Practical quality improvement changes for a low-resourced pediatric unit

Background This work describes a sustainable and replicable initiative to optimize multi-disciplinary care and uptake of clinical best practices for patients in a pediatric intensive care unit in Low/Middle Income Countries and to understand the various factors that may play a role in the reduction in child mortality seen after implementation of the Quality Improvement Initiative. Methods This was a longitudinal assessment of a quality improvement program with the primary outcome of intubated pediatric patient mortality. The program was assessed 36 months following implementation of the quality improvement intervention using a t-test with linear regression to control for co-variates. An Impact Pathway model was developed to describe potential pathways for improvement, and context was added with an exploratory analysis of adoption of the intervention and locally initiated interventions. Results 147 patients were included in the sustainability cohort. Comparing the initial post-implementation cohort to the sustainability cohort, the overall PICU unexpected extubations per 100 days mechanical ventilation decreased significantly from baseline (6.98) to the first year post intervention (3.52; p < 0.008) but plateaued without further significant decrease in the final cohort (3.0; p = 0.73), whereas the mortality decreased from 22.4 (std 0.42) to 9.5% (std 0.29): p value: 0.002 (confidence intervals: 0.05;0.21). The regression model that examined age, sex, diagnosis and severity of illness (via aggregate Pediatric Risk of Mortality (PRISM) scores between epochs) yielded an adjusted R-squared (adjusting for the number of predictors) value of 0.046, indicating that approximately 4.6% of the variance in mortality was explained by the predictors included in the model. The overall significance of the regression model was supported by an F-statistic of 3.198 (p = 0.00828). age, weight, diagnosis, and severity of illness. 15 new and locally driven quality practices were observed in the PICU compared to the initial post-implementation time period. The Impact Pathway model suggested multiple unique potential pathways connecting the improved patient outcomes with the intervention components. Conclusion Sustained improvements were seen in the care of intubated pediatric patients. While some of this improvement may be attributable to the intervention, it appears likely that the change is multifactorial, as evidenced by a significant number of new quality improvement projects initiated by the local clinical team. Although currently limited by available data, the use of Driver Diagram and Impact Pathway models demonstrates several proposed causal pathways and holds potential for further elucidating the complex dynamics underlying such improvements.