Smart Discharges to improve pediatric post-discharge survival

<br/><strong>Background:</strong> In Sub-Saharan Africa, pediatric post-discharge death is increasingly recognized as an important contributor to mortality. Current studies evaluating interventional approaches for post-discharge mortality focus on pharmacologic therapy, though only malaria prophylaxis post-discharge appears effective. Approaches to reduce vulnerability through health system strengthening approaches may further help to improve outcomes. This study aimed to evaluate the impact of a risk-differentiated approach to improved peri-discharge care on post-discharge mortality among children under 60 months.<br /> <br /><strong>Methods:</strong> We conducted a prospective parallel cluster crossover trial at 6 hospitals in Uganda. Children <60 months admitted due to suspected infectious illness were eligible for enrollment. Phase 1 was a comparative control. During phase 2, enrolled children were screened for post-discharge mortality risk at admission using a multivariable risk algorithm. All children received counselling on post-discharge care practices during admission and at discharge. High-risk children received referrals and automated SMS engagement at 2, 7 and 14 days at a clinic of their choice, or by a community health worker. Survival analysis, adjusting for age, sex, site, period time and predicted risk of mortality was used to estimate the effect of the intervention on 6-month all-cause post-discharge mortality.<br /> <br /><strong>Findings:</strong> 13,050 patients were enrolled (phase 1: n=6954; phase 2: n=6096) and had complete 6-month follow-up. Baseline characteristics were similar between groups. The median age was 0.8 months (IQR: 0.2-1.7), with 56% of participants male. The multivariable risk algorithm gave a mean predicted risk of post-discharge mortality of 6.1% in phase 1 and 5.9% in phase 2. The rate of post-discharge mortality was 6.0% during phase 1 and 4.9% during phase 2, with an adjusted hazard ratio of 0.77 (95% CI – 0.90), favoring the intervention. Additional sensitivity analysis using different sets of covariates in the model showed similar results. <br /> <br /><strong>Ethics Declaration:</strong> These studies were approved by the Mbarara University of Science and Technology (No. 15/10-16), the Uganda National Council for Science and Technology (HS 2207), and the University of British Columbia (H16-02679).<br />