Introduction
Computer system validation in the Life Sciences is a process to ensure data integrity, particularly patient-related data, that is accumulated and stored in a computer system. One aspect of data integrity is ensuring that data is collected accurately – that is, the value of a data point is transferred to the system in a manner that ensures the same value (as well as metadata, such as units of measure, etc.). A primary objective of validation, then, is to verify that data is entered into the system accurately – that is, that the value of the data is stored in the system as it was intended to be.
Once the data is stored in the system accurately, the validation process then aims to ensure that any changes to the data value while it resides in the computer system are comprehensively monitored and logged, so that at any point in time, it is clear to a reviewer what the datapoint’s value was upon entry, and that each time the datapoint value changed there is a clear record of the user who modified the value, when it was modified, why it was modified, and what the previous and new values where.
There are two primary methods in which data points can be changed – via an internal user or external user.
- The internal user is the system itself. The computer system, which is governed by sets of instructions encoded in the software, has the capability to modify any data point in its control. These programmatic changes may be intended – changing a temperature value entered as Fahrenheit to the Celsius equivalent, for example – or unintended – as when an application erroneously changes a temperature value entered in Fahrenheit to a parts-per-million value.
- The external user is an end-user of the system who interacts with it via the user interface (UI). As with the system user, changes to data from external users can be intentional or unintentional. At a certain level, it really does not matter how the data value changes. The key is that all data changes must be captured and available for review.
Audit History
This is why every system must have a robust method of monitoring and tracking all changes that are made to any data point’s value, regardless of the actor or the intent. As long as the system captures the metadata around the change – the who, what, when, and why – it is possible to know the history of that data point and see how it came to its current value. Because computer systems in the life sciences often process data that are critical to people’s health and safety, having the assurance that the history of each data point is comprehensive is a critical aspect of validation testing.