Effect of Real-time Patient-Centered Education Bundle on Administration of Venous Thromboembolism Prevention in Hospitalized Patients

IMPORTANCE Numerous interventions have improved prescription of venous thromboembolism (VTE) prophylaxis; however, many prescribed doses are not administered to hospitalized patients, primarily owing to patient refusal. OBJECTIVE To evaluate a real-time, targeted, patient-centered education bundle intervention to reduce nonadministration of VTE prophylaxis. DESIGN, SETTING, AND PARTICIPANTS This nonrandomized controlled, preintervention-postintervention comparison trial included 19 652 patient visits on 16 units at The Johns Hopkins Hospital, Baltimore, Maryland, from April 1 through December 31, 2015. Data analysis was performed from June 1, 2016, through November 30, 2017, on an intention-to-treat basis. INTERVENTIONS Patients on 4 intervention units received a patient-centered education bundle if a dose of VTE prophylaxis medication was not administered. Patients on 12 control units received no intervention. MAIN OUTCOMES AND MEASURES Conditional odds of nonadministration of doses of VTE prophylaxis (primary outcome) before and after the intervention on control vs intervention units. Reasons for nonadministration (ie, patient refusal and other) and VTE event rates (secondary outcomes) were compared. RESULTS A total of 19 652 patient visits where at least 1 dose of VTE prophylaxis was prescribed were included (51.7% men; mean [SD] age, 55.6 [17.1] years). Preintervention and postintervention groups were relatively similar in age, sex, race, and medical or surgery unit. From the preintervention period to the postintervention period, on intervention units, the conditional odds of VTE prophylaxis nonadministration declined significantly (9.1% [95% CI, 52%-16.2%] vs 5.6% [95% CI, 3.1%-9.9%]; odds ratio [OR], 0.57; 95% CI, 0.48-0.67) compared with no change on control units (13.6% [95% CI, 9.8%-18.7%] vs 13.3% [95% CI, 9.6%-18.5%]; OR, 0.98; 95% CI, 0.91-1.07; P < .001 for interaction). The conditional odds of nonadministration owing to patient refusal decreased significantly on intervention units (5.9% [95% CI, 2.6%-13.6%] vs 3.4% [95% CI, 1.5%-7.8%]; OR, 0.53; 95% CI,0.430.65) compared with no change on control units (8.7% [95% CI, 5.4%-14.0%] vs 8.5% [95% CI, 5.3%-13.8%]; OR, 0.98; 95% CI, 0.89-1.08; P < .001 for interaction). On intervention units, the conditional odds of nonadministration owing to reasons other than patient refusal decreased (2.3% [95% CI, 1.5%-3.4%] vs 1.7% [95% CI, 1.1%-2.6%]; OR, 0.74; 95% CI, 0.58-0.94), with no change on control units (34% [95% CI, 2.7%-4.4%] vs 3.3% [95% CI, 2.6%4.2%]; OR, 0.98; 95% CI, 0.87-1.10; P = .04 for interaction). No differential effect occurred on medical vs surgical units (OR, 0.86; 95% CI, 0.60-1.23; P = .41 for interaction). There was no statistical difference in the proportion of VTE events among patients on intervention vs control units (0.30% vs 0.18%; OR, 0.60; 95% CI, 0.16-2.23). CONCLUSIONS AND RELEVANCE In this study, a targeted patient-centered education bundle significantly reduced nonadministration of pharmacologic VTE prophylaxis in hospitalized patients. This novel strategy improves health care quality by leveraging electronic data to target interventions in real time for at-risk patients.