The reproducibility of ophthalmic utility values.

Abstract

PURPOSE: Utility values have been used in the ophthalmic literature to measure the quality of life associated with a health state. By convention, a utility value of 1.0 is associated with perfect health, and a value of 0.0 is associated with death. Construct validity of utility values has been demonstrated, particularly in regard to decreasing utility values as the vision decreases in the better seeing eye, but long-term test-retest reliability has not been demonstrated. The purpose of this study was to demonstrate the test-retest reliability of ophthalmic utility values. METHODS: One hundred fifteen patients with ophthalmic diseases and stable visual acuity underwent time trade-off utility analysis with retesting at various intervals ranging from 1 month to 2 years. The results were analyzed using the Wilcoxon signed rank test. The study was designed to have an 50% power, using a two-sided alpha of 5%, to be able to detect a 10% difference between the test and retest groups. RESULTS: The mean time from testing to retesting was 0.87 years, with a median time of 1.0 year and range of 1 month to 2 years. The mean utility value in the test group was 0.766 (SD = .21; 95% CI, 0.730-0.802), while the mean utility value in the retest group was 0.763 (SD = .22; 95% CI, 0.724-0.802). The difference between the means of the test-retest groups was not significant (P = .99). The intraclass correlation between the initial and follow-up utility scores was .5246 (P < .00005). CONCLUSIONS: Ophthalmic utility values appear to have good test-retest reliability over prolonged periods of time. This information is important because it gives researchers increased confidence in the validity of basic tools for ophthalmic cost-effective (cost-utility) analyses.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 56 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.