Minimal important difference thresholds and the standard error of measurement: is there a connection?

Area of Science:

• Health Outcomes Research
• Measurement in Health Sciences
• Clinical Significance

Background:

• Recent studies show varying relationships between standard error of measurement (SEM) and minimal clinically important difference (MCID) or minimal important difference (MID) for health status measures.
• Discrepancies exist in established SEM criteria for MCID/MID across different health-related quality of life (HRQOL) and health status instruments.

Purpose of the Study:

• To compare and review two sets of studies examining the SEM criteria for MCID/MID.
• To investigate the relationship between SEM and MCID/MID in chronic disease and musculoskeletal pain populations.

Main Methods:

• Review of three studies on disease-specific HRQOL measures in chronic disease (heart, COPD, asthma) showing a 1 SEM correspondence with MCID/MID.
• Review of three studies on health status measures in patients with back, lower extremity, and neck pain, indicating approximately 2.3 SEMs for MCID.
• Analysis of patient global change ratings on a 7-point Likert scale to define MCID/MID thresholds in both populations.

Main Results:

• Chronic disease patients achieved MCID/MID with a change of 1, 2, or 3 on a 7-point scale, corresponding to approximately 1 SEM.
• Back pain patients required average global transition scores of 5, 6, or 7 (indicating significant improvement) to reach MCID, necessitating approximately 2.3 SEMs.
• The SEM-based criteria for MCID/MID differ significantly between chronic disease populations and those with musculoskeletal pain.

Conclusions:

• The SEM required to reach MCID/MID is population- and measure-dependent.
• Findings suggest that 1 SEM may be a suitable criterion for MCID/MID in chronic disease HRQOL measures.
• Approximately 2.3 SEMs may be a more appropriate criterion for MCID/MID in health status measures for patients with back, lower extremity, and neck pain.