This blog was co-authored by Dominic Furniss and Julie Avery.
1. The Crisis Is Here – The Regulatory Gap Costing Pharma Billions and Impacting Patients
Data from the OECD shows that drug shortages have been rising since 2019 and are accelerating.[1] Many are due to manufacturing issues and product quality deficiencies.
The industry numbers tell a stark story.
In fiscal year 2024, pharmaceutical manufacturers voluntarily initiated 260 recall events, up 15% on FY 2023.[2] A total of 421 products were subject to recall. Fifty per cent of these recalls were due to failure in current Good Manufacturing Practice (cGMP).[2] The data also shows that some companies were responsible for many recalls, showing a systemic breakdown in their system. For example, in 2023 one company was responsible for 67 recalls.[3]
For the same year, the FDA issued 105 warning letters to human drug manufacturing sites based on inspection and record requests and product testing. This was the highest number of human drug quality-related warning letters issued in the past five years.[2]
Between 2012 and 2023 there have been approximately 330 recall events per year for products regulated by FDA.[4] However the actual number of products recalled is much larger with individual recall events identifying the need for multiple product recalls due to systemic issues. 19% of these recalls attributed to labelling and packaging errors – fundamentally human performance issues.[4]
Top violations appear repeatedly in inspection findings (483s) including warning letters:
- Inadequate quality control procedures (laboratories)
- Incomplete Production Record Reviews (21 CFR 211.192)
- Failure to thoroughly investigate batch discrepancies and deviations
- Lack of comprehensive documentation for production processes
- Insufficient Written Procedures and Deviation Management (21 CFR 211.100)
- Absence of detailed SOPs for critical operations
- Inadequate handling and documentation of process deviations
- Data integrity lapses (still in the top 10 despite 15 years of industry focus)
- Inadequate Control of Microbiological Contamination (21 CFR 211.113)
- Deficiencies in aseptic practices, gowning procedures, and environmental monitoring
- Instances of poor aseptic techniques and inadequate media fill validations
But the numbers don’t capture the full picture.
Behind these statistics lie pharmaceutical manufacturers struggling with systemic quality problems – not isolated incidents, but persistent patterns they cannot seem to break. Some are caught in exhausting reactive cycles: deviation occurs, investigation follows, corrective and preventive action is implemented but not a system solution, and then the pattern repeats with a different task or different operator. Others have reached a troubling realisation: when “human error” becomes the most common root cause in your deviation database, your investigation process has stopped generating insights. It focuses on the problem rather than the underlying causes, often seeing the person as the problem.
For emerging sectors, the stakes are even higher. Cell and gene therapy start-ups and manufacturers face operational complexities that can threaten organisational viability – especially given the high cost of each dose/therapy. The contract manufacturing services market for these therapies alone is projected to grow from USD 7.52 billion to USD 88.84 billion by 2034 (28.1% CAGR),[5] reflecting surging demand for outsourced production as companies struggle with manufacturing scale and complexity. Getting human performance right isn’t just about compliance – it’s existential.
The common thread? These aren’t random errors occurring despite well-designed systems. They are predictable patterns that emerge when human performance is treated as an afterthought rather than systematically designed into quality operations.
The crisis isn’t coming. It’s already here.
2. The Technological Mirage: Why AI Isn’t The Answer
When faced with persistent human performance problems, many pharmaceutical organisations naturally look to technology for solutions. The promise is compelling: artificial intelligence that can predict deviations before they occur, machine learning algorithms that identify quality trends invisible to human analysis, digital twins that simulate processes to reveal vulnerabilities, and real-time monitoring systems that alert supervisors to developing problems.
Yet despite substantial investments in these technologies, the fundamental patterns persist.
Why AI Alone Fails
First, AI cannot design human-centred work. Predictive analytics might tell you that an error is likely to occur – perhaps even when and where – but it cannot tell you why or how the work is fundamentally designed to create that error. It cannot redesign workspace layouts that create confusion, rewrite procedures that don’t reflect actual work practices, or restructure competence systems that fail to prepare operators for critical decisions.
Second, there is the nuance problem. Work-as-imagined (as described in SOPs) and work-as-done are rarely the same thing. Operators develop informal practices – workarounds, double-checks, memory aids – that help them succeed despite system inadequacies. These resilience strategies are often tacit, undocumented, and invisible to formal procedures. AI systems trained on documented procedures will miss these entirely, potentially optimising against the very adaptations that currently prevent failures. AI cannot therefore understand the risks people manage, often successfully, in daily work because it is not written down anywhere.
Third, technology without understanding generates expensive noise. More dashboards don’t create insight. They are often lag metrics – status measures about yesterday, not actual key performance indicators that give insight into where you can drive effective change and efficiencies. More alarms (red signals) create alarm fatigue. More data doesn’t mean better understanding. Without systematic analysis of how work actually happens, how things go right, and why failures occur, technology investments simply help organisations fail faster – albeit with better data visualisation.
Fourth, AI learning from records of investigations and solutions (Corrective and Preventive Actions, or CAPAs) is very limited. AI can only read the words input into the Quality Management System (QMS) from process deviations and CAPAs, which are often documented inadequately (see FDA top inspection findings). Also, QMS systems require you to identify cause (not causes) from a drop-down menu of categorised headings. These options are often too restrictive, only telling a small part of the story.
What AI Can’t Replace
There is no algorithm that can sit with operators to understand the texture of their actual work. No machine learning model that can facilitate the consensus-building required to create genuine ownership of improvements. No digital system that can identify the subtle environmental factors – poor lighting, confusing labelling, inadequate workspace, ambiguous instructions, the need to multi-task just to achieve the task in a timely manner and manage changes – that increase error likelihood.
This isn’t an argument against technology. AI has tremendous power to enhance well-designed systems: monitoring for deviations from understood norms, identifying patterns across datasets too large for human analysis, providing decision support at critical moments. But that power only manifests when applied to systems where human factors have already been systematically addressed.
The pharmaceutical industry needs to fix the human factors first, then apply AI to enhance what already works.
3. What High-Hazard Industries Know (That Pharma Doesn’t Yet)
Pharmaceutical manufacturing is not the first high-hazard industry to confront the challenge of human performance in complex technical systems. Others learnt through tragedy, then systematised their response.
Lessons Written in Disaster
Process safety industries, nuclear power, and aviation don’t treat human performance as a behavioural problem requiring motivational posters and cultural campaigns. They treat it as an engineering problem requiring systematic frameworks.
The structure looks remarkably similar across sectors:
- Proactive risk assessment before incidents occur
- Systematic analysis of how work actually happens (not how procedures say it should happen)
- Integration across safety, quality, and operational functions
- Measurable success criteria that can be inspected and benchmarked
- Continuous improvement cycles built into the management system
The Pharmaceutical Gap
Pharmaceutical manufacturing has extensive regulatory human factors guidance – but nearly all of it focuses on medical device user interfaces and patient-facing combination products.[6] This makes sense: ensuring patients can correctly use insulin pens or inhalers is critical. The FDA’s human factors guidance in this domain is mature and comprehensive.
What doesn’t exist is equivalent regulatory systematic guidance for human performance in manufacturing operations themselves. The Quality System Regulations (21 CFR 210 and 211) as described in ICH Q-10[7] require a fit-for-purpose Quality Management system with adequate procedures, appropriate training, and investigation of deviations and CAPAs (Corrective & Preventive actions) – but provide no structured framework for proactively managing human performance in quality-critical tasks.
There is an expectation set around system thinking and human performance in European Commission regulations. For companies manufacturing and marketing drug products in Europe regulations that are covered in EudraLex. Human performance gets a mention in Vol. 4 Chapter. 1.4: “Where human error is suspected or identified as the cause, this should be justified having taken care to ensure that process, procedural or system-based errors or problems have not been overlooked, if present.”[10] But again, it does not provide a structured framework for proactively managing human performance in quality-critical tasks.
The gap isn’t knowledge. Human factors science is well-established. Methodologies for systematic task analysis, failure mode identification, and performance influencing factor assessment have been proven across multiple high-hazard domains.
The gap is structure. Pharmaceutical quality needs its own Human Factors Delivery Guide as exists in other high-hazard industries.
4. The COMAH Model: A Proven Template
The Control of Major Accident Hazards (COMAH) regulations provide the most relevant template for pharmaceutical quality, not because pharmaceutical product quality and cGMP compliance is identical to process safety, but because the regulatory challenge is analogous: how to systematically prevent low-probability, high-consequence events driven by complex interactions between human performance and technical systems. The framework is defined in the Human Factors Delivery Guide for COMAH sites.
Background
The HF Delivery Guide emerged from the UK’s response to major industrial disasters, most notably Piper Alpha (1988). The realisation was disturbing: major accidents weren’t simply the result of individual errors or isolated equipment failures. They emerged from systemic vulnerabilities in how human performance was managed across the entire operation. Cultural factors, organisational pressures, design inadequacies, and procedural gaps interacted in ways that created “accident pathways” invisible to traditional risk assessments.
The framework has been developed and refined since approximately 2010, with the most recent update to the Human Factors Delivery Guide released in December 2023.[8] Its adoption extends beyond the UK – Singapore’s Ministry of Manpower (MOM) has implemented similar requirements, Brazil is developing comparable frameworks, and multinational companies often apply the COMAH HF Delivery Guide principles across their global operations.
Three Key Characteristics of the COMAH Approach
1. A Structured Topic Framework
Rather than general exhortations about human factors, COMAH HF Delivery Guide provides a comprehensive six-topic structure ensuring all critical aspects of human performance are systematically addressed:
- Managing Human Performance
- Human Factors in Process Design
- Critical Communications
- Procedures Design and Management
- Critical Competence Mapping
- Managing Organisational Factors
2. Auditable Success Criteria
For each topic, specific success criteria define what “good” looks like. These aren’t aspirational goals – they’re measurable standards that regulators use during inspections. Sites can benchmark their performance, identify gaps, and demonstrate continuous improvement. This transforms human factors from abstract concept into concrete management requirement.
3. Integration of Psychology and Hierarchy of Control
COMAH HF Delivery Guide guidance explicitly combines psychological understanding of human performance (why errors occur, how people make decisions under pressure, what influences behaviour) with the hierarchy of control principles (eliminate hazards through design first, then engineer controls, with administrative controls and training as last resorts). This integration means human factors interventions target system design, not just individual behaviour modification.
The Translation Opportunity
The COMAH HF Delivery Guide focuses on preventing major accident hazards – fires, explosions, toxic releases. A pharmaceutical equivalent would focus on quality-critical tasks and patient safety. The principles remain the same; the application context differs.
We don’t need to reinvent the wheel. We need to adapt a proven framework for pharmaceutical quality’s specific challenges.
5. A Seven-Topic Framework for Pharmaceutical Quality: A Proposal for Consideration
Building on the COMAH HF Delivery Guide structure but adapted for pharmaceutical manufacturing’s unique context, we propose a seven-topic framework. The addition of a seventh topic – Leadership and Quality Culture – reflects pharmaceutical manufacturing’s particular challenge: quality isn’t just about preventing catastrophic single events (as in process safety), but about sustained excellence across every batch, every day, with patient safety dependent on consistent performance. We also added this to show that the content of the topics could change as well as the topics themselves – perhaps we could end up with seven plus or minus two topics.
Topic 1: Managing Human Performance in Quality-Critical Tasks
Proactive Assessment:
Systematic identification and analysis of quality-critical tasks before deviations occur. This means applying structured methodologies (such as Safety Critical Task Analysis) to understand how work actually happens, where vulnerabilities exist, and what factors influence likelihood of failure. The goal is to identify quality-critical tasks, prioritise them based on consequence and vulnerability, and conduct systematic human reliability assessments. These would reflect on Quality Critical Parameters and product licence details.
Reactive Investigation:
When deviations do occur, investigation should move beyond identifying “human error” as a root cause to understanding why and how the error occurred and what system factors made it likely. Human Reliability Assessment methodologies exist (such as Task Analysis Based Incident Evaluation, or TABIE) that provide systematic approaches to incident investigation, ensuring that learning generates actionable insights about system improvements, not just individual remediation.
Success Criteria:
- Quality-critical tasks identified and prioritised with clear rationale
- Evidence of systematic human reliability assessment methodology
- Incident investigations reveal system factors and generate design-level improvements
- Integration between proactive assessment and reactive learning
Topic 2: Human Factors in Process and Equipment Design
It is important to apply this thinking to legacy systems and equipment, but it is especially important in the start-up phase, where new manufacturing processes are being developed, scaled, and implemented, and in the design of new production lines and factories.
Design Philosophy:
The hierarchy of control applies: eliminate hazards through design first, then engineer controls, then administrative controls, with personal protective equipment and training as last resorts. For quality, this means: design processes so the right action is intuitive and the wrong action is difficult or impossible.
Key Applications:
- Human-Machine Interface design for quality operations (clear displays, intuitive controls, effective alarms)
- Equipment design that prevents misidentification (colour coding, physical incompatibility, forcing functions, feedback, cues)
- Workspace layout that reduces error opportunities (logical flow, adequate space, good lighting)
- Thoughtful automation allocation: understanding which tasks benefit from automation (repetitive, precisely controlled, requiring continuous monitoring) versus which require human judgement (pattern recognition, adaptation to unusual circumstances, ethical decisions). This includes designing automation that supports rather than undermines human understanding of the process, avoiding “out of the loop” problems where operators lose situation awareness.
Success Criteria:
- Human factors integrated during design and process validation phase with documented assessments
- Evidence of error-proofing in equipment specifications
- Work environment designed to support quality task performance
- Automation decisions based on human factors analysis, not just cost
- Automated systems designed to maintain operator engagement and process understanding
Topic 3: Quality-Critical Communications
Key Systems:
- Shift handover protocols ensuring quality status, ongoing deviations, and critical monitoring requirements transfer effectively
- Batch record transitions with clear ownership and quality accountability
- Deviation and CAPA communication across shifts and functional boundaries
- Cross-functional coordination (QA, Production, QC, Maintenance, Engineering) with defined information flows
Success Criteria:
- Formal communication protocols for quality-critical information with evidence of use
- Shift handover processes demonstrably effective at transferring critical information
- Quality information reaches decision-makers with appropriate urgency
- Communication breakdowns identified as system issues, not individual failures
Topic 4: Design and Management of Risk-Informed Procedures
Procedure Philosophy:
Procedures should reflect work-as-done, not just work-as-imagined. They should provide risk information at point-of-use, be proportional to task complexity and consequence, and be genuinely usable during actual work (not just filed as reference documents).
Key Principles:
- Operator involvement in procedure development ensures procedures match reality
- Critical steps clearly identified with no interpretation required, with appropriate warnings and cautions for safety and quality
- Visual aids and decision support embedded where they add value, with full use of available technology (e.g. videos, tablets, QR code links to specific tasks, context-driven)
- Regular review cycles ensure the group of process procedures remains current with actual practice (no siloed review of individual documents)
- Format and detail level proportional to task risk, complexity, and operator experience
Success Criteria:
- Procedures based on systematic task analysis, not just subject matter expert recollection
- Evidence of operator input and ownership (they recognise their work in the procedure)
- Critical information appropriately emphasised and positioned
- Procedures demonstrably usable in the actual work environment
Topic 5: Critical Competency Management
Moving Beyond Apprenticeship Models:
Traditional “see one, do one, teach one” approaches create unacceptable variability. One operator’s understanding becomes another’s training, perpetuating both good practices and hidden vulnerabilities. Knowledge remains tacit, assessment remains informal, and competence verification depends on who happened to provide the training.
Task-Based Competency Mapping:
Systematic task analysis reveals what knowledge and skills each quality-critical step actually requires. This enables competency mapping: explicit identification of what operators must know and be able to do for each stage of the process. Operators can be formally assessed and signed off against these specific requirements, ensuring consistent competence regardless of who provided training. Ongoing competency is supported by this map.
Key Elements:
- Knowledge requirements identified for each quality-critical task (what must operators understand?)
- Skill requirements defined (what must operators be able to do?)
- Experience events needed to embed the skill and knowledge
- Training effectiveness measured against task performance and competence standard, not just completion rates
- Formal sign-off against specific competency requirements and standards for critical steps
- Ongoing competence verification for highest-risk activities, not just initial qualification
Success Criteria:
- Competence standards explicitly linked to quality-critical tasks through documented task analysis
- Training content demonstrably aligned with actual work demands derived from systematic analysis
- Assessment methods validate ability to perform critical tasks, not just theoretical knowledge
- Formal competency mapping with sign-off requirements for critical operations
- Competence management system adapts as processes and tasks evolve
Topic 6: Managing Organisational Factors in Quality
Management of Change:
How do process changes, personnel changes, or organisational restructuring impact quality system performance? Success requires systematic assessment before changes are implemented, not reactive investigation after problems emerge.
Fatigue and Workload:
Shift patterns, production pressure, and resource constraints directly affect human performance. These aren’t just “nice to have” considerations – they’re quality risk factors requiring active management.
Staffing Levels:
Are there adequate qualified resources for quality-critical operations? Not just minimum staffing, but sufficient capacity to perform tasks thoroughly without cutting corners during high-demand periods?
Production Pressure:
The eternal tension between efficiency and thoroughness. Success requires transparency about these trade-offs and organisational commitment to quality when conflicts arise, including demonstrated querying of a successful batch – how did things go right?
Success Criteria:
- Systematic process for assessing change impacts on quality with documented decisions
- Workload and fatigue risks identified and actively managed
- Resource allocation decisions consider quality task demands, not just throughput. Do people have the capacity to complete the tasks safely and compliantly?
- Evidence that quality considerations influence production planning and scheduling
Topic 7: Leadership and Quality Culture
This seventh topic addresses pharmaceutical manufacturing’s unique challenge: sustaining quality excellence requires leadership behaviours and cultural norms that support continuous learning and improvement.
- Learning Teams:
Leaders regularly engaging with frontline operators to understand work-as-done, not issuing directives based on work-as-imagined. These aren’t “gemba walks” where leaders observe passively – they’re structured conversations where leaders learn about system constraints operators navigate daily.
- Pre- and Post-Job Reviews:
A critical element for pharmaceutical manufacturing. Leaders and teams discuss what is required before the task, anticipate potential challenges, and then review how things went afterwards to identify learning opportunities.
- Quality Coaching:
Supervisors asking better questions when problems arise. Instead of “Why didn’t you follow the procedure?” asking “What makes following the procedure difficult?” “What adaptations do you need to make?” “How did you know what to do?” This shift – from blame to understanding – generates different information and different solutions.
- Psychological Safety:
Creating an environment where errors can be discussed without fear of punishment to enable improvement. This isn’t about reducing individual accountability; it’s recognising that learning requires transparency, and transparency requires trust. People need resource, autonomy, and authority to improve their work if you want to hold them accountable.
Human and Organisational Performance (HOP) Principles:
Understanding error as a normal part of human performance, not a character flaw. Focusing on the system conditions that make errors more or less likely, not just individual actions. How context influences our behaviour. Error is a signal from the system.
Cultural Elements:
- Blame-free reporting that encourages learning from near-misses and deviations
- Frontline voice genuinely valued in quality improvement (not just formally solicited then ignored)
- Continuous learning mindset embedded in operations (not just compliance-driven)
- Quality excellence seen as everyone’s responsibility (not just Quality Assurance’s)
Success Criteria:
- Evidence of leadership participation in learning teams with operators, with demonstrated competence in human factors
- Operators report psychological safety for raising quality concerns
- Quality culture systematically measured (not just assumed)
- HOP principles visible in investigation approaches and improvement initiatives
- Incident reviews focus on system factors, with individual accountability reserved for rare wilful violations
This framework isn’t theoretical abstraction. Each topic addresses specific challenges visible in FDA warning letters, recall patterns, and the daily struggles of pharmaceutical quality organisations. It provides comprehensive coverage of human performance factors whilst remaining practical, scalable, and implementable.
6. The Path Forward: Who Does What
Creating a Human Factors Delivery Guide for pharmaceutical quality requires coordinated action across multiple stakeholders. No single entity can drive this transformation alone. Help might be needed from professional bodies, regulatory authorities, industry groups (e.g. BioPhorum Human Performance workstream , “… concentrates on embedding human performance practice and assessment into drug substance operations” [9], academic institutions, and consultancies who specialise in GxP and Human Reliability.
Two key stakeholders:
For Pharmaceutical Regulators
The Immediate Opportunity:
Regulatory authorities – particularly the FDA and EMA – have extensive human factors expertise focused on medical device user interfaces. That same expertise could be directed towards manufacturing operations. The precedent exists: FDA already has comprehensive human factors guidance for patient-facing products.[6] The framework requires adaptation, not invention.
Specific Actions:
- Consider developing human factors guidance specifically for manufacturing operations (not just device use)
- Examine COMAH structure as a potential regulatory model with proven effectiveness
- Establish success criteria for human performance in pharmaceutical quality systems
- Include human factors assessment in inspection protocols alongside traditional quality system review
The regulatory approach needn’t be immediately prescriptive. Initial guidance could be advisory, allowing industry to develop best practices that inform eventual requirements – much as COMAH evolved from voluntary adoption to regulatory requirement.
For Pharmaceutical Manufacturers
Don’t Wait for Regulation:
By the time regulatory requirements appear, leading organisations will have already built capability. They won’t be scrambling to comply – they’ll be helping write guidance based on their experience.
The Beachhead Approach for your organisation:
- Phase 1 – Prove Value:
Identify 3-5 highest-deviation tasks or processes causing greatest quality concern. Apply systematic human factors analysis to these tasks. Measure outcomes: deviation reduction, operator satisfaction, cost of quality improvements. This typically requires 3-6 months and generates clear evidence of value.
- Phase 2 – Build Capability:
Train internal champions in human factors methodologies or establish partnerships with specialists. Expand analysis to broader set of quality-critical tasks. Begin integrating human factors into design processes for new equipment and procedures.
- Phase 3 – Systematic Integration:
Embed human factors into quality management system. Establish success criteria aligned with proposed framework. Include human factors considerations in management review, new product introduction, change management, and investigation and CAPA processes.
Investment Perspective:
Proactive human factors investment reduces downstream costs: fewer deviations requiring investigation, fewer CAPAs consuming resources, fewer regulatory observations requiring response, fewer batch release delays, and fewer product recalls with associated costs and reputational damage. The business case isn’t difficult – it requires commitment to upfront investment rather than perpetual reactive expenditure. From experience, each deviation can cost between $10k-$15k in administrative costs alone.
Competitive Advantage:
First movers won’t wait for regulatory pressure. They’re building capability now, before their competitors notice the gap. When regulatory requirements eventually arrive, they’ll be helping write the standards, not struggling to meet them.
Conclusion
The pharmaceutical industry stands at a crossroads. FDA warning letters are rising. Recall events persist at 330 per year. Stock-outs driven mainly by delays in batch release due to manufacturing issues are the highest they’ve ever been, with the same human performance patterns recurring. Cell and gene therapy complexity is exploding. And yet, whilst other high-hazard industries systematised their response to human performance decades ago, pharmaceutical quality manufacturing continues without equivalent structure to manage the risks.
This isn’t a problem that technology alone can solve. AI, data analysis dashboards, machine learning algorithms, and digital twins are powerful – but only when applied to well-designed systems. First, we must address the human factors.
The seven-topic framework proposed here isn’t a prescription – it’s an invitation for consideration. It builds on proven principles from COMAH whilst recognising pharmaceutical quality’s unique challenges. Each topic addresses real FDA citations, actual recall patterns, and the daily frustrations of quality professionals struggling with “human error” as their most common root cause.
The question isn’t whether pharmaceutical quality needs systematic human factors management. The evidence is overwhelming. The question is: who will lead?
Will regulators develop guidance before the next major quality crisis forces their hand?
Will manufacturers build capability now, collaboratively shaping standards based on their experience rather than scrambling to meet requirements written in response to significant failure?
Will industry bodies facilitate the conversation, creating consensus before regulation makes it mandatory?
The framework exists. The need is urgent. The opportunity is now.
Feedback & Connection
Thoughts? Are you seeing these patterns in your organisation? What’s stopping us from adopting systematic approaches?
Contact us and connect as it would be great to hear from you if you had feedback or thoughts about progressing this project. Julie Avery and Dominic Furniss.
References
[1] Organisation for Economic Co-operation and Development (OECD). (2022). Shortages of Medicines in OECD Countries. OECD Publishing, Paris. Available at: https://www.oecd.org/health/shortages-of-medicines-in-oecd-countries-b5d9b14f-en.htm European Commission, Health Emergency Preparedness and Response Authority. Critical Medicines Alliance. Available at: https://health.ec.europa.eu/health-emergency-preparedness-and-response-hera/overview/critical-medicines-alliance_en
[2] U.S. Food and Drug Administration, Center for Drug Evaluation and Research (CDER). (2024). FY2024 Report on the State of Pharmaceutical Quality. Available at: https://www.fda.gov/media/185592/download
[3] U.S. Food and Drug Administration, Center for Drug Evaluation and Research (CDER). (2023). FY2024 Report on the State of Pharmaceutical Quality. Available at: Fiscal Year 2023 Report on the State of Pharmaceutical Quality
[4] Ghijs, S., Wynendaele, E., & De Spiegeleer, B. (2024). The continuing challenge of drug recalls: Insights from a ten-year FDA data analysis. Journal of Pharmaceutical and Biomedical Analysis, 249, 116349.
[5] Nova One Advisor. (2025). Cell and Gene Therapy CDMO Market Size, Share & Trends Analysis Report 2025-2034. Market size estimated at USD 7.52 billion in 2024, projected to reach USD 88.84 billion by 2034, at a CAGR of 28.1%. Retrieved from https://www.novaoneadvisor.com/report/cell-and-gene-therapy-cdmo-market
[6] U.S. Food and Drug Administration. (2016). Applying Human Factors and Usability Engineering to Medical Devices: Guidance for Industry and Food and Drug Administration Staff. Available at: https://www.fda.gov/regulatory-information/search-fda-guidance-documents/applying-human-factors-and-usability-engineering-medical-devices
[7] International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH). (2008). ICH Harmonised Tripartite Guideline: Pharmaceutical Quality System Q10. Available at: https://database.ich.org/sites/default/files/Q10%20Guideline.pdf
[8] Health and Safety Executive (HSE). (2023). Human Factors Delivery Guide for COMAH Sites (December 2023 edition). Available at: https://www.hse.gov.uk/comah/assets/docs/hf-delivery-guide.pdf
[9] BioPhorum Human Performance Workstream overview. Available at: https://www.biophorum.com/workstream/human-performance/
[10] EudraLex. The Rules Governing Medicinal Products in the European Union (Vol. 4). Chapter 1 (1.4). Available at: https://health.ec.europa.eu/document/download/e458c423-f564-4171-b344-030a461c567f_en?filename=vol4-chap1_2013-01_en.pdf
Acknowledgements
This blog was developed collaboratively with Claude (Anthropic’s AI assistant, Claude Sonnet 4.5). Claude assisted with:
- Research and synthesis of FDA regulatory data, pharmaceutical recall statistics, and cell and gene therapy market projections
- Structural development of the seven-topic framework adapted from COMAH principles
- Analysis of the gaps between current pharmaceutical quality practices and systematic human factors approaches used in other high-hazard industries
- Drafting and iterative refinement of the content based on human factors principles and pharmaceutical manufacturing context
The core concepts, framework structure, and strategic insights reflect expertise in human factors and pharmaceutical quality systems, whilst Claude provided research support, structural organisation, and drafting assistance. All factual claims have been verified against cited sources, and the authors take full responsibility for the content and recommendations presented.
About the Authors
Dominic Furniss is a Senior Human Factors Consultant at Human Reliability Associates, specialising in Safety Critical Task Analysis across high-hazard industries including pharmaceutical manufacturing. His work focuses on bridging proactive risk assessment and reactive incident investigation through systematic human factors methodologies. Dominic leads the Human Reliability Academy and its associated training programmes.
Connect on LinkedIn: https://www.linkedin.com/in/dominicfurniss/
Julie Avery has over 30 years of experience in quality systems, , pharmaceutical and cosmetics manufacturing. She is a Six Sigma and Lean Black Belt. Julie has successfully led and delivered strategic and tactical integration of human factors programmes into Quality Management Systems on both a global and site level. She has a wealth of experience in pharma and biopharma quality and associated risk management systems, including remediation programmes following regulatory observations. Julie has designed and delivered comprehensive educational and training GxP programs.
Connect on LinkedIn: https://www.linkedin.com/in/julieavery1/
