OBJECTIVES : In response to wide inter-state variability and lack of national guidance surrounding the use of newborn screening (NBS) tests, the American College of Medical Genetics (ACMG) was commissioned to develop a uniform panel of conditions for NBS. ACMG surveyed a group of subject matter experts to determine a recommendation based on responses and an ACMG algorithm. ACMG ignored uncertainty due to missing survey responses. This research demonstrates how missing survey data may impact the ACMG screening recommendations; one of the 84 conditions, CPTII, is used as an example.

METHODS : Data were available for two survey questions. Missing responses for each question were determined based on the total expected responses for the CPTII condition. Next, upper and lower boundary estimates were calculated by substituting the highest and lowest possible response scores for the missing data. Using the established upper and lower bounds, the bound width for each question was determined. Finally, the bound widths for both questions were combined to evaluate their collective impact.

RESULTS : For CPTII, the two questions relate to Incidence of the condition (3 missing data points) and Prevention of Mortality (2 missing data points). Upper and lower bound estimates generated a bound width of 6.67 points for the incidence question and 4.44 points for the prevention of mortality question. With a combined bound width of 11.11 points, the range for the total score was 1166 – 1178 compared to the original total score of 1169.

CONCLUSIONS : The overall score shift from these two questions alone was not large enough to alter ACMG screening recommendations (critical value ≥1200). However, if data were available for additional questions or if other conditions were considered, findings may have been altered. These results suggest that missing data can influence conclusions and underscore the importance of carefully evaluating ACMG screening recommendations.