Impact of Medication Dose Optimization on Heart Failure Outcomes in African-American Female Patients: A Safety-Net Hospital Experience

Figures

• 

  Figure 1. Odds ratios (ORs) for 30-day readmission across medication classes RAS, Renin-angiotensin system; MRA, mineralocorticoid receptor antagonists; SGLT2i, Sodium-glucose cotransporter 2 inhibitors. The figure illustrates the ORs for 30-day hospital readmission across medication classes and their interactions. The X-axis shows ORs (0.0 to 2.5), with 1.0 as the reference line for no effect. The Y-axis lists: RAS inhibitor dose, Beta Blocker dose, MRA dose, SGLT2i, RAS × Beta Blocker (p=0.04), RAS × MRA (p=0.08), RAS × SGLT2i, Beta Blocker × MRA, and Beta Blocker × SGLT2i. Each point marks the OR, with the horizontal bars representing confidence intervals. Most ratios are near 1.0, indicating little effect on readmission risk. The RAS × Beta Blocker interaction (p=0.04) and RAS × MRA (p=0.08) show slight deviations, suggesting potential impacts with marginal statistical significance.

• 

  Figure 2. RAS inhibitor vs. beta blocker dose interaction RAS, Renin-angiotensin system. Without beta blockers, higher RAS inhibitor doses increase 30-day readmission risk. At goal beta blocker doses, higher RAS inhibitor doses reduce readmission risk. Low beta blocker doses show no significant readmission change with RAS inhibitor dose. Medium beta blocker doses lower readmission risk only at goal RAS inhibitor dose levels.

• 

  Figure 3. MRA vs. RAS inhibitor dose interaction chart with 30-day probabilities MRA, Mineralocorticoid receptor antagonist; RAS, Renin-angiotensin system. Without RAS inhibitors, increasing MRA dose from none to medium raises 30-day readmission risk. At medium or goal RAS inhibitor doses, higher MRA doses reduce readmission risk. At low RAS inhibitor doses, goal MRA dose levels significantly lower readmission risk compared to lower MRA doses.

• 

  Figure 4. MRA vs. Beta blocker interaction chart with 30-day readmission probabilities MRA, Mineralocorticoid receptor antagonist. The MRA x beta blocker interaction chart shows that when MRA dose is 0, predicted probabilities of 30-day readmission are similar across all beta blocker doses. When MRA dose is low, predicted probabilities of 30-day readmission are markedly lower when no beta blockers are given compared to when beta blockers are given at any dose level.

• 

  Figure 5. Odds ratios (ORs) for 90-day readmission across medication classes RAS, Renin-angiotensin system; MRA, mineralocorticoid receptor antagonist; SGLT2i, Sodium-glucose cotransporter 2 inhibitor. This figure shows ORs for 90-day hospital readmission across medication classes (RAS inhibitors, beta blockers, MRA, SGLT2i) and their interactions (RAS × Beta Blocker, RAS × MRA, RAS × SGLT2i, Beta Blocker × MRA, Beta Blocker × SGLT2i). Points represent ORs, with 95% CIs ranging from 0.5 to 2.0. OR<1.0 indicates reduced risk and OR>1.0 indicates increased risk. The line at OR=1.0 shows no effect. RAS × MRA (p=0.07) and Beta blocker × SGLT2i (p=0.08) interactions are near significance.

• 

  Figure 6. Odds ratios (ORs) for one-year mortality across medication classes RAS, Renin-angiotensin system; MRA, mineralocorticoid receptor antagonist; SGLT2i, Sodium-glucose cotransporter 2 inhibitor. This forest plot displays ORs for one-year mortality across various medication classes. Beta blocker dose (p=0.02), MRA dose (p=0.03), and Beta blocker_x_SLGT2i (p<0.01) showed statistically significant reductions in mortality, with OR<1. RAS inhibitor dose, SLGT2i, RAS_x_Beta_blocker, RAS_x_MRA, and RAS_x_SLGT2i showed no significant effect, with ORs close to 1. CIs were narrow, indicating precise estimates.

• 

  Figure 7. Guideline-directed medical therapy composite This bar chart illustrates the average predicted logits for the three clinical outcomes: 30-day readmission, 90-day readmission, and one-year mortality, across nine categories (zero to nine, each uniquely colored). The Y-axis spans average predicted logits from -3.50 to 1.00. For 30-day readmission, logits are near zero, with category zero being slightly positive. For 90-day readmission, categories zero and two show negative logits around -0.50. For one-year mortality, all categories have negative logits, with category seven at -3.50, indicating a strong negative prediction. The chart reveals distinct predictive patterns across the outcomes and categories.