Autonomic quality management system ledger

Abstract

A computer-implemented method of recording information from a quality management system digitally, e.g. using blockchain. The information retrieved from the documents is assigned a value and these val

Description

This invention provides a solution to digitalisation of Quality Management System (QMS) states within systems of systems, such as those found in Good Manufacturing Practice (GxP) and other environments. A Quality Management System within an organisation is typically based upon an applicable standard, such as ISG9001:2015. These standards are implemented as a set of manual documentation controls around systems and processes within an organisation. This documentation both defines the design of the system and captures any associated activities. QMS documentation is a regulatory requirement for GxP compliant organisations, in order to provide evidence to regulatory and intercompany auditors that the processes in place are being followed and improved. It is important that the records are original and truthful. The Quality System can become one of the largest cost centres for some organisations to maintain and there is typically a high human labour effort requirement. When automated, computerised or other electronic recording systems have been integrated, the level of complexity in quality management rises, but it increases still further when considered within a smart factory or Industry 4.0 paradigm. A framework for management of electronic systems and records is set forth in Good Automated Manufacturing Practice (GAMP) published by International Society for Pharmaceutical Engineering (ISPE). GAMP5 is subtitled “A risk based approach” which seeks to manage this complexity through the grading of activities according to relative risks as part of the process. Nevertheless, this approach within the GAMP framework still remains a manual process and is an additional layer of manual controls. As has been discussed in some academic research, QMS was not designed with managing these levels of complexity in view and can for this reason be described as fundamentally incompatible with QMS. This inventions solution to the already briefly described rising complexity problem, is adopting the representation of the QMS as digital states and corroborating those states alongside QMS documented activities by capturing them in a time-series ledger. The digital QMS states are represented, maintained and distributed throughout the system by employing the concept of Autonomic Manager (AM) agents. An AM is a software construct in the form of a control loop, which is defined within the study of Autonomic Computing(AC) to realise systems with self-managing properties and is underlying this inventions methodology. This invention is within the domain of QMS, but not associated with the so-called electronic QMS, or eQMS, which is specifically the electronic management of QMS documents. There is no claim here being made with respect to the existing concept of an Autonomic Manager. The invention specifically concerns utilising the same as a design component with respect to achieving digitalisation of system states in a QMS system. Capturing quality system state data and activities overtime by means of an integrated ledger, scheduling and self-management of quality activity allowing for the selfidentification and self-correction of quality gaps by autonomous systems are features this inventions approach uniquely facilitates in QMS. This has the overarching goal of demonstrating compliance with the relevant guidance and regulations in line with policies by automatically capturing the present QMS documentation state at any given endpoint and its activities. The invention will now be further described by way of example and referring to the accompanying drawing. Figure 1 shows the integration of dual-purpose AM’s into the QMS system and how the digital representation of QMS states is being achieved. An AM which provides a touch-point for the QMS system to any individual quality endpoint and / or quality component through any supported sensor with respective value(s) and communicates those state changes to a supervisor AM, is hereafter referred to as an Autonomic EndPoint (AEP) (1). In this context, the endpoint refers to any element of the quality system which is providing feedback to this invention through its associated AM, whether hosted directly upon a full computing endpoint proper, or facilitating a collection of individual interfaces utilising smaller computing devices, such as Internet-of-Things devices. A supervising AM monitors state and event messages from the registered AEP’s is hereafter referred to as the Autonomic Ledger(AL) (2), so called because its design comes with an integrated on-loop blockchain ledger (3), which provides immutable corroboration of quality activities in given states by recording these events. AM control loops (4) take a 4-phase form which contains Monitoring, Analysis, Planning and Execution, to which is added a Knowledge layer shared and updated throughout the loop (MAPE-K). This loop is informed by taking inputs from Sensors with their respective values and outputs results to the system touchpoint through a mechanism known as Effectors. This loop form provides the structure for an approach to system self-management which takes place as follows. During monitoring the AM loop takes in messages as one group of sensors and any other defined sensor inputs on the given endpoint. These sensors may include information and values provided by interfacing with a computing endpoint operating system, such as the the system identification, presently logged on person, or identification, current hardware state, running processes, executables and other such relevant examples. In the analysis phase, the inputs are taken and processed through policies which may be fixed or dynamic and seek to optimise where possible based upon new or comparative information. The planner can also be thought of as a scheduler and comprises response activities whether immediate, or scheduled at a pre-defined time base. The planner is afterward processed through the execution phase, as are any pending recordings and / or knowledge updates. The AL represents all registered AEP’s to itself within a node registration state table (5), which contains the AEP node identification, alongside their current document state values, as well as network registration information so that the AEP nodes can be tracked and updated by the AL. Whereas, the AEP represents its own self-state in an individual state record with no knowledge of other nodes states (6). (5) on (2) functions as the primary record of present digital states, such that any temporary state changes to (6) on (1) do not result in loss of this valid state. The state information itself, which is to be represented may chiefly be the reflection of QMS documentation states and their respective sub-states. Any state can be represented that has been defined in the AM policy and may vary between organisations and industries, but an example of how these states could be defined will now be provided. If using GAMP as a reference, the qualification phases may be used, such as Design Qualification (DQ), Install Qualification (IQ), Operational Qualification (OQ) and Performance Qualification (PQ). When a system is fully qualified it would enter into its Standard Operating Procedure (SOP) training state and any faults would for example be shown as a Non-Conformance (NO) or fault state. Other corrective quality interventions could be captured in a variance, or deviations (DEV) state. The invention represents these QMS states as digital values upon the AEP, relayed to the AL as activities. AEP to AL internode communication is in the form of a multipart message (7) which must satisfy GAMP requirements of being attributable (e.g. to a person or identification), being human readable and legible, containing a confirmed contemporaneous time stamp, providing a reason for the record, as well as the underlying information or data that is to be recorded. It’s suggested that this can be suitably represented in the form - who, what, where, when, why (5W). The AL upon receipt of this message, analyses it through a response policy (8), updates the node registration state table (5) as applicable and then communicates that state back to the AEP nodes in the form of a command message, which is conceived at a minimum should be 2-part message indicating 1) that a state is to be updated and 2) what that state is (9). Both the AEP and AL implement a subset of messages which are treated as commands by the respective onboard policies and call functions upon the respective AM utilising or facilitating the QMS state management (10). An example of such a command may be to lock further activities upon the end-point because it has been found to be out of compliance according to some policy definition, effectively allowing the system to self-protect the QMS. Per GAMP and 21 CFR part 11 standards employed in GxP industry, no quality record may be lost and in order to ensure this, the invention implements non-volatile messaging buffers at both ends of the node communication (11). The invention is therefore suitable for networked systems, with each endpoint having an associated AM comprising a quality node and can resume following any interruption, recovering all state and message information from non-volatile storage. Finally, though having the goal of system self-management, the invention can accommodate the notion of human-on-the-loop by integrating the required response policies within (8)

Claims

The claims for this invention are:

1. Representing Quality Management System (QMS) states digitally upon computing endpoints.

2. Related to claim 1, endpoint digital states monitored and maintained by an Autonomic Manager (AM) control loop being an Autonomic End-Point (AEP) which can represent and effect any stateful element of the QMS digitally.

3. AEP quality states recorded and corroborated by an Autonomic Ledger (AL) utilising a blockchain time-series ledger as long-term non-volatile and immutable memory of those states and any associated recorded quality system events.

4. Related to claim 1, the AL maintains and updates any connected AEP states, such that any temporary changes to AEP internal quality states do not override the validated and documented quality state upon the AL, which is provided by means of an internal QMS node registration state table.

5. Related to claim 3, the appropriate response of the AL to AEP quality state changes is guaranteed by quality validated response policies.

6. Related to claim 5, appropriately timed response of the AL to AEP quality state changes and scheduling of any other quality system activity takes place through the AM’s internal scheduler / planner taking into consideration policy adjustments.

Images