Effectiveness of pneumococcal vaccine in preventing hospital admissions for pneumonia among the elderly

Dissertation (MSc (Clinical Epidemiology)--University of Pretoria, 2019.Introduction Hospitalisations and prolonged hospital stays impose great economic burden especially at the present time when resources are limited. Community-acquired pneumonia (CAP) is a common and costly illness associated with considerable morbidity and mortality. Other than children, the elderly are the most vulnerable to CAP due to reduced immunity and comorbid chronic conditions. Streptococcus pneumoniae (S. pneumoniae) has been identified as the most common culprit encountered in cases of CAP with the incidence of CAP peaking during the annual influenza season. There is a known synergistic pathogenesis between the influenza virus and S. pneumoniae. Vaccination against invasive pneumococcal disease (IPD) is established in children. However, the burden of pneumonia has remained high in the elderly. This study sought to explore primarily the effectiveness of pneumococcal vaccination, as administered in a South African Medical scheme population, in the elderly who are aged 65 years and older; and secondly to explore the effectiveness of influenza vaccination in the same age group in preventing hospital admissions due to pneumonia (all causes). Methods The study population consisted of 34 068 beneficiaries of Medihelp medical aid scheme, and the outcome measures were investigated for years 2017/2018. The researcher has conducted a case-control study using cross-sectional secondary data with 1:1 matching. The sample size consisted of 800 pairs of case and control for primary and secondary exposures (pneumococcal vaccine and influenza vaccine, respectively). ICD-10 (International classification of diseases .10th revision) coding was used to identify study cases based on hospital admission claims and was matched for covariates age, sex and important comorbidities: ischaemic heart disease (IHD), chronic obstructive pulmonary disease (COPD), asthma and diabetes mellitus (DM). For the primary outcome, we adjusted for the use of influenza vaccine and for the secondary outcome, we adjusted for the use of pneumococcal vaccine. McNemar’s odds ratio (OR) and its 95% confidence interval (CI) was used to measure the association between vaccination and hospitalisation for CAP. Sensitivity analyses by means of propensity score matching (PSM) were also performed to estimate the OR. In addition, subgroup analyses were performed by estimating the odds ratios in participants who have used 23-valent pneumococcal polysaccharide vaccine (PPSV-23) and 13-valent pneumococcal conjugate vaccine (PCV-13) respectively for the primary exposure by PSM. Results All participants had claimed only one type of pneumococcal vaccine in this study. Vaccine uptake for pneumococcal vaccine and influenza vaccine in the study population were 0.9% and 16.6% respectively. For the primary exposure, 15 (1.9%) cases were exposed to pneumococcal vaccine compared to nine (1.1%) in controls with an OR of 1.67 (95% confidence interval (CI), 0.683 - 4.319) (P= 0.308). Propensity score matching revealed similar estimates, although closer to the null value, with OR of 1.05 (95% CI, 0.991 - 1.121) (P= 0.095). For the secondary exposure, 140 (17.5%) cases were exposed to influenza vaccination compared to 152 (19.0%) controls with an OR of 0.90 (95% CI, 0.683 - 1.178) with a (P= 0.460). Using PSM the OR was 0.99 (95% CI, 0.983 - 0.994) (P