Evaluation of a simple scoring system for use of the one-hour glucose tolerance test

BACKGROUND: Recently, Naylor and colleagues proposed a simple numeric scoring system by which women could be selectively excluded from having a one-hour glucose tolerance test (GTT) based solely on their age, race and body mass index (BMI) (NEJM, 1997; 337:1591-6). This study concluded that use of the scoring system could exclude one-third of all women from screening while improving the sensitivity and specificity of the test as a screening instrument. The following analysis applies the scoring system to a heterogeneous sample of postpartum women, fifty percent of whom are Hispanic. METHODS: All women delivering at a large university hospital over a 14-week period (N=646) had their medical records reviewed and were interviewed by bilingual researchers. Collected data included information regarding demographics, prenatal care, delivery and birth outcomes. One-hour GTT levels greater than 140mg/dL were considered abnormal. Comparisons were made using the Chi-square test. RESULTS: Full data regarding one-hour GTT testing, BMI, race and age were available for 477 (74%) of patients. Of these, 56 (12%) would have been excluded from testing under the scoring criteria. Of the persons in the exclusion group, 7 (12%) had an abnormal one-hour GTT compared to 89 (19%) of those in the screening group (p=0.20). CONCLUSIONS: These results suggest that among this heterogeneous population, the proposed scoring system may be effective in excluding low risk women from screening with the one-hour GTT. The scoring system may lack clinical significance given the small difference between groups and the difficulties involved in changing the previously established standard of universal screening.