Use of administrative records to assess pneumococcal conjugate vaccine impact on pediatric meningitis and pneumonia hospitalizations in Rwanda

Abstract

BackgroundOngoing surveillance is critical to assessing pneumococcal conjugate vaccine (PCV) impact over time. However, robust prospective studies are difficult to implement in resource-poor settings. We evaluated retrospective use of routinely collected data to estimate PCV impact in Rwanda. MethodsWe collected data from admission registers at five district hospitals on children age <5yearsadmitted for suspected meningitis and pneumonia during 2002–2012. We obtained clinical and laboratory data on meningitis from sentinel surveillance at the national reference hospital in Kigali. We developed multivariable logistic regression models to estimate PCV effectiveness (VE) against severe pneumonia and probable bacterial meningitis and Poisson models to estimate absolute rate reductions. Haemophilus influenzae type b vaccine was introduced in January 2002, PCV7 in April 2009 and PCV13 in August 2011. ResultsAt the district hospitals, the severe pneumonia and suspected meningitis hospitalization rates decreased by 70/100,000 and 11/100,000 children for 2012 compared to baseline, respectively. VE against severe pneumonia calculated from logistic regression was 54% (95% CI 42–63%).In Kigali, from 2002 to 2012, annual suspected meningitis cases decreased from 170 pre-PCV7 to 40 post-PCV13 and confirmed pneumococcal meningitis cases from 7 to 0. VE against probable bacterial meningitis was 42% (95% CI −4% to 68%). ConclusionIn a resource-poor African setting, analysis of district hospital admission logbooks and routine sentinel surveillance data produced results consistent with more sophisticated impact studies conducted elsewhere. Our findings support applying this methodology in other settings and confirm the benefits of PCV in Rwanda.