ALCOA-C, eSource and Electronic Data Capture (EDC)

Marrying Quality and Credibility with eSource and EDC

Data collected in human trials, whether they are clinical, nutritional or public health-related should meet sufficient quality standards. These standards are outlined by the Food and Drug Administration (FDA) and through the international principles of good clinical practice (GCP). Any research conducted which reports outcomes, findings and claims according to the best practices, must be robust and credible. The foundations of these outcomes are the data itself, built on a bedrock of quality. Data must therefore be:

‍

• Attributable
• Legible
• Contemporaneous
• Original
• Accurate
• Complete
• Consistent, Enduring, Available (ALCOA+ terms used often)

Let's visit each of these themes and see how we can implement this best in ePRO and EDC systems.

‍

Attributable

‍

As the name suggests, we must be able to attribute data to an original (I.e. an individual). Who were they? When did they submit this data? Are we sure it was actually them? Have the data been edited? If they have, can we justify why this was needed, what was edited, and by whom?

Implementing these in EDC/ePRO:

• Secure login with password and/or 2-factor authentication
• Password policies
• Single sign-on (SSO) options
• Timestamp
• Edit logs and notetaking

Legible

‍

Electronic data capture (EDC) probably serves this principle best. Since defined keyboards, letters, languages and fonts are used, the risk of illegible records written by hand becomes much less possible. However, in the instance that some hand-written records need to be captured electronically (by scanning, photographing, etc.) transcription of these items at a later date becomes tricky and risky.

‍

Implementing these in EDC/ePRO:

• Be sure to use validated data types and field limitations
• Care taken when transcribing

‍

Contemporaneous

‍

Data should be collected in a timely manner. With paper records, there is always a risk that someone completes the study and just fills in all the data at the end. While it’s very difficult to know that this was the case when we receive data from participants, electronic data collection once again proves incredibly effective in this situation. Robust and reliable systems will have timestamped events which record every item that was entered.

Implementing these in EDC/ePRO:

• Timestamps
• Limitations on data entry (time windows for permitting data entry)

‍

Original

‍

Data should always be original. You might have heard of the terms ‘at source’ and ‘eSource’ (we covered this in a recent post). The very first instance of data being recorded is called the source. Whether this was on a post-it or on an electronic system – this is the source. The original must be kept, and, while it might not be relevant any longer, for example, if a mistake was rectified and the original is no longer appropriate, it must be kept for audit purposes.

Implementing these in EDC/ePRO:

• Keep all records for audit purposes
• Keep timestamps to set aside originals from amendments
• Never overwrite original records
• Use versions numbers if and when appropriate

‍

Accurate

‍

For trust in results and reporting, data collected must be accurate. There must be no room for error and mistakes must be identified and rectified immediately.

Implementing these in EDC/ePRO:

• Take care when entering data
• Field limitations
• Data validation
• Changes permitted with audit logs

Completeness

‍

Lastly, are your records complete? A data table of the results is perfect for reporting, but we must know where it came from. This includes all the metadata surrounding staff and participants and when they logged in to collect the data. In addition, any supporting material like files and protocols which are also all version controlled should be stored where they are relevant to this process.

Implementing these in EDC/ePRO:

• Adherence to the protocol
• User access roles for everyone involved in the trial
• Defined roles of all staff involved
• Training on ePRO/EDC systems
• Flexibility

‍

‍

Consistent, Enduring, Available - ALCOA+

As research hinges more on technology and on better quality data than ever, three additional terms are floated in the ALCOA space

‍

‍

Consistent

‍

Records must be consistent such that any reader can make sense of them. For example, in the case of an audit, can events be traced in a sensible order which depict the trial journey? A common reference point, such as a time on an EDC server can be a good starting point.

‍

Implementing these in EDC/ePRO:

• Ensure items are timestampes
• Where manual records are used and are later digitized, record the time and date with great detail - even timezone can be useful where international/multi-site studies are undertaken

‍

‍

Enduring

When we want to check old records, it makes sense that we are selecing a medium that is future-proofed in some way. Paper records can discolour, and room humiditiy can smudge or smear inks. What is the best way to store such records longitudinally? Even digital records stored on hard-drives are at risk of disk failure.

‍

Implementing these in EDC/ePRO:

• CDs/DVDs are not reliable as long term storage
• SSD hard-drives with multiple backup systems in place (sometimes on the cloud) can be very effective
• For scanning and copying original paper records, avoid thermal papers which can be voided and damaged

‍

‍

Available

‍

Availability of data is an absolute must - particularly in high risk hower phase trials. When the risks to participants health and safety are highest, accessing that data and that participant journey is key to taking action as soon as it is needed. This might unblinding that participant or bridging trials data to electornic health records. Ensuring sensible and human readible records can be produced without delay are very important.

‍

Implementing these in EDC/ePRO:

• With EDC vendors, ask whether there is a standard audit export opion
• Get examples of exports - and even add some dummy data to see an example of how critical data can be exported effectively

‍

As long as we adhere to these core 6 principles (and others in ALCOA+), we can be confident that any results, publications and claims coming from the study (at least from a data quality perspective) are trustworthy. When we are dealing with pharmaceuticals which are potentially lifesaving, this is essential.