As clinical informatics systems incorporate broader domains of clinical data, it is important to assess the quality of these data for routine assessment of perinatal services.

Hypothesis: Compared to the computerized database (PC-Log), there is no difference in data elements obtained by hand abstraction.

Methods: Random selection of 99 charts from a total of 893 births at a tertiary perinatal center during 1995. Of 310 fields in the database, 32 variables were compared to a hand abstraction of the medical charts.

Results: PC-Log had 100% positive predictive value (PPV) for eclampsia, prolonged rupture of membranes, pre-existing diabetes, fetal distress, cesarean section, and neonatal transports. The sensitivity, specificity,and PPV for other variables (table) showed moderate agreement.

Table 1
Full size table

Compared to hand abstraction, PC-Log had no recorded cases of identified substance abuse, antenatal steroids, hyaline membrane disease, circumcision, maternal and infant length of stay. The sensitivity, specificity, and PPV for maternal smoking were 0.64, 0.95, and 0.80; but the correlation for number of cigarettes smoked was r=0.697. Means for birthweight and five minute Apgar scores did not differ, but the correlations were r=0.982 and r=0.436.

Conclusions The PC-Log showed good agreement for many but not all variables of clinical interest. Cost control, consolidation, integration, accountability and quality improvement efforts at all levels of the health care system will demand improved quality and utilization of the computerized database.