DESIGN OF A CLINICAL COMPUTER DATABASE FOR A NEONATAL INTENSIVE CARE UNIT |[lpar]|NICU|[rpar]|

Hospitalization in a NICU averages 1 mo. but may extend to 8-9 mos. Voluminous amounts of data makes it difficult to systematically study disease processes with the conventional medical record. A computer system is being developed to aid in the analysis of the incidence, course, interrelationships and demographic features of the various newborn illnesses seen in a referral NICU averaging 600 admissions/year. How does this computer system provide information about newborn problems which would otherwise be difficult to acquire? Each infant's database is divided into 4 categories: 1) Admission history and physical exam 2) In-hospltal clinical events 3) Laboratory data 4) Outpatient follow-up. Complete information from each category is entered into separate computer files. Information is assessed via bit-mapped (inverted) files which allow acquisition of single or multiple pieces of data common to some subset of the total of patients. Laboratory data collected on a daily basis is available for analysis as an interface with clinical data. These file structures favor rapid searches of the entire database but will still permit display of all or part of an individual patient's complex hospital course. The major advantage of this system is that it permits the user to explore rapidly a large number of possibly significant relationships in a large population of newborn infants.