Minimal important change (MIC) based on a predictive modeling approach was more precise than MIC based on ROC analysis

Minimal Important Change and Minimal Detectable Change in Activities of Daily Living in Community-Living Older People

OBJECTIVE

To estimate the minimal important change (MIC) and the minimal detectable change (MDC) of the Katz-activities of daily living (ADL) index score and the Lawton instrumental activities of daily living (IADL) scale.

DESIGN

Data from a cluster-randomized clinical trial and a cohort study.

SETTING

General practices in the Netherlands.

PARTICIPANTS

3184 trial participants and 51 participants of the cohort study with a mean age of 80.1 (SD 6.4) years.

MEASUREMENTS

At baseline and after 6 months, the Katz-ADL index score (0-6 points), the Lawton IADL scale (0-7 points), and self-perceived decline in (I)ADL were assessed using a self-reporting questionnaire. MIC was assessed using anchor-based methods: the (relative) mean change score; and using distributional methods: the effect size (ES), the standard error of measurement (SEM), and 0.5 SD. The MDC was estimated using SEM, based on a test-retest study (2-week interval) and on the anchor-based method.

RESULTS

Anchor-based MICs of the Katz-ADL index score were 0.47 points, while distributional MICs ranged from 0.18 to 0.47 points. Similarly, anchor-based MICs of the Lawton IADL scale were between 0.31 and 0.54 points and distributional MICs ranged from 0.31 to 0.77 points. The MDC varies by sample size. For the MIC to exceed the MDC at least 482 patients are needed.

CONCLUSION

The MIC of both the Katz-ADL index and the Lawton IADL scale lie around half a point. The certainty of this conclusion is reduced by the variation across calculational methods.