ArticlesClinical signs that predict severe illness in children under age 2 months: a multicentre study
Introduction
An estimated 4 million babies die every year during the first 28 days of life (ie, in the neonatal period); about 75% of them die in the first week of life.1, 2 To achieve Millennium Development Goal 4—to reduce the number of deaths of children under age 5 years by two-thirds between 1990 and 2015—neonatal mortality in countries of low or middle income must be reduced. The main causes of neonatal deaths are infections, prematurity, birth asphyxia, and congenital malformations.1 Most deliveries, especially in low-income countries, still occur in the home, assisted by traditional birth attendants or family members.3 Thus, sick newborn children could have their first encounter with the formal health-care system when they are brought to health workers at first-level health facilities. Improved management of neonates, including identification of infants with potentially life-threatening illnesses warranting referral by these first-level health workers, represents a critical step towards reducing mortality in this vulnerable population.
To obtain information on clinical signs of sepsis in young infants aged under 2 months, WHO did a large multisite study of the clinical features and causes of serious bacterial disease from 1990 to 1992.4, 5, 6, 7 This information contributed to the development of the Integrated Management of Childhood Illness (IMCI) algorithms during the mid-1990s, which standardised the management of sick young infants at first-level health facilities.8, 9, 10
The original IMCI guidelines did not cover the first week of life, when most neonatal mortality occurs.11, 12 Despite the paucity of evidence, several countries have adapted and modified the IMCI algorithms to include the first week of life. Furthermore, the IMCI algorithms currently in use result in high referral rates, including substantial numbers of well children (because of high sensitivity at the cost of low specificity), which could overburden weak health systems in high-mortality settings where the algorithm is needed most, and unnecessarily expose some infants to nosocomial pathogens.13
We designed the multicentre study presented here to assess the performance of simple clinical signs, alone and in combination, when used by first-line health workers to identify those with severe illness requiring hospital admission (excluding jaundice), as judged by an expert paediatrician. Our aim was to provide evidence to support an IMCI referral checklist for sick neonates in the first week of life, and—if possible—to improve the existing guidelines for infants aged 7–59 days.
Section snippets
Patients
Study sites were Dhaka Shishu Hospital in Dhaka, Bangladesh; Hospital del Niño and Hospital Materno-Infantil in La Paz, Bolivia; Komfo Anokye Teaching Hospital in Kumasi, Ghana; Postgraduate Institute for Medical Education and Research and General Hospital, Sector 16 in Chandigarh, India; All India Institute of Medical Sciences and Safdarjung Hospital in Delhi, India; three primary health clinics established for the study by the Department of Paediatrics, Aga Khan University in Karachi,
Results
3177 children aged 0–6 days and 5712 aged 7–59 days were enrolled (table 1). Across sites, the proportion of 0–6-day-old infants who needed admission varied between 13% and 70%, and in the older age-group between 6% and 24%. Case fatality varied between 0% and 12% in the younger group, and between 0% and 7% in the older infants. Overall, about 2% of infants who did not need admission in both age-groups were lost to follow-up, and fewer than 1% in either group deteriorated after being sent home
Discussion
We have developed an algorithm that includes seven signs or symptoms—history of difficulty feeding, movement only when stimulated, temperature below 35·5°C or 37·5°C or more, respiratory rate over 60 breaths per minute, severe chest indrawing, and history of convulsions—to predict the need for hospitalisation in young infants presenting to health facilities, particularly in the first week of life, so that they can receive prompt treatment. This referral decision algorithm for the first week of
References (18)
- et al.
4 million neonatal deaths: when? Where? Why?
Lancet
(2005) - et al.
Hospital-acquired neonatal infections in developing countries
Lancet
(2005) World health report 2005: make every mother and child count
(2005)World health statistics 2006
(2006)- et al.
Bacterial and viral etiology of serious infections in very young Filipino infants
Pediatr Infect Dis J
(1999) - et al.
The aetiology of serious infections in young Gambian infants
Pediatr Infect Dis J
(1999) - et al.
Etiology of pneumonia, sepsis and meningitis in infants younger than three months of age in Ethiopia
Pediatr Infect Dis J
(1999) - et al.
Bacterial and viral etiology of severe infection in children less than three months old in the highlands of Papua New Guinea
Pediatr Infect Dis J
(1999) Integrated approach to child health in developing countries
Lancet
(1999)
Cited by (0)
- ‡
Members listed at end of paper