Smart Pharma: description and data-driven softwares

From a description-driven approach, this means describing each activity performed by each stakeholder in a diagram. Example of a schema (“JSON”) at the start of production (the green step in the workflow above):

The excel version of this schema simplifies the collaboration between the business and the developers while offering a documentary traceability. The excel document will be used as the basis for the software specification.

The diagram is a representation of your business know-how and your automation gained from your experience. It allows you to capitalize on all the information necessary for the smooth running of your business. Based on this data schema, the user interface will be generated automatically. This is a low-code approach.

After release, the “IoT” object (batch) will look like this :

In this approach, particular attention is given on the description phase of your data models that will make up your business. This phase of business analysis is crucial to achieve a robust data model that can obviously evolve thereafter. The software is tailor-made in its configuration, but not in the development of the code itself, in order to limit costs.

The interconnection between your business objects is an equally important component in this approach. For example, the “batch” object will be connected to the “packaging” object, which in turn will be connected to the “packaging production line” object. Thus the context of your data will be used for further analysis such as traceability, quality analysis, …

The advantages are numerous:

• It’s easy to extend the schema of an object a posteriori without having to question the whole software implementation.
• It’s easy to add a task to a workflow without having to question the entire structure of the user screens.
• The evolution of the software is guaranteed so that it accurately represents your business and important data
• The integrity of your data is ensured (1) through descriptive schemas that “force” data quality and (2) through the representation of your business process by a workflow.

This approach is particularly powerful for the life-science and agri-food industries where government regulations (FDA, etc.) impose very high requirements in terms of data governance and traceability (ALCOA+).

The software frameworks built on this approach show a series of strong points for pharmaceutical production monitoring:

• Their development is agile (addition of business functional bricks in rapid increments).
• The traditional V-shaped development applied in the pharmaceutical world as recommended by GAMP5 is respected while being accelerated.
• GxP software qualification is simplified by the design of the software itself.

The framework used was invented at CERN in order to supervise the development of the particle accelerator crystals. Its robustness is not to be demonstrated!!

More information on the crystal open source framework invented at CERN

A summary on Cristal technology

Many projects that want to take advantage of new technologies such as artificial intelligence (AI) or machine learning (ML) fail because the data source is not sufficiently “right”. A prerequisite for any data analysis project is to have a clean, reliable and anomaly-free database.

There is also increasing talk of Data Driven Business (DDB) or the data-driven enterprise. The descriptive software approach presented here is a solid foundation for being able to generate added value on the basis of your data in order to gain a competitive advantage and thus move your business towards a Data Driven Business.