Interpreting pathology test result values with comparators (< , >) in Electronic Health Records research: an OpenSAFELY</,>...

Area of Science:

• Health data research
• Clinical pathology
• Electronic health records

Background:

• Numeric pathology test results are often reported as ranges (e.g., "<10") rather than precise values.
• Comparators (<, >) attached to numeric results are crucial for categorizing patients based on clinical thresholds but are often stored separately and overlooked.
• The interpretation of health data can be influenced by the presence and type of these comparators.

Purpose of the Study:

• To identify common pathology tests with attached comparators in routine clinical data.
• To quantify the prevalence of comparators across different test types and geographic regions.
• To analyze the specific numeric values and reference limits associated with comparator-qualified results.

Main Methods:

• Utilized routine clinical data from 24 million patients via the OpenSAFELY platform, with NHS England approval.
• Identified pathology test codes frequently associated with comparator values.
• Reported the proportion of results with comparators, categorized by comparator type and region, along with specific values and reference limits.

Main Results:

• Eleven common test codes were identified where at least 25% of results included comparators.
• Glomerular filtration rate (GFR) tests showed 31-45% of results with "≥" comparators, with values 60 and 90 often representing ranges.
• Other affected tests (blood, urine, faecal) predominantly used "≤" comparators (29-86%).

Conclusions:

• Comparators are prevalent in specific pathology tests within electronic health records.
• While generally having minimal impact for most research, careful consideration of comparators is needed, especially around clinical threshold values.
• The presence of comparators, particularly in GFR tests, can limit the reliable distinction between different disease stages or health statuses.