This article describes a scheduling-approach developed for managing workloads and resources in meeting analyses completion dates in the Quality Control (QC) laboratories of a client in the Pharmaceuticals industry.  

The diagram (below) represents a greatly simplified model of the flow of work through a Quality Assurance or QC analytical laboratory.

In actuality, these laboratories operated multiple HPLC instrument from various manufacturers, had a range of other instrumentation and were staffed by analysts each with different skill sets and qualified to perform specific test-types.

Each of the major steps shown in the diagram break down into further sub-processes; each with their own rules. These and other complexities were taken into account by the scheduling model discussed here.

The work flow model is a scheduling perspective which was used by Btt's automated system to build a more highly optimised* plan for the allocation of work and resources to meet required completion dates in an environment with uncertainties such as: the arrival date/time of samples, variability in both workloads and in the availability of resources (eg. analysts and instrumentation).

Laboratory Work Flow Model (simplified)
 

The Workflow
Samples for analysis arrive at the left hand side where their release into the laboratory processes is controlled by the computed schedule according to the set of delivery requirements, priorities and operational constraints.

Controlling the release of samples for analysis into the Laboratory work flow is important. It is a valuable mechanism for regulating the workloads throughout the process and thereby avoiding overload of any stage.

Overloads means losing close management control of the schedule and plunging into fire-fighting. The challenge is how to get control of the schedule in the first place in a fast changing environment! That's where the approach discussed here comes into its own in bringing stability and responsiveness in a changeable world.

Planning and Scheduling the Work
There are two key methods of planning and scheduling the work required to complete the analysis of each sample: -

1. Forward-Scheduling
Analysis jobs with high-priority urgency (ie. required as-soon-as-possible) are forward-scheduled. That is, they are planned to complete each of the tasks in their analysis as early as possible. The result is that cycle times for these samples are minimised and these samples take precedence in the laboratory schedule over those with lower priorities. However...

2. Backwards-Scheduling
Sample analysis to meet specified deadlines are backwards-scheduled from their required completion dates. Activities across the laboratory are synchronised, and resource allocation decisions are computed in order to meet the deadline/due-date and priority requirements. The output schedule shows the system-planned completion dates for all the analysis work within the defined planning period (4 weeks into the future in this case).

The scheduling system accommodated sample analysis with any combination of forward-scheduled and backwards-scheduled work.

To synchronise the activities across the laboratory to meet planned completion dates,  work-plans are derived for each area of the laboratory from the overall computed schedule. The work plans show the sequence in which to carry out tasks and they show the planned completion dates/times. The work plans indicate which resources are necessary for each activity; resource contention conflicts are automatically prevented in these plans. Where applicable, the work plans also show which samples should be combined ('campaigned' or 'batched') into set-ups or assays to meet the overall plan.

Parallel versus Serial working
Where resource availability and processing rules allow, tasks will be carried out in parallel. The scheduling system seeks to carry out as much parallel work as possible within the constraints of resources availability and according to the specified rules for which sample types can or cannot be combined in campaigns/batches. .

Where resource limitations dictate, tasks are planned to be carried out sequentially. In effect, the system-generated plan configures; the flow of work, the sequencing of tasks and the allocation of resources as required to best meet the daily operational goals of the laboratory.
 

Notes: the exact steps, their sequencing and the analytical techniques/working practices used in the process flow will of course be specific to each laboratory. There may be steps for samples which must be carried out sequentially whilst there may be others which can be carried out in parallel if workloads and resource availability allow.

The challenge for laboratories that operate in an environment with varying sample types, resource constraints and uncertainty as to the arrival date/time of incoming samples is... how to organise the work to make most effective use of resources and facilities to meet required deadlines with the requisite dependability of analysis results?  

The generated Schedule
In real-life application, the scheduling system demonstrated its ability to generate, effective plans projected out, on a rolling basis, over the coming 4 weeks showing: -

• analysis completion dates for each sample

• resource utilisation levels and how much they were loaded day-by-day (or hour-by-hour)

• where and when any bottlenecks would occur (the system-generated schedule takes them into account)

• how samples would be campaigned/batched in each of the laboratory operations

• the planned start/end dates for intermediate work stages (such as the stages shown in the diagram)

Considerations in the Schedule
In generating this plan, the system took into account factors such as: -

• the planned and actual resource availabilities (eg. analysts holidays/training time, instrumentation due for calibration/maintenance, etc)

• retests where they had arisen, or cancellations, and their subsequent effects on other planned work

• the actual arrival of samples versus their expected arrival dates

• the actual progress of samples in work-in-progress versus their planned progress

• rules by which samples can/cannot be campaigned or batched for say, HPLC runs, or other instruments such as MS

• the capacities of each stage, such as the number of samples that an HPLC run can accommodate

• the forward-scheduled and the backwards-scheduled work and the priority of each sample relative to the other laboratory work

Client comments:

“Btt introduced planning logic and disciplines to the laboratory reducing the need for continual rescheduling and firefighting”

"The Btt consultant was able to interact effectively with staff at bench level and with the laboratory management team, readily gaining their confidence and support"

This project was presented by our client at international pharmaceutical industry conferences in Berlin and London. Also presented by Btt at the international LIMS Conference in Barcelona, Spain 2004

* we recognise that no one can ever truly claim to have optimised a process or organisation since it can never be proved.

About the author: Chris Dale

Related articles: Btt Scheduling Engine, Btt in Laboratory Management, European Pharmaceutical Review June 1998

Books: The scheduling system mentioned in this article was based on an automated application of Eli Goldratt's Critical Chain principles. For information on Critical Chain and its underlying Theory of Constraints, see these books: Critical Chain, Theory of Constraints, The Race, The Haystack Syndrome.

Other references, amongst many, were: Project Management in the Fast Lane, Critical Chain Project Management

For general background: Btt principles

For training in Critical Chain using ProChain software (used in the Btt scheduling system): Focus 5 Systems

To raise questions or discuss the subjects further...  questions.