Minimal change is sensitive, less specific to recovery: a diagnostic testing approach to interpretability

Area of Science:

• Clinical Outcome Measurement
• Longitudinal Data Analysis
• Physiotherapy Research

Background:

• Interpreting longitudinal patient outcomes is crucial for assessing treatment effectiveness.
• Several methods exist to define minimal important change or final states, but their comparative utility is unclear.
• Direct comparison of these methods using diagnostic utility is needed.

Purpose of the Study:

• To directly compare different methods for defining important changes in patient outcomes.
• To evaluate the diagnostic utility and usefulness of various interpretation methods.
• To identify the most clinically sensible approach for longitudinal outcome assessment.

Main Methods:

• Secondary analysis of longitudinal cohort data from physiotherapy patients with shoulder pain.
• Utilized the Disability of the Arm, Shoulder, and Hand (DASH) outcome questionnaire at baseline and 3 months.
• Compared methods defining positive response based on minimal change, final state, and combined criteria using Kappa statistics and diagnostic testing.

Main Results:

• Moderate agreement (Kappa=0.47) was observed between different interpretation methods.
• Minimal clinically important differences showed high sensitivity but low specificity.
• Final state scores demonstrated lower sensitivity, higher specificity, and superior accuracy compared to change-based methods.

Conclusions:

• A novel approach was employed to evaluate the relative merits of different longitudinal outcome interpretation methods.
• A combination of significant change (greater than error) and/or a final score within general population norms appears most clinically sensible.
• This combined approach offers strong diagnostic accuracy for assessing patient outcomes in physiotherapy.