Data Integrity in Testing Laboratories – More Than Just Good Documentation

A laboratory result may ultimately appear as a single number on a report.

But how many steps did that number pass through before it reached the customer?

The result may begin with an instrument reading, move through software, be copied into a spreadsheet, undergo calculations, be transferred into a LIMS, be reviewed, and finally appear on a laboratory report.

Every transfer, calculation, correction, and manual entry creates an opportunity for error—or, in more serious situations, inappropriate alteration.

That is why data integrity is about much more than keeping good records.

For a laboratory to have confidence in its results, it must also have confidence in the data behind those results.

What Does Data Integrity Mean?

At its simplest, data integrity means that laboratory data remains complete, consistent, accurate, and reliable throughout its lifecycle.

This applies from the moment data is generated until it is reviewed, reported, archived, and eventually disposed of according to applicable requirements.

A commonly used way of describing good data-integrity principles is ALCOA+.

Data should be:

• Attributable – Who performed the activity or generated the data?
• Legible – Can the information be read and understood?
• Contemporaneous – Was it recorded when the activity occurred?
• Original – Is the original data, or an appropriately controlled true copy, retained?
• Accurate – Does the record correctly represent what happened?

The “+” expands the concept to include characteristics such as being complete, consistent, enduring, and available.

These principles apply to both paper and electronic records.

The Final Report Is Only the End of the Data Trail

During audits, it is easy to focus on the final laboratory report.

Is the correct result reported?

Was the report reviewed?

Does it contain the required information?

Those are important questions, but they only address the end of the process.

A stronger review follows the result backwards.

Reported Result → Calculation → Processed Data → Raw Data → Sample

Could someone trace the reported value back through each step?

Could the laboratory explain who generated the data, what calculations were performed, whether anything was changed, and who reviewed it?

If not, the problem may be much larger than documentation.

Manual Data Transfer Creates Risk

One area I pay particular attention to during laboratory audits is manual transcription.

Consider a result that is:

• Generated by an instrument
• Written on a worksheet
• Entered into a spreadsheet
• Transferred into a LIMS
• Copied into a final report

Even when every employee acts carefully and honestly, every manual transfer creates another opportunity for a transcription error.

Where possible, laboratories should consider reducing unnecessary manual data entry.

Where manual transfer cannot be avoided, appropriate verification becomes important.

The question should not simply be:

“Did someone enter the result?”

It should also be:

“How do we know the result was transferred correctly?”

Spreadsheets Deserve More Attention

Spreadsheets are extremely useful in laboratories.

They can also create significant data-integrity risks.

A spreadsheet may contain formulas used to calculate final results, detection limits, recoveries, dilution factors, QC acceptance, or other critical information.

What happens if someone accidentally changes a formula?

Can cells containing formulas be overwritten?

Who can modify the spreadsheet?

Are changes controlled?

Has the spreadsheet been verified for its intended use?

Is the current approved version being used?

A beautifully formatted spreadsheet is not necessarily a controlled spreadsheet.

When spreadsheet calculations influence reported laboratory results, their reliability deserves the same level of attention as other parts of the testing process.

Electronic Systems Introduce Different Questions

Moving from paper to electronic systems does not automatically eliminate data-integrity risk.

It changes the nature of the risk.

Laboratories should consider issues such as:

• Individual versus shared user accounts
• User access levels
• Ability to modify or delete data
• Audit trails
• System backups
• Data retention
• Changes to calculations or processing parameters
• Software configuration
• Electronic approvals
• Protection of original instrument data

For example, if several analysts use the same instrument login, it may become impossible to determine who performed a particular activity.

The data exists—but part of its traceability has been lost.

Corrections Should Tell the Story

Mistakes happen.

An analyst may enter an incorrect sample ID, use the wrong dilution factor, transpose digits, or make another recording error.

Data integrity does not mean records can never be corrected.

It means corrections should be controlled and transparent.

Someone reviewing the record later should be able to understand what changed, who changed it, when appropriate why it changed, and what the original information was.

A correction should clarify the history of the data—not erase it.

Data Integrity Is Also a Culture Issue

Procedures and software controls are important, but laboratory culture matters too.

Employees need to feel comfortable reporting mistakes.

If employees believe that admitting an error will automatically result in punishment, there may be pressure to quietly “fix” records rather than document what actually occurred.

That creates a much more serious problem.

A strong laboratory culture distinguishes between an honest mistake, poor system design, inadequate training, and deliberate misconduct.

The objective should always be to protect the reliability and transparency of the data.

Questions to Consider

Laboratories can ask:

• Can every reported result be traced back to the original data?
• Where is data manually transferred?
• How are manual entries verified?
• Are critical spreadsheet formulas protected and verified?
• Do employees have individual system accounts?
• Can electronic data be changed or deleted without traceability?
• Are original instrument files retained?
• Are corrections transparent?
• Can reviewers reconstruct what actually happened during testing?
• Would we know if critical data were missing?

These questions often reveal risks that are not obvious from reviewing the final laboratory report.

Key Takeaway

Data integrity is not simply about having complete paperwork.

It is about maintaining confidence in laboratory data throughout its entire lifecycle—from sample and raw data to calculations, review, reporting, and retention.

A laboratory should be able to demonstrate not only what result it reported, but also how that result was generated and why it can be trusted.

Conclusion

Reliable laboratory results depend on reliable data.

Procedures, worksheets, spreadsheets, instrument software, LIMS systems, calculations, corrections, access controls, and employee practices all contribute to data integrity.

The more complex the path between the sample and the final report becomes, the more important it is to understand where data can be lost, altered, incorrectly transferred, or inadequately controlled.

A useful question for any laboratory is:

“If a customer challenged this result two years from now, could we reconstruct exactly how we produced it?”

If the answer is uncertain, the laboratory may have a data-integrity risk worth investigating.