What is the minimal clinically important difference (MCID) for the neck dosability index in radiation therapy?

Neck Disability Index MCID

Direct Answer

The minimal clinically important difference (MCID) for the Neck Disability Index (NDI) is 7.5 points on the 0-50 scale (or 15 points on the 0-100 scale), representing the smallest change that patients perceive as clinically meaningful. 1, 2

Understanding the NDI Scale and MCID Values

The NDI can be reported on two different scales, which is a critical source of confusion:

• The NDI consists of 10 items yielding a total score out of 50, but some users double the scale to report total scores out of 100 3
• The most commonly used MCID is 7.5 points on the 0-50 scale, which corresponds to 15 points on the 0-100 scale 1, 3
• Anchor-based ROC curve analysis established the MCID at 7.5 points (0-50 scale) by identifying the change score with equal sensitivity and specificity to distinguish patients reporting "Somewhat Better" from "About the Same" on health transition assessments 2

Substantial Clinical Benefit Threshold

Beyond minimal improvement, a higher threshold represents substantial benefit:

• A 10-point decrease on the NDI (0-50 scale) represents substantial clinical benefit (SCB), distinguishing patients who report feeling "Much Better" from those who are only "Somewhat Better" 2
• The SCB threshold of 9.5 points (rounded to 10) indicates a more robust improvement that clinicians should target when evaluating treatment effectiveness 2

Critical Implementation Considerations

Scale Congruence is Essential

• Inappropriate implementation of the NDI MCID occurs in 10% of studies, while uncertain implementation due to poor reporting occurs in 66% 3
• Users must ensure the magnitude of scales used for NDI data and MCID are congruent—applying a 7.5-point MCID to data reported on a 0-100 scale is incorrect 3
• When NDI scores are reported on the 0-100 scale, the MCID should be 15 points, not 7.5 3

Range of MCID Values by Calculation Method

• Distribution-based methods yield MCID values ranging from 6 to 43 points (0-100 scale equivalent), with a median of 18 points 1
• The standard error of the mean produces the smallest MCID values, while minimum detectable change (MDC) yields the largest 1
• Anchor-based methods are most clinically relevant and easily understood by clinicians, whereas distribution-based MCIDs are useful for understanding population-level changes 1

Clinical Application Algorithm

For screening or low-risk interventions:

• Use the lower MCID threshold of 7.5 points (0-50 scale) to identify any clinically detectable improvement 1

For evaluating treatment effectiveness or high-risk interventions:

• Target the median or higher MCID values, or preferably the SCB threshold of 10 points (0-50 scale) 1, 2

When interpreting published studies:

• Verify which NDI scale (0-50 or 0-100) was used for data collection 3
• Confirm the MCID value matches the scale used (7.5 for 0-50 scale; 15 for 0-100 scale) 3
• Be cautious with studies that do not clearly report both the NDI scale and MCID value used 3

Important Caveats

• MCID values are population-derived and the threshold for perceived benefit varies between individuals 4, 5
• The MCID is not perfect in detecting patients experiencing clinically important improvement, and its accuracy (sensitivity and specificity) should be considered 5
• MCID values may vary depending on the patient population and clinical context in which the NDI is administered 5