Introduction: The DECIDE-LVAD Trial demonstrated the benefit of a decision support tool for patients considering destination-therapy left ventricular assist device (DT LVAD) implantation. We hypothesized that patients with low health literacy or numeracy may have benefited most from the intervention.
Methods: We used the Rapid Estimate of Adult Literacy in Medicine (REALM-R) and the Subjective Numeracy Scale (SNS) to assess patients’ health literacy and numeracy, respectively. A REALM-R score of six or less out of eight identified patients at risk for poor literacy based on the original measure validation. An SNS score less than four out of six identified patients with poor subjective numeracy based on the mean score of the DECIDE-LVAD cohort. We assessed the effect of the decision support tool on LVAD knowledge and values-treatment concordance—concordance between patients’ stated values and their treatment outcome of LVAD or continued medical therapy—and their interaction with health literacy and numeracy. This interaction was assessed using linear mixed models for LVAD knowledge and the difference in Kendall’s tau correlation coefficient for values-treatment concordance.
Results: Of the 248 DECIDE-LVAD patients, 228 with complete literacy and numeracy data were analyzed: 51% (116) had high literacy and numeracy; 18 (8%) had low literacy and high numeracy; 55 (25%) had high literacy and low numeracy; 39 (17%) had low scores for both measures. Average age was 63, 15% were female, 19% were racial/ethnic minorities. Patients with high scores for both measures were older, had more formal education, and higher total household income compared to the other 3 groups. Patients with low literacy and/or numeracy had significantly lower LVAD knowledge scores at baseline. In all groups, LVAD knowledge improved over time. The difference in LVAD knowledge between those with low literacy and high literacy decreased over time (baseline difference 9.4%, p=0.002; 6 month follow-up difference 1.4%, p=0.65). In those with high literacy, the decision support tool intervention increased the LVAD knowledge score by 4.2% relative to the control (p=0.15), while in those with low literacy, the intervention increased the score by 10.6% (p=0.04). For patients with low literacy, values-treatment concordance improved with use of the decision support tool (control vs. intervention tau: 0.234 vs 0.673, p=0.028). We did not observe differences in the effect of the intervention by participant numeracy score.
Conclusion: All patient groups in the DECIDE-LVAD cohort of patients considering DT LVAD had improved LVAD knowledge with the use of the decision support tool. We did not observe a differential effect of the intervention by numeracy score. Compared to those with higher health literacy, patients with low health literacy improved their LVAD knowledge and values-treatment concordance after the intervention.