­ings and subsequent survey insight con­firms these trends and expectations in terms of digital maturity.

As with the DPMM, there is typically an aspirational level (5), which is not what every organization may strive for. Level 5, real-time data sharing of manufacturing data, is not needed and is not useful in many situations. Instead, auto-ingest (level 4) is currently the main aspiration in terms of quality, supply chain, and manufactur­ing. Additionally, automation of data ex­change is a lower priority for less-frequent activities like Tech Transfer.

Another significant takeaway from BioPhorum’s report is mismatched expec­tations between sponsors and contract or­ganizations due to differences in maturity levels. The biggest mismatch occurs when large BioPharma sponsors desire real-time data from the contractor, whereas the con­tractor’s standard processes only provide batch records in PDF or Excel files for weekly or monthly submissions. Con­versely, a biotechnology start-up with lim­ited IT capability might be collaborating with an experienced and digitalized con­tract organization.

In both cases, there will be a signifi­cant difference between the visibility of and sophistication of process-related data and the ability to reap benefits from data shar­ing. This disconnect can be exacerbated by the lack of clarity in the contractual frame­work to convey the sponsor’s expectations.

When data sharing is left as a “detail to be worked out later,” subsequent unap­proved requests from the sponsor, which may seem trivial to a more digitally mature organization, may require partners to re­vamp their entire process.

Frequently, manual extraction still plays an important part in data exchange. Organizations working with multiple sites, partners, and/or sponsors often use data lakes to aggregate process data. Instead of allowing external parties to directly ac­cess this data or the original data sources, which would compromise the confidential­ity of other partners or sponsors, they may have to manually extract a modified ver­sion of sponsor-specific data.

However, through this process, infor­mation exchange is predisposed to cyber­security risks, human errors, and potential delays. Furthermore, data misinterpreta­tion can happen due to the variances in terminologies used by multiple parties. Ul­timately, this could jeopardize data in­tegrity and compliance, affecting the product release timelines and delivery to patients.

Sponsors and partner manufacturers have attempted to remediate this chal­lenge by implementing point solutions, but eventually, these solutions failed due to the complexity of the data shared, challenges in managing, and syncing changes on both sides and lack of scalability. Such point solutions struggle to support a range of digital platforms and siloed manufac­turing and lab applications across partner networks.

SHOULD ORGANIZATIONS STRIVE FOR HIGHER DIGITAL MATURITY LEVELS?

Ideally, all organizations could up­grade their manufacturing systems or de­ploy novel digital integration tools for a nearly homogenous digital maturity level. This solution is not realistic as it would force organizations to delay deploying ca­pacity while upgrading a facility and incur capital expenditures beyond the value of many projects.

In particular, achieving a digital ma­turity of level 4 (auto-ingest) and beyond can be costly both in terms of time and capital, especially when both the sponsor and partner must be at the same level of maturity.

Costs aside, an industry-wide up­grade is not necessarily constructive in every situation. Paper batch records and manual review may suffice for infrequent or small batch manufacturing to support early clinical trials. On the sponsor side, new companies with new molecules may receive little benefit from higher levels of automation until their products are pro­duced at large scales. Therefore, aspiring for the most advanced digital maturity level across the entire biopharmaceutical industry could be counterproductive.

The better alternative is to standardize data sharing in a way that accounts for sponsors and partners with varying digital maturity levels.

The first step to standardizing digi­tized data exchange between partnering organizations is for both parties to align expectations and requirements. The con­tract agreement must clearly define the scope of data to be included and excluded as well as the expected timeline for data exchange. This may also need to take into account regional differences in data gov­ernance as determined by the geographic locations of the parties involved and the intended markets. The GxP compliance re­quirements must also be established as well as the plan of action in case of revi­sions to previously shared data.

Addressing the fundamental ques­tions of data management will help ensure the partner organization – whether it is a CDMO for the development of the thera­peutic or a CMO for its mass production – will provide the optimum solution for the project proposed by the sponsor while at­tenuating the process interruptions caused by miscommunication.

DIGITAL INTEGRATION: AN EFFECTIVE STRATEGY TO BRIDGE THE GAPS BETWEEN VARIED DIGITAL MATURITY LEVELS

For partnerships between organiza­tions with varying maturity levels, shared digital solutions can provide tremendous value.

Cloud-native software that can offer seamless data integration and sharing across the manufacturing network is emerging as a promising solution. Such platforms can enable the partners to log in and enter, upload, or query data within a shared process context, while allowing the receiving end to pull the data directly when needed, maximizing efficiency and transparency.

Cloud-based software can satisfy sev­eral principles of digital integration. Data contextualization is a key strength that en­ables coherent capture of all critical processes and quality data with higher data integrity. Thus, the data is immedi­ately rendered meaningful, scientifically relevant, and interpretable for both the sponsors and manufacturing partners.

Furthermore, contextualized data can easily be converted into analytical output compliant with FDA regulatory require­ments, such as Continued Process Verifi­cation (CPV) and Annual Product Quality Review (APQR). The ability to track data entry and approvals and add electronic signatures supports data GMP principles and regulatory compliance with 21 CFR Part 11 and Annex 11.

Accommodating varied levels of digi­tal maturity, cloud-based digital integra­tion and data management software can establish streamlined communication be­tween sponsors and their partners, which helps build trust between the organizations and is one of the key ingredients of an agile manufacturing network.

Designed to give visibility to both par­ties, these systems enable them to jointly optimize manufacturing processes. Im­proved data digitalization capability also reduces the need for manual activities and removes human errors. In addition, spon­sors and manufacturing partners can view process data cooperatively for a systemic evaluation of performance against the quality assurance targets. This allows early prediction of manufacturing pain points and timely intervention to mitigate them, ultimately reducing batch failure and re­leasing product for patients faster.

Perhaps one of the most noteworthy premises of cloud-based digital integration is its flexibility to accommodate partners regardless of their digital maturity level and data management strategies. CMOs can continue to use their streamlined processes and systems to remain efficient in their manufacturing while replacing manual data file generation, review, and delivery with a focused, purpose-built tool to simplify and speed sharing with their customers.

For the sponsor, applications focused on process context can automatically map the data structures of multiple partners along with internal data to a standard, user-friendly context that can be inter­preted by all authorized parties and imple­mented with minimal training.

Cloud-based data contextualization and management tools will create new possibilities in the biopharmaceutical in­dustry by minimizing hindrances caused by complicated data exchange require­ments and facilitating collaboration. Rather than imposing the most advanced digital data-sharing systems on their part­ners, sponsors must encourage active par­ticipation in standardized digital integration. Meanwhile, contract organi­zations can maintain efficient operations and meet their customers’ data needs by supporting standardized digital integra­tion.

The collective efforts from drug spon­sors and their various partners will broaden the limits of the pharmaceutical industry and make it adaptable to the emergence of new therapies and health crises requiring the rapid deployment of agile manufacturing networks to efficiently manufacture novel therapeutics and vac­cines.

David Brick is Senior Director of Data Services at IDBS and has more than 30 years of experience in consulting, project management, data management, and data warehousing for reporting and analytics applications. He has spent more than 20 of these years focused on pharmaceutical and biotech manufacturing and process development. As part of Skyland Analytics and IDBS, he has been responsible for data management, connectivity, and sharing aspects of PIMS as well as technical implementation project delivery. Prior to joining Skyland Analytics, he served as Director, Professional Services for Dassault Systèmes BIOVIA (and its predecessors Accelrys and Aegis Analytical) where he had responsibility for all implementation activities for Discoverant, the world’s leading informatics software for Life Science manufacturers, and for the Nexus data access and aggregation components of the product. His clients included more than 50 process development and manufacturing facilities worldwide. He earned a BS in Applied Mathematics with University Honors and a MS in Statistics from Carnegie Mellon.

Harlan Knapp is Head of Digital Manufacturing Solutions at IDBS and drives digital solutions for tech transfer and manufacturing control strategies to help accelerate approvals of therapeutics. He has spent the past 20 years in the biopharmaceutical industry in a variety of roles, including R&D, manufacturing, product development, management, sales, and C-Suite business development. He earned his Master of Science in Biology from Eastern Washington University.