Objectives: This study aimed to describe the population of people who acquired ventilator acquired pneumonia and determine the feasibility of a larger scale study to assess the degree to which bundle compliance reduces or even eliminates, the risk of ventilator acquired pneumonia.
Research methodology/design: A retrospective matched case note review was conducted to scrutinise 10 VAP cases. Cases were matched with two controls for age, gender, APACHE score and number of ventilated days. Compliance with the VAP bundle was determined by extracting data on compliance from case notes. Binary logistic regression was used to calculate odds ratios with confidence intervals which were utilised to determine numbers needed for a larger study.
Setting: A general intensive care unit within a 750 bedded district general hospital, serving a population of approximately 270,000 people in Scotland.
Main outcome measure: The outcome variable of interest was ventilator acquired pneumonia and the independent variable was ventilator acquired pneumonia bundle compliance.
Results: Binary logistic regression suggested that cases which did not receive the bundle reliably were more likely to develop ventilator acquired pneumonia (OR 1.33, confidence interval (CI) 0.28-6.30). Statistical results should be interpreted with caution due to the small sample size, which is demonstrated with the wide ranging confidence intervals (CIs).
Conclusion: Wide confidence intervals enable only a cursory impression as to numbers that would be required for a full scale trial. Nonetheless, the effect size indicated in this paper contributes towards consideration as to numbers needed for future studies.
Copyright © 2012 Elsevier Ltd. All rights reserved.