Meaningful Change

What Is Meaningful Change?

Often we compare two scores. When we calculate the difference between scores (e.g., Score A – Score B = difference between scores), we want to know if that difference is meaningful to patients. Even if the difference is statistically significant, it may not be clinically meaningful. It may not support a determination of treatment benefit or harm and may not identify when clinical action is needed.

Therefore, you should identify if the magnitude of change you see is considered meaningful by patients.

Comparing Groups vs. an Individual

Identify if you are comparing groups of people or one person over time. The magnitude of change that is considered meaningful varies depending on who you are comparing.

Group comparisons include:

• Assessing two or more groups and determining if the difference between the groups’ average score (e.g., means) reaches a magnitude patients consider meaningful (“between-group difference”).
  • Example of a group comparison: Group A’s PROMIS® Physical Function is 60. Group B’s Physical Function is 64. A researcher wants to know if the difference between these two groups is meaningful.
• Assessing a single group of patients over time and determining if the difference in means between two time points reaches a magnitude that patients consider meaningful (“within-group change”).
  • Example of a group comparison: The mean PROMIS Fatigue score for a group of people worsened from 55 to 60. A researcher wants to know if this worsening is meaningful.

For an individual:

• Determining if a person has improved or deteriorated (“within-person change”).
  • Example of individual change: A patient’s Anxiety score improved from 65 to 58 two months later. A clinician wants to know if this improvement is real and meaningful.

The same meaningful change thresholds recommended for PROMIS measures (2-6 T-score points for group comparisons, 5-7 T-score points for individual change) also apply to Neuro-QoL measures.

Additional studies have estimated within-person change for specific Neuro-QoL applications.

Individual Change

Individual change values were estimated for 14 Neuro-QoL measures. Estimates vary based upon the patient’s symptom/dysfunction severity. Index tables and an interactive Excel workbook that calculates individual change values are available (Kozlowski et al., 2016).

Idio Scale Judgment

In 2017, Cook and colleagues published a new method, idio scale judgment, designed to estimate the amount of Neuro-QoL Fatigue score change that is meaningful for people with multiple sclerosis.

• Improvement
  • Mean = -3.5 points (SD=2.6)
  • 50th percentile = -3.3
  • 75th percentile = -5.1
  • 90th percentile = -7.3
• Worsening
  • Mean = 3.2 points (SD=2.0)
  • 50th percentile = 3.2
  • 75th percentile = 4.5
  • 90th percentile = 5.6

Evidence continues to accumulate regarding reasonable estimates of meaningful change score thresholds. There are several things you should keep in mind when using these estimates.

• There is no single number that reflects a meaningful or important change for all patients in all circumstances for all purposes.
• There are different methods for calculating change thresholds (see Meaningful Change), and they will yield different estimates. Some of these approaches are called Minimally Important Difference (MID), Minimally Important Change (MIC), Minimal Clinically Important Difference (MCID), and Meaningful Within-Person Change (MWPC). These approaches typically produce a range (e.g., 3 to 5 points). Additionally, analyses in different samples and contexts will lead to different estimates. Compile results from multiple methods (triangulation) and studies if you can.
• Selecting a value requires a judgment on your part. You should consider, for example, the use of the value. The lower estimate on the reported important change range might be appropriate for group comparisons. A higher estimate is often more appropriate to categorize change in an individual (within-person change). So, if the meaningful change range is 3 to 5 points, it may be okay to use 3 points for a group and 5 points for an individual. Consider the consequences of how the estimate will be used (e.g., what is the cost-benefit ratio of a lower versus a higher threshold for a given context). Evaluate the relative risk of missing people who had meaningful change but were below the selected threshold versus identifying people as having experienced meaningful change when they did not (false positives).
• Values for improvement may vary from values for deterioration.
• Finally, you should keep in mind that most estimates of meaningful or important change are an average across the range of change scores from a group of patients. It may be that some people require more or less change to consider it meaningful. This could in part depend on where they started. In addition, especially in the context of clinical monitoring of individuals, an average change score may or may not be considered important to any given person.

In summary, to select a threshold for comparing scores:

1. Identify how you will use the threshold.
2. Triangulate available data (e.g., recommended thresholds on HealthMeasures.net, published studies).
3. Use your judgment.