QBD 2.0: An Integrated Five-Pillar-Based Conceptual Predictive Lifecycle Framework for Next-Generation Pharmaceutical Development

Over the last 20 years, the pharmaceutical industry has been moving away from trial-and-error development and toward the science- and risk-based paradigms set out in the International Council for Harmonisation (ICH) Quality Guidelines Q8–Q11.(1–4) In this Quality by Design (QbD) framework, product and process understanding is attained by systematically correlating Critical Material Attributes (CMAs) and Critical Process Parameters (CPPs) with Critical Quality Attributes (CQAs), thus facilitating robust process design, control strategies, and lifecycle management of pharmaceutical products (5,6). Systematic reviews also show that QbD is widely accepted in theory, but in practice it is still not used consistently and is often limited by problems with organizations, technology, and regulations (5,6).

Even though QbD clearly makes processes more stable and less variable, there are some problems that make it hard to use its ideas in everyday manufacturing. Here are some of the issues: broken data architectures, limited end-to-end connectivity between development and commercial manufacturing, and mostly reactive instead of proactive use of Process Analytical Technology (PAT) (5,6). Traditional QbD methodologies often do not expand the design space to include downstream clinical performance metrics. This makes the feedback loop between real-world therapeutic outcomes and upstream decisions about formulation or process less strong (5,6). These kinds of limits show how important it is to have a system that brings together development, manufacturing execution, and patient-centered performance.

Digital and cyber-physical technologies have also come with the rise of Industry 4.0 and its adaptation to the pharmaceutical industry, Pharma 4.0. These technologies make the pharmaceutical value chain more open, responsive, and long-lasting. Some examples are the Internet of Things (IoT), cyber-physical production systems, advanced analytics, and smart automation (7). Digital twins have become more popular in this context as constantly changing, virtual representations of processes, equipment, or even whole buildings. This allows for simultaneous monitoring, diagnosing, and enhancing of manufacturing operations (8,9). According to the literature, digital twins can effectively combine mechanistic and data-driven models, link Process Analytical Technology (PAT) with multiscale process modelling, and enable scenario analysis throughout the product lifecycle (8–10). Complementary analyses of the application of artificial intelligence (AI) in advanced pharmaceutical manufacturing settings demonstrate how machine learning, optimization algorithms, and autonomous control systems can transform digital twins from mere simulators into intelligent agents that assist in decision-making or even execute decisions (10,11).

These converging advancements support the groundwork for the evolution of traditional Quality by Design (QbD) into an Integrated Predictive Lifecycle (IPL) framework, referred to as “QbD 2.0.” The main ideas behind this new framework are structured risk assessment and design-space definition (5,6). Digital twins, AI, IoT-enabled infrastructures, and autonomous analytics are some of the Industry 4.0 enablers that are also included (7–11). It can help with processes that improve themselves, testing releases in real time, and making sure that manufacturing performance and patient outcomes are more in line with each other. The aim of this review is to rigorously evaluate and synthesize the nascent paradigm of QbD 2.0/IPL, highlighting its theoretical foundations, enabling technologies, and practical ramifications for the creation of pharmaceuticals that are predictive, sustainable, and centered on patient needs.

The Figure 1. outlines the foundational elements of the Quality by Design (QbD) framework—Critical Material Attributes (CMAs), Critical Process Parameters (CPPs), Critical Quality Attributes (CQAs), control strategy, and lifecycle management—alongside the major implementation limitations, including data silos, reactive PAT usage, and weak clinical linkage. Industry 4.0 enablers such as IoT, cyber-physical systems, advanced analytics, and the QbD 2.0/IPL framework are shown as solutions to address these gaps.

The framework proposed in this review is presented as a conceptual synthesis of existing pharmaceutics and regulatory literature, intended to support lifecycle-oriented quality thinking rather than introduce a new regulatory classification.

Figure 1. Current QbD Elements, Key Limitations, and Industry 4.0 Enablers

PILLAR 1 – AI-POWERED PREDICTIVE PROCESS CONTROL

The pharmaceutical industry is transitioning from conventional observational process control to data-driven prediction methodologies that integrate artificial intelligence (AI) inside the Quality by Design (QbD) framework. In this context, Process Analytical Technology (PAT) is evolving from a passive monitoring instrument to a prescriptive control system that can autonomously adapt and make real-time decisions. The use of AI transforms PAT into a proactive component that not only monitors but also forecasts and dynamically modifies process parameters to sustain product quality within design specifications (12,13).

The usage of digital twins, or virtual copies that mimic physical processes in silico, is one of the most encouraging developments that may facilitate this change. In pharmaceutical development and continuous production, digital twins enable real-time duplication of unit activities, which helps with process scale-up, failure prediction, and early anomaly detection. This, in turn, reduces experimental costs and risks (9). Their comprehensive approach to data integration, predictive control, and process optimization is invaluable throughout the product lifecycle, which begins with drug development and continues through commercial-scale continuous production (9).

Reinforcement Learning (RL) emerged as an effective artificial intelligence technique for adaptive and self-governing process control. Algorithms for RL determining optimal control policies by interacting with real-world situations, and their performance is continuously improved through feedback. Recent advances, such as Deep Reinforcement Learning (DRL) and Control-Informed Reinforcement Learning (CIRL), improve the understanding and safety of AI-driven recommendations in chemical and pharmaceutical process control (12,13). DRL helps the system handle complicated and constantly changing pharmaceutical processes by learning from many possible situations. CIRL adds basic scientific principles into this learning so the system’s actions always make physical sense (12).

The United States Food and Drug Administration (FDA) has stressed the necessity of verified AI systems and their incorporation into frameworks for continuous improvement and lifecycle management, as defined in ICH Q12(14). This regulatory reinforcement shows a greater acknowledgment of AI's role in guaranteeing product quality, consistency, and compliance while also assisting the industry's transition to digital and continuous manufacturing paradigms (15). The Figure 2. illustrates the progression of Process Analytical Technology (PAT) from passive monitoring to proactive prediction and prescriptive AI-driven control. The central panel depicts the interaction between the physical process and its digital twin through real-time data exchange. The right panel shows the reinforcement learning (RL/DRL/CIRL) loop consisting of process data acquisition, agent-based decision-making, and adaptive control actions.

Therefore, AI-powered predictive process control, which is based on PAT, digital twins, and reinforcement learning, is a big change in the way drugs are made. It makes operations more efficient, products more reliable, and regulators more confident.

Figure 2. Evolution of PAT, Digital Twin Architecture, and Reinforcement Learning Control Loop

PILLAR 2 – CLOSED-LOOP LIFECYCLE INTEGRATION

A data design based on Industry 4.0 connects research and development, manufacturing, regulatory oversight, and post-market surveillance in a digital continuum called Closed-Loop Lifecycle Integration. Setting up an integrated data infrastructure makes the pharmaceutical value chain more open, makes it easier to share information, and lets people make decisions in real time. This is in line with the digital transformation principles of sustainable pharmaceutical operations (7).

One thing that makes this approach different is that it has a constant feedback loop that links how well a product works in the real world to its manufacturing and quality functions higher up the chain. Companies can link safety problems that happened after a product was sold to specific production batches thanks to strong lot-level tracking in global safety reporting systems. This helps find the root cause, lower risks, and always make things better (16).

Artificial intelligence (AI) and natural language processing (NLP) make this closed-loop system even better by automatically pulling out new safety trends and adverse drug events from large, unstructured datasets. Many people have discovered that supervised learning models powered by NLP are very useful for spotting signs of bad events in clinical narratives. This makes pharmacovigilance much more responsive and changes drug safety tracking from a reactive to a predictive process (17).

Real-World Evidence (RWE) makes this loop more regulatory. It is becoming more important in shaping the decision-making frameworks of major global agencies. Using real-world evidence in lifecycle management helps keep evaluations of risks and benefits, regulatory assessments, and quality oversight up to date with the most recent clinical performance after the product is on the market. (18)

Lifecycle with a Closed Loop Integration makes a single digital system out of scientific, operational, and regulatory areas. In this system, real-world data is very important for making product design better all the time, making manufacturing processes more efficient, and, in the end, providing relevant outcomes(7,16–18). The Figure 3. presents a digitally integrated, closed-loop pharmaceutical lifecycle connecting research and development, manufacturing, quality assurance, regulatory oversight, pharmacovigilance, AI/NLP-based adverse event detection, and real-world evidence (RWE) generation. Bidirectional data flows enable traceability, continuous improvement, and incorporation of post-market insights into upstream design and regulatory decision-making.

Figure 3. Closed-Loop Pharmaceutical Lifecycle Integrating Manufacturing, Quality, Safety, and RWE

PILLAR 3 – QUANTITATIVE PATIENT-CENTRIC MODELLING

The third pillar of the new QbD 2.0 concept is Quantitative Patient-Centric Modelling, an unconventional technique that connects process and product knowledge to measurable patient outcomes known as Critical Clinical Performance Attributes (CCPAs). These CCPAs function as clinical counterparts of CQAs, representing measurable markers of therapeutic efficacy and safety that can be traced back to manufacturing and formulation characteristics (8,19).

Quantitative Systems Pharmacology (QSP) is a significant innovation that enables this integration by combining mechanistic modelling of drug-disease-patient interactions with molecular and system-level insights. QSP serves as a translational bridge between CQAs and in vivo performance, allowing developers to simulate how process-induced changes, such as protein glycosylation patterns, influence pharmacodynamics and clinical outcomes (20). For example, antibody Fc glycosylation heterogeneity has been recognized as a key quality attribute (CQA) with important implications for immune effector function, therapeutic efficacy, and safety profiles (21). This mechanistic mapping of CQAs to CCPAs is an important regulatory and scientific objective for improving predictive pharmacological design (20).

The integration of systems biology data with patient-reported outcomes (PROs) strengthens this patient-centered framework even more. The use of omics-derived biomarkers and real-world patient experience data improves the model's ability to capture inter-individual variability and illness heterogeneity. Wang et al. (2022) emphasize the necessity of incorporating PROs into electronic health records to establish a continuous feedback loop between patient experience and therapy design, thereby enabling adaptive, evidence-based optimization (22).

Simultaneously, the reverse-engineered design philosophy—beginning with the intended patient outcome and working backward to establish the ideal CQA—represents a paradigm shift in QbD application.  Pharmaceutical scientists can use digital twins to simulate process-product-patient interdependencies and iteratively improve Critical Process Parameters (CPPs) to obtain the desired CCPAs before conducting physical experiments (8).  Chen et al. (2020) define digital twins as a dynamic, data-driven virtual duplicate of manufacturing systems that allows for continuous learning, risk assessment, and real-time control across the product life cycle (8).

In conclusion, Pillar 3 integrates patient-centric outcome modelling, digital twin technology, and quantitative systems pharmacology into a unified paradigm that goes beyond traditional QbD limitations. This method achieves the ultimate objective of QbD, which is to design quality into products that are scientifically optimized for patient benefit, by mechanistically connecting molecular CQAs to clinical performance through CCPAs and incorporating patient feedback through PROs (8,19–22).

This Figure 4. shows the integration of Critical Quality Attributes (CQAs), Critical Clinical Performance Attributes (CCPAs), and Patient-Reported Outcomes (PROs) via a Quantitative Systems Pharmacology (QSP) framework. QSP acts as a translational bridge, informing the digital twin, which generates digital outcomes to support predictive assessment of clinical performance and patient-centered design.

Figure 4. Quantitative Patient Centric Modelling Linking CQAs, CCPAs, PROs and Digital Twin Outputs

PILLAR 4 – RESILIENT-BY-DESIGN SUPPLY CHAIN

A resilient pharmaceutical supply chain combines digital technology to improve visibility, traceability, and responsiveness throughout the distribution network.  The Internet of Things (IoT) allows for real-time monitoring of important characteristics, including temperature, humidity, and logistics conditions, maintaining product integrity throughout the transportation and storage phases (23).  The use of IoT-based cold chain systems in container ports has shown the ability to identify deviations instantly, reducing the risk of temperature excursions that could compromise drug quality (23).

A key tool for predicting possible problems in global supply chains is predictive analytics that is backed by machine learning (ML) algorithms. ML models can predict risks and make the best decisions for proactive actions by processing multidimensional datasets that include weather patterns, port congestion, and supplier reliability. This intelligence based on data makes pharmacy networks more flexible and quicker to change, reducing downtime and making sure products are delivered on time (24,25).

Blockchain technology improves the supply chain by using an immutable ledger system to ensure complete transparency and security against counterfeiting operations. The decentralized nature of blockchain ensures that each transaction, from production to distribution, is verifiable and tamper-proof, hence increasing stakeholder trust (25,26). Furthermore, the use of blockchain-based smart contracts automates compliance verification and payment processing based on IoT-generated quality signals (27). These self-executing digital contracts eliminate the need for manual monitoring, speed up logistics, and ensure consistent compliance with regulatory standards (27).

Pillar 4 emphasizes the integration of four key components: IoT monitoring, blockchain technology, a system of smart contracts, and predictive analytics, all of which enhance the pharmaceutical supply chain by promoting stronger management practices, improving quality, and supporting a resilient-by-design ideology.

This addition seems to improve the management system from being a reactive management system to a more data-driven system, which is proactive in nature. Thus, this shift may improve the supply chain's operational efficiency and reliability and, in the end, improve patient safety.

The Figure 5. presents an IoT-driven supply chain architecture where sensor-generated environmental data feed into cloud analytics and machine learning models for predictive risk assessment. A blockchain ledger ensures secure traceability and anti-counterfeiting, while the downstream logistics and distribution system supports efficient, compliant product movement.

Figure 5. IoT-Enabled Supply Chain with Cloud Analytics, Blockchain Traceability, and Smart Logistics

PILLAR 5 – SUSTAINABILITY INTEGRATION

Sustainability integration in the pharmaceutical Quality by Design (QbD 2.0) system highlights the importance of making products and processes more environmentally friendly. The environmental performance of a process is shown by its green critical quality attributes (gCQAs), which are measurable eco-metrics like Process Mass Intensity (PMI), energy efficiency, and liquid safety. It has been debated for a long time how to choose the right green metrics. However, Constable et al. pointed out that the most useful sustainability indicators are those that integrate material efficiency with real-world use in synthetic route optimization (28).

PMI has become one of the most important ways to measure how well pharmaceutical manufacturing uses materials and how it affects the environment. Jimenez-Gonzalez et al. showed that it is widely used in the industry as a standard for judging the sustainability of a process because it is easy to use and can lead to measurable changes in reducing waste and maximizing resource utilization (29). This alignment of PMI with QbD principles makes sure that environmental efficiency isn't just an afterthought, but is instead a design variable that affects how stable and scalable the process is.

The concept of a "Green Chemistry Continuum" takes this philosophy even further by supporting sustainability from the very beginning of route design all the way through large-scale API production. Koenig et al. said that this continuum is necessary to keep a strong and environmentally friendly supply chain that balances green performance with business stability (30). Roschangar et al. also came up with the Green Aspiration Level™ (GAL) method to measure the difference between current and desired green performance. This gives us a structured way to decide which greener process innovations to focus on first (31).

Lifecycle Assessment (LCA) serves as a complementary tool, integrating environmental impact data within process simulations to guide the selection of low-impact synthetic routes. Rose et al. emphasized the importance of LCA-guided green-by-design strategies for small-molecule APIs, enabling proactive decision-making in early R&D phases (32). Nevertheless, as Satta et al. noted, the practical application of LCA in pharmaceuticals still faces challenges related to data availability, system boundary definitions, and methodological standardization (33).

Embedding sustainability into pharmaceutical R&D is not only a scientific imperative but also a cultural transformation. Sneddon underscored that achieving such integration requires institutional commitment, interdepartmental collaboration, and incentive structures that reward sustainable innovation (34). Ultimately, aligning QbD 2.0 with sustainability principles leads to synergistic outcomes—reducing environmental burden while enhancing process efficiency, cost-effectiveness, and regulatory goodwill.

The Figure 6. illustrates the green chemistry continuum, beginning with eco-centric process design using gCQAs and sustainability metrics such as PMI and solvent safety. API production is followed by lifecycle assessment (LCA), Green Aspiration Level (GAL) benchmarking, and institutional commitment—enabling improved process efficiency, scalability, cost savings, and environmentally responsible pharmaceutical manufacturing.

Figure 6. Green Chemistry Continuum for Sustainable API Development and Lifecycle Assessment

BARRIERS TO IMPLEMENTATION AND MITIGATION STRATEGIES

The adoption of an Integrated Pharma Line (IPL) framework requires alignment of digital, organisational, and regulatory capabilities. Evidence from Industry-4.0-focused analyses consistently demonstrates that interoperability limitations, workforce capability gaps, cultural resistance, and evolving regulatory expectations hinder modernisation efforts in pharmaceutical manufacturing (35–37). Economic modelling further reveals that transitions to advanced or continuous systems are highly sensitive to capital investment assumptions, which constrains large-scale adoption (38). Table 1. summarises these barriers, the evidence supporting their impact, and structured mitigation approaches.

Table 1. Evidence-Aligned Barrier and Mitigation Matrix for IPL Framework Adoption