Interpreting score differences in the SF-36 Vitality scale: using clinical conditions and functional outcomes to define the minimally important difference.

OBJECTIVE: To propose the minimally important difference (MID) for the SF-36 Vitality (VT) scale by evaluating the association of score differences with clinical conditions and functional outcomes.
METHODS: Analyses were performed on data from the Medical Outcomes Study (n = 3445). The first analyses regressed VT scores (0-100 scale) on chronic conditions that cause fatigue in order to determine the impact of each condition on VT. The second set of analyses examined the relationship between baseline VT scores and other outcomes at baseline, 1-year, and 7-year follow-up.
RESULTS: VT scores were significantly reduced in patients with anemia [5 points (95% CI 2-9 points)], CHF [6 (3-9) points], and COPD [6 (3-9) points]. Decreases in VT score were significantly associated with increased odds of negative outcomes, including inability to work due to health at baseline [OR (5 points) = 1.27 (95% CI 1.24-1.31), OR (10 points) = 1.62 (1.54-1.71)], job loss at 1 year [OR (5) = 1.13 (1.08-1.19), OR (10) = 1.28 (1.17-1.41)], hospitalization at 1 year [OR (5) = 1.08 (1.05-1.11), OR (10) = 1.17 (1.10-1.23)], short-term mortality [0-18 months-Hazard Ratio (HR) (5) = 1.10-1.71, HR (10) = 1.21-2.39, depending on VT level] and long-term mortality [19+ months-HR (5) = 1.05-1.31, HR (10) = 1.10-1.54]. The mortality risk increase was largest at low VT levels.
CONCLUSIONS: VT decrements of 5-10 points were seen for diseases known to cause fatigue. Further, differences of 5-10 points in the VT score were associated with significant increased risk of negative outcomes. We recommend an MID of 5 points for analyses of groups with VT scores below average. For follow-up of individual patients, we recommend a 10-point difference as important.