SF-6D utility index as measure of minimally important difference in health status change.

OBJECTIVES

To combine anchor- and distribution-based approaches to identify minimally important differences (MIDs) for the short-form six-dimension utility index (SF-6D) and to identify variables associated with self-reported health status change.

DESIGN

Descriptive, exploratory, nonexperimental study.

SETTING

United States between April 1, 1999, and October 31, 1999.

PATIENTS: 2,317 participants of SADD-Sx (Study of Antihypertensive Drugs and Depressive Symptoms), aged 50 years or older and with hypertension and coronary artery disease.

INTERVENTION: Patients were randomized into a verapamil SR- or atenolol-led hypertensive treatment strategy and mailed baseline and 1-year surveys.

MAIN OUTCOME MEASURE: SF-6D utility scores for patients completing both surveys.

RESULTS

The pooled mean (±SD) MID change on the SF-6D of patients whose health status minimally changed was 0.035 ± 0.095. The anchor-based change scores had a median value of 0.036 (interquartile range -0.03 to 0.10). One-third and one-half of the SD of SF-6D change scores were 0.035 and 0.053, respectively. Whites were less likely to report minimally improved health status compared with nonwhites (odds ratio 0.59 [95% CI 0.40-0.88]). Change in SF-6D scores improved prediction of health status change.

CONCLUSION

We recommend using the MID range based on all patients combined (-0.03 to 0.10) to interpret SF-6D scores. These estimates can be used in conjunction with other measures of efficacy to determine meaningful changes. SF-6D demonstrates potential utility in predicting minimally important improvement or worsening among patients receiving different pharmacologic medications.