Study designs of evaluations included in the review
Studies in which some symptom, sign or laboratory test was compared with a reference test were eligible. No other inclusion criteria relating to the study design were specified. One study was excluded because it was a retrospective chart review, while a further 10 were excluded because they had a non-independent comparison of a test with a gold standard of uncertain validity.
Specific interventions included in the review
Studies comparing paediatric dehydration with a symptom, sign or laboratory value were eligible. The included studies evaluated individual clinical signs and symptoms (such as capillary refill time, respiratory pattern and skin turgor), groups of signs or clinical prediction models, and laboratory tests (such as blood urea nitrogen and bicarbonate).
Reference standard test against which the new test was compared
The 'gold' standard chosen for percentage of volume lost was the difference between the rehydration weight and acute weight divided by the rehydration weight. Studies using solely examination signs or a general dehydration assessment were excluded.
Participants included in the review
Studies of dehydration in children aged 1 month to 5 years were eligible for inclusion. The authors stated that studies enrolling 'some children' outside that age range were not excluded, and in 2 studies the age ranged up to adolescence. The included studies were carried out in emergency departments, hospitals and a gastroenteritis clinic in the USA, Australia, Egypt, Kenya, Columbia, Indonesia and Turkey. In most of the studies, the children were admitted to hospital with diarrhoea.
Outcomes assessed in the review
The measures of diagnostic accuracy calculated for each test, for both 5% and 10% dehydration, were the sensitivity, specificity and likelihood ratios (LRs).
How were decisions on the relevance of primary studies made?
Each of the three authors reviewed the titles and abstracts obtained from the searches.