Ordering Laboratory Tests

To the Editor.—  Of the reasons cited by Dr Lundberg for ordering laboratory tests (1983;249:639), perhaps the least supportable and of least potential benefit to the patient is habit.Before September 1982, our laboratory, supporting a 160-bed, tertiary care, pediatric teaching hospital, performed an average of four urine chloride determinations daily. While fluid and electrolyte derangements are common in our patient population, disorders in which urine chloride analysis is said to be helpful^1,2 are not common and could not account for the number ordered. This phenomenon has been previously documented in a study of 47 requests for urine chloride determinations, of which only one was clinically indicated.³ Despite my addressing this discrepancy informally among the clinical house staff, no reduction was achieved.Most house staff wrote orders for "urine lytes," a phrase translated by the nurses and ward clerks (who actually submitted the laboratory requests) to mean the