The Integrated Management of Childhood Illness (IMCI) approach helps community nurses and medical assistants assess and treat sick children at primary health facilities in poor countries. They use a combination of symptoms, signs and investigations to decide on treatment and referral. Could this approach also be used for initial diagnosis in hospitals?

Nurses and medical assistants often conduct the preliminary assessment of children arriving at hospital, with limited supervision from senior staff and a lack of reliable laboratory services. Researchers from the KEMRI/Wellcome Trust Research Laboratories in Kenya developed an assessment protocol involving replies to eight questions, examination for 12 physical signs, measurement of oxygen levels and microscopy for malaria. These results guide the choice of six different treatment strategies:

• antibiotics
• anti-malarials
• intravenous fluids
• specific nutritional support
• oxygen
• blood transfusion.

They tested the protocol at Kilifi District Hospital in Kenya’s Coast Province and compared the results with final diagnoses by a paediatrician. The study included 3 705 children under 13 years-old who did not have an obvious simple diagnosis, such as sickle cell disease. They found that:

• 63 per cent fit the definition for at least one severe syndrome. Mortality in this group is 7.8 per cent.
• the 1 378 children without a severe syndrome most commonly have a final diagnosis of malaria (48 per cent), pneumonia (13 per cent) or gastroenteritis (19 per cent)
• these children have milder disease, lower mortality (one per cent) and short inpatient stays (average two days)
• mortality is lower among children who need fewer treatments. Two-thirds of all deaths in the severe syndrome group are in those who need two or three treatments
• positive predictive value (the likelihood that a child really has a certain disease if they receive that initial diagnosis) for the IMCI approach ranges from 46 to 70 per cent for severe pneumonia, diarrhoea and malaria
• the IMCI approach picks up at least 96 per cent of children who have these three syndromes as their final diagnosis. But it identifies only 56 per cent of children who have meningitis.

The researchers conclude that this approach may give clearer guidelines for targeting treatment and making decisions about admission. Syndrome definitions should minimise the risk from failure to treat, promote the rational use of scarce resources and prevent unnecessary invasive procedures. But they emphasise that health planners introducing this approach will need to:

• balance the risk of not admitting children who are less severely ill with the improved quality of care resulting from having fewer admissions
• define the syndrome hierarchy – which to tackle first
• focus on a limited number of features to ensure it works under operational conditions
• develop a well-organised record form for the protocol to guide the choice of treatment
• consider adding two additional simple laboratory tests (microscopy of spinal fluid and measurement of blood haemoglobin) to the protocol, which would considerably improve practice.

Source(s):
‘Hypothetical performance of syndrome-based management of acute paediatric admissions of children aged more than 60 days in a Kenyan district hospital’, Bulletin of the World Health Organization 81: 166-173, by M. English et al, 2003 Full document.

Funded by: KEMRI; Wellcome Trust

id21 Research Highlight: 2 June 2003

Further Information:
Mike English
Centre for Geographic Medicine Research
KEMRI/Wellcome Trust Research Laboratories
PO Box 230
Kilifi
Kenya

Contact the contributor: menglish@kilifi.mimcom.net

Kenya Medical Research Institute - Wellcome Trust Collaborative Programme

Other related links:
'Rapid response – diagnosing bacterial meningitis in primary healthcare settings'

'Self-diagnosis of schistosomiasis by Tanzanian schoolchildren'

'Children in developing countries face new health threats' >

See id21's collection of links relevant to maternal and child health.