SURE Test Accuracy for Decisional Conflict Screening among Parents Making Decisions for Their Child

Abstract

Background. We aimed to validate the SURE test for use with parents in primary care. Methods. A secondary analysis of cluster randomized trial data was used to compare the SURE test (index, higher score = less conflict) to the Decisional Conflict Scale (DCS; reference, higher score = greater conflict). Our a priori hypothesis was that the scales would correlate negatively. We evaluated the association between scores and estimated the proportion of variance in the DCS explained by the SURE test. Then, we dichotomized each measure using established cutoffs to calculate diagnostic accuracy and internal consistency with confidence intervals adjusted for clustering. We evaluated the presence of effect modification by sex, followed by sex-specific calculation of validation statistics. Results. In total, 185 of 201 parents completed a DCS and SURE test. Total DCS (mean = 4.2/100, SD = 14.3) and SURE test (median 4/4; interquartile range, 4–4) scores were significantly correlated (ρ = −0.36, P < 0.0001). The SURE test explained 34% of the DCS score variance. Internal consistency (Kuder-Richardson 20) was 0.38 ( P < 0.0001). SURE test sensitivity and specificity for identifying decisional conflict were 32% (95% confidence interval [CI], 20%–44%) and 96% (95% CI, 93%–100%), respectively. The SURE test’s positive likelihood ratio was 8.4 (95% CI, 0.1–17) and its negative likelihood ratio was 0.7 (95% CI, 0.53–0.87). There were no significant differences between females and males in DCS ( P = 0.5) or SURE test ( P = 0.97) total scores; however, correlations between test total scores (–0.37 for females v. for –0.21 for males; P = 0.001 for the interaction) and sensitivity and specificity were higher for females than males. Conclusions. SURE test demonstrated acceptable psychometric properties for screening decisional conflict among parents making a health decision about their child in primary care. However, clinicians cannot be confident that a negative SURE test rules out the presence of decisional conflict.