The objective was to assess whether Baby Check could be used in a different cultural setting to where it was developed. Babies were scored using Baby Check and then assessed by a physician who had no knowledge of the score. Out-patient clinics were held at the Child In Need Institute, Calcutta, India. The subjects were 100 babies aged between 0 and 12 months. Ninety-five babies were successfully scored and independently graded by a doctor. The total scores ranged 0 to 53 with a median of 11. Physicians graded two babies as seriously ill that required urgent admission; eight required admission for observation and/or treatment; 51 were thought to require careful observation and treatment at home; and 34 were considered to be mildly ill or well. There were several difficulties using Baby Check in Calcutta. Baby Check is designed to assess the severity of acute systemic illness in babies. Many babies seen had chronic conditions such as malnutrition, as well as an acute illness. There were difficulties translating some of the checks. In the study design there is no 'gold standard' against which to test physicians' grading of illness or the Baby Check score. Despite these problems Baby Check detected the majority of ill babies in the study: with a total score cut-off of 12-13 Baby Check had a positive predictive value of 86 per cent for detecting babies that required medical attention. Of the babies scoring less than 8, 97 per cent did not need admission to hospital.(ABSTRACT TRUNCATED AT 250 WORDS)