1	Design and implementation of a high throughput sample processing operation: Learning lessons from manufacturing industry Paul Downey Operations Director, UK Biobank paul.downey@ukbiobank.ac.uk
2	Protocol development – science defines the “what”? Prof. Paul Elliott chaired a Sample Handling and Storage development group in December 2003 Its remit was to make recommendations on: Blood collection protocols and volumes Transport of blood from collection centres Blood processing protocols Archiving of fraction aliquots
3	The following tubes will be used to collect blood from participants
4	The operation will be high throughput, reproducible and standardized….. defining the how?
5	Learning the lessons from the past – What we shouldn’t do! UK Biobank is a new organisation, no legacy ANYTHING, a great opportunity to get it RIGHT! Recognised Industry Errors: Focus on high utilisation and nominal installed capacity Reactive process with no control of number or quality of inputs/outputs or constraints of the supply chain No standardisation of process Unsuitable facilities/staff (research culture)/organization structure Focus on point optimisation, whole process view On line process development and testing of b-technology All In-house, Why? – no outsourcing or use of off the shelf components, DIY approach Obsession with technology
6	Industrialised implementation – adopting manufacturing engineering principles
7	Organizational Change: Establish the correct team and style of management UK Biobank followed a characterised organisational evolution: Understand the difference between project and operation Establish the team to deliver Project leader Scientists Engineers Project managers Many precedents for successful project teams: Kelly Martin - Lockheed Martin Skunkworks General Groves – Manhattan Project
8	Understanding the system in the first order – before you invest Understand the system and its goals and how this will deliver the organizational objectives before investment is committed
9	Then design the process in the first order Base on net realisable throughput. Nominal peak component speed is irrelevant. Capability: how the component must perform Process repeatability and tolerances Reliability (MTTF, MTTR, unmanned operation) Peak and net throughput Set-up and maintenance Redundant, de-coupled parallel cells Reliability of a cell based system is the sum of its parts Reliability of a serial system of single units system is the product of its parts
10	Design the process in the first order Recognise the complexity and point of hysteresis
11	Design and prototype – build in process quality assurance Whole system quality and in-process control FMEA Taguchi – Robust design
12	Design - -80^oC automated sample archive
13	Design - -80^oC automated sample archive
14	Design – automated blood fractionation platforms
15	Design – automated blood fractionation platforms
16	Prototype - risk elements of the -80C store
17	Design and prototype – build in quality assurance Poke Yoke Contact defects – correct inputs Constant number defects – correct processing Performance sequence defects – correct sequence Error handling strategies: Shut out Attention Total system validation and revalidation Large, complex processes are prone to propagation of small errors SPC - Measure your process, data driven process improvements
18	UK Biobank bespoke laboratory facility
19	Build fit for purpose facilities
20	Implementation and commissioning – transition from type 2 to type 1 at the appropriate time The challenge: A fixed process (inc. outsourcing/ISO) New and established technologies New facilities New people New systems Transition from project to operation (often with the same staff) Integration with the whole process Statutory obligations – H&S, HTA Managing change
21	Operating an industrialized process – type 1 operation Running the process as an operation, you must implement: Cellular manufacturing approach Controlled, fully understood processes/technology Clear predictability/control of the supply chain Implement product quality process – QA & QC schedule Proper objectives and review Scheduled maintenance New technologies/capacity developed off line (CCP) Proper financial control and reporting Focus on key measures Kaizen – continuous improvement (vs BPR) – bottom up
22	Continuous Improvement: Optimising Batching – product quality Identification of optimal run time – balance of throughput, batch size, downtime, automation initialisation and lability of product
23	Optimising Batching – product quality
24	Operating an industrialized process
25	Summary Learn lessons from other industries Avoid a DIY approach; outsource Recognise project and operation phases Understand the whole system; plan for uncertainty Prototype and build quality into cellular processes – risks fully understood and mitigated Implement the appropriate culture Throughput and quality in a professional cellular manufacturing environment Kiazen - Continuous incremental improvement (TQM)