Digital Transformation Dissertations
Permanent URI for this collection
Browse
Recent Submissions
Item Evaluating Efficiency Gains and Security of LLM-Driven Test Generation for Computerised System Validation: A Compliance-Focused Analysis of Life Sciences Testing Processes(2025) Vladimirov, DaniilAbstract Pharmaceutical computerized system validation remains documentation-intensive, consuming substantial project effort and impeding Pharma 4.0 adoption. The CSV market grew to $3.92B in 2024 and is projected to reach $14.02B by 2037, highlighting the scale of optimization opportunity. This thesis addressed the tension between regulatory assurance and agility by developing and empirically evaluating a compliance-aware framework that uses Large Language Models to automate Operational Qualification (OQ) test generation from User Requirements Specifications (URS) under GAMP 5 (2nd ed.), 21 CFR Part 11, EU Annex 11, and ALCOA+ constraints. The methodology employed a five-agent, event-driven architecture (GAMP classifier, context provider, research analyst, SME consultant, OQ generator) with confidencegated handoffs, a fail-closed no-fallback policy, and full audit trails; evaluation used 30 synthetic URS spanning GAMP Categories 3–5, K=5 self-consistency, risk-based scoring aligned to ALCOA+, and predefined quantitative metrics. Results demonstrated 96.7% requirements coverage (target ≥95%), 91.3% categorization accuracy, and 7.4 minutes average processing per document. Migration to the open-source DeepSeek model reduced cost by 91% while preserving performance. Security controls achieved 100% semantic preservation with zero unsafe transformations; however, end-to-end completion was 76.7%, below the 90% reliability target, indicating variance and edge-case sensitivity. This research contributes the Compliance-Aware AI Engineering paradigm, establishing regulatory constraints as first-class design parameters, and validates a practical multi-agent architecture for auditable, GxP-aligned OQ generation. In practice, the framework offers a staged implementation path with measurable efficiency gains and clear governance (traceability, authority checks, documentation) suitable for regulated deployment. Future work should focus on variance reduction via reproducible multi-run protocols, expanded adversarial testing, and extension beyond OQ to IQ/PQ and multilingual corpora.Item Measuring Digital Maturity in QA/RA and it's Regulatory Readiness in Pharmaceutical Industry: A cross organizational Study(2025) Ragu, GowthamThis study looks at how digital maturity in pharma QA/RA work shows up in day-to-day practice and how it links to regulatory readiness. A 10-dimension Digital Maturity Index (DMI) was used as the overall lens; six dimensions were studied live in interviews (Strategy & Governance, Technology Infrastructure, Data Integrity & Records, Process Automation & Workflow, Compliance Readiness/Assurance, Continuous Improvement), and four were considered in the background (People & Capability, Supplier/Cloud Assurance, Master Data/Interoperability, Performance & Insight). Conducted six semi-structured interviews (P1– P6) across QA/MES and RA operations. Transcripts were coded line-by-line; themes were compared across cases; each dimension was rated 1–4 with a one-line reason. Findings are clear. Controls are steady: people describe draft → review → approve, role-based access, “supersede, not delete”, and audit-trail checks. These behaviours sit at level-3 and relate to smoother submissions and stable audits. Differences in readiness come from flow, not from the basic control layer. Where MES/workflows stop on exceptions and manual pockets sit outside systems (e.g., RA reuse/notifications in spreadsheets), cycle-time increases and people do double work. Where teams added small pre-checks or used auto-routing, clock-stops reduced and movement was faster. Technology is mostly stable, but integration gaps (e.g., RIM↔EDMS duplicate typing) and one paper-heavy site explain delays. In RA, one case showed level-4 assurance due to a full change → submission → approval trace. The study suggests practical fixes that can be done this quarter: write a short rationale in periodic reviews; add pre-checks at the common stop points; align 8–10 master-data fields between systems; keep a tidy supplier pack; and make the weekly pipeline data-led with two simple KPIs. These small steps connect digital maturity to everyday readiness: fewer queries, fewer clock-stops, shorter cycle-time, and smoother approvals.Item Public Trust and Perception of AI-Powered Health Chat Bots: A Qualitative Study of Digital Symptom Checkers in Ireland(2025) Pradeep, TheerthaThe growing popularity of artificial intelligence (AI) in the healthcare industry has facilitated the implementation of AI-powered symptom checkers (SCs) that can be used to give patients immediate access to health-related information. Although they have potential, little is known about the issues of trust, usability, and integration into the current healthcare systems in Ireland. This paper explored the attitudes and perception of SCs among young adults in Ireland, with the view to determine the factors that affect their uptake and sustained use. The qualitative interpretivist design was used with semi-structured interviews of participants aged 18-35. Data were analysed thematically to identify both cognitive and affective side of trust. The results indicated that, although SCs are valued as convenient, discrete and a source of reassurance about minor ailments, participants were skeptical about complex or urgent care. It was revealed that trust is a complex phenomenon, which depends on accuracy, emotional appeal, transparency, personalization, and past medical experiences. Users often used chatbot suggestions in conjunction with other sources or consultations with professionals and did not regard SCs as alternatives to healthcare providers. The obstacles to adoption were unclear or over-cautious outputs, lack of empathy in design, issues of data privacy and no formal regulation. On the other hand, integration into government healthcare systems and acceptance by HSE were regarded as ways of becoming more accepted. This study adds to the knowledge of technology acceptance and trust in medical AI and provides recommendations on empathetic design, transparent communication, regulatory oversight, and prudent personalization as ways to promote sustainable adoption.Item How Digital Technologies are changing drug discovery: A comparative analysis with focus on Alphafold and other digital tools in India.(2025) Vasant Tambe, SwapnilThis thesis discusses the impact of digital technologies, specifically artificial intelligence (AI), on drug discovery in India, especially the target identification phase. AlphaFold, developed by DeepMind, is mentioned as one of the greatest innovations in protein structure prediction and its significance is compared to those of other AI-based applications including IBM Watson, Atomwise, and BenevolentAI. The study is positivist, quantitative in nature and based on an online survey administered to students, early-career researchers, and industry professionals in India. Responses of 129 participants gave insights on awareness, adoption, and benefits as well as challenges of using AI in drug discovery. Data were analyzed with descriptive statistics and visual comparison of students and professionals. The results indicate that students possess more theoretical knowledge about the AI tools like AlphaFold, whereas professionals reveal more variated but restricted practical use, with most of them relying on in-house tools. The benefits noted by the two groups were that of speed in target identification, better accuracy and lower costs. Nonetheless, low skills, prohibitive costs, data availability, and regulatory uncertainty remain some of the challenges that limit the wider adoption in India. The research finds that, although the pharmaceutical industry in India has started to experiment with AI in research, its wide-scale integration has been low. It recommends that more investments in AI training, improvement of data infrastructure, and friendly regulations should be introduced to encourage adoption. In academia, the gap between awareness and practice can be filled by integrating AI tools into the curricula. On the whole, this dissertation helps us to realize how AI can help speed up the pace of earlystage drug discovery in India and provides viable suggestions to researchers, industry professionals, and policymakers to enhance digital transformation in the pharmaceutical sector.Item Evaluating the Role of Digital Literacy in the Adoption of AIEnabled Health Technologies for Dementia Care: Perspectives from Professional and Family Caregivers in Ireland(2025) Muriuki, Nellius WambuiThis study investigates how digital literacy influences the adoption and use of artificial intelligence AI-enabled health technologies among professional and family caregivers involved in dementia care in Ireland. Through a mixed-methods approach combining a structured online survey (n = 318) and semi-structured interviews, the research explores caregivers’ digital skills, attitudes towards AI, training needs, institutional support, and adoption barriers and enablers. Quantitative analysis revealed that caregivers with higher digital literacy reported greater readiness to adopt AI tools, higher trust in technology, and stronger perceived usefulness of AI applications such as remote monitoring, predictive alerts, and virtual assistants. Professional caregivers demonstrated higher digital proficiency and AI exposure than family caregivers, who nonetheless expressed strong interest in training and support. While 73% of participants believed AI could improve dementia care quality, over 90% stated that formal training would be essential for effective adoption. Qualitative findings highlighted concerns about data privacy, usability, and the emotional impact of relying on AI in caregiving. Many family caregivers described feelings of isolation and lacked institutional support, whereas professional caregivers cited inconsistent workplace policies and insufficient resources. Both groups emphasised the importance of trustworthy, user-friendly, and ethically designed technologies tailored to their distinct needs. The study concludes that digital literacy is a key enabler of AI adoption in dementia care, but structural disparities,such as training access and institutional backing, create unequal opportunities between caregiver groups. Recommendations include community-based training programmes, inclusive technology co-design, and clearer national policy frameworks. These findings inform strategies to ensure equitable, effective, and human-centred digital transformation in dementia caregiving.