Date of Award

2024

Degree Type

Dissertation

Degree Name

Master of Medicine (MMed)

First Advisor

SAMUEL NGUKU

Second Advisor

BERNADETTE MUTHEE

Third Advisor

KEVIN OMBATI

Department

Imaging and Diagnostic Radiology (East Africa)

Abstract

Introduction; Pneumonia causes significant morbidity and mortality in children globally, emphasizing the need for accurate diagnosis. While chest radiographs have traditionally been the gold standard for diagnosis, recent international guidelines discourage their routine use in paediatric community acquired pneumonia (pCAP). Despite this, chest radiographs are still regularly acquired for pneumonia diagnosis, resulting in practice variations and imaging overutilization. This study aims to evaluate the diagnostic utility of chest radiographs in children with clinically diagnosed pneumonia as per the Kenya Basic Paediatric Protocol (KBPP).

Materials and methods: Data was retrospectively reviewed of 159 paediatric patients at Aga Khan University Hospital, Nairobi (AKUHN) with chest radiographs performed for clinical diagnosis of pneumonia and compared with clinical diagnosis of pneumonia as per KBPP. Statistical performance measures of chest radiography were calculated through the sensitivity, specificity, negative predictive value, positive predictive value, and accuracy presented as frequencies and percentages. Univariate analysis was used to compare the clinical parameters with the outcomes of chest radiograph pneumonia. The interobserver agreement was computed based on Kappa Statistic.

Results: Using the KBPP as the reference standard, chest radiographs showing primary end point pneumonia and/or pleural effusion demonstrated 40% sensitivity, 65.4% specificity, positive predictive value (PPV) of 67.6%, negative predictive value (NPV) of 37.7 % and accuracy of 56.6% in the diagnosis of pneumonia. No individual symptom or physical examination finding was strongly associated with radiographic diagnosis of pneumonia. A fair inter-observer agreement was observed upon comparison of the predominant chest radiographic finding, Cohen’s Kappa of 0. 223(p < 0.001).

Conclusion: Our study supports the current guidelines suggesting that chest radiographs should not be routinely acquired in non-severe cases in view of the low to moderate diagnostic accuracy compared to KBPP. The low to moderate efficacy of chest radiographs in diagnosing pneumonia may result in under and over diagnosis of pneumonia.

Download

Share

COinS