This study reviewed hospital and genetics clinic records of 411 patients evaluated in our department from 1986 to 1997 inclusive. Major objectives were to establish how often and under what circumstances a specific genetic/syndrome diagnosis was made and to determine the value of laboratory tests in the hope of gaining a more selective approach to referral, evaluation, and use of the laboratory. A specific genetic/syndrome diagnosis was made in 19.9% of cases, and in a further 4.4% the referring diagnosis was eliminated but no new diagnosis made. There was a significant excess of affected males (277:134) and of affected male sib pairs over expectation, suggesting an additional, potentially important, contribution from nonspecific X-linked mental retardation (MR). Factors associated with making a diagnosis included referral from a pediatrician or neurologist, absence of cerebral palsy, presence of more than three minor anomalies and/or an unusual appearance, a recognizable Gestalt or key anomaly. There was a linear relationship between the likelihood of making a diagnosis and the number of minor anomalies. Factors not associated with making a diagnosis included the year when the patient was seen, degree of MR, number of prior specialists seen, presence of a major malformation, occurrence of seizures, and a head circumference either <3rd or >97th centile. Although chromosome studies were somewhat less likely to be ordered in patients with less severe MR, the positive rate was unaffected by the severity of the MR. The rate of abnormal results was positively correlated with the presence of minor anomalies and/or an unusual appearance. None of 134 studies carried out on patients with β€3 minor anomalies alone were positive. Fragile X studies were less likely to be ordered with increasing levels of MR and 10 of 14 positive test results were among those with mild delay. Thirteen of 14 diagnoses were based on Gestalt. Molecular and fluorescence in situ hybridization studies had a positive rate of >60% when ordered by a clinical geneticist compared with 0% when ordered by other physicians. Results showed that use of the laboratory was inconsistent and not clearly based on the findings in a particular child. Significant changes in patterns of referral and the evaluation process could be made that would result in significant economies of time and laboratory use and a minimum level of missed diagnoses. Am.