Validity and responsiveness of EuroQol 5 dimension 5 level in the abortion care setting

Utility preference scores are standardized generic health-related quality of life measures that quantify disease burden, severity, and treatment impact on a scale where 0 represents death and 1 represents optimal health. They are widely used in cost-effectiveness research but have not been validated in the abortion care setting.

We evaluated the responsiveness and validity properties of the EuroQol 5 dimension 5 level (EQ-5D-5L), one of the most commonly used utility systems, in individuals undergoing first-trimester medication or procedural abortion.

We completed baseline (N=242) and 30-day questionnaires (N=212) with individuals undergoing first-trimester abortion at 4 sites in the Northeast United States. These included the EQ-5D-5L, Center for Epidemiological Studies Depression Scale, and Perceived Stress Scale. Responsiveness was assessed using EQ-5D-5L baseline and 30-day change scores, effect size, and standardized response mean. Validity properties were assessed by analyzing the EQ-5D-5L score's relationship with depression; stress; patient experience; and postabortion pain, bleeding, and cramping symptoms.

Of the 242 individuals included in this analysis, the mean (standard deviation) age was 27.8 ± 5.9 years, the median parity was 1 (range 0-6), 64.3% were Black, and 14.9% were Hispanic. The mean EQ-5D-5L scores improved from 0.851(95% confidence interval, 0.824-0.879) at baseline to 0.884 (95% confidence interval, 0.858-0.911, P=0.028) at 30 days postabortion; the change, 0.031 (95% confidence interval, 0.004-0.058), was greater than the suggested minimal important difference for preference-based utility measures from other populations. The effect size and standardized response mean were in the very small range overall (effect size 0.145, standardized response mean 0.152) and in the small rage for participants with improvement in the depression scale (effect size 0.473, standardized response mean 0.465) and Perceived Stress Scale (effect size 0.337, standardized response mean 0.392) at 30 days, demonstrating responsiveness of the EQ-5D-5L similar to other conditions. Low (worse) utility scores were highly correlated with high (worse) depression scale (rho>0.5, P<0.001) and Perceived Stress Scale (rho>0.5, P<0.001) scores at 30 days. For the depression scale categories, mean EQ-5D-5L scores ranged from 0.714 (severe depression) to 0.965 (no depression), (P<0.001). For the Perceived Stress Scale categories, mean EQ-5D-5 L scores ranged from 0.605 (high stress) to 0.966 (low stress), (P<0.001). Increased levels of postabortion symptom severity were associated with worse EQ-5D-5L scores: bleeding (P=0.003), pain (P=0.002), and cramping (P=0.008). EQ-5D-5L scores also tracked with participant expectations of bleeding (P=0.006), pain (P=0.003), and cramping (P<0.01). The mean EQ-5D-5L scores were better for participants who had a positive versus a negative experience (P<0.001) and for those who received adequate versus inadequate counseling (P=0.003).

The EQ-5D-5L has good validity properties and is a responsive preference-based utility and general health-related quality of life measure in the abortion-seeking population. It can be used for future cost-effectiveness analyses and other research in this setting.