Stephen J.LurieMD, PhD, Senior Editor
To the Editor: In their meta-analysis, Dr Bhutta and colleagues1 found cognitive and behavioral deficits among school-age children who were born preterm. However, we have concerns about the scoring system used to assess quality of the studies included in the meta-analysis. The 10-point score had 6 items (Table 1) that appear to take on values of either 0, 1, or 2, yielding a possible range of 0 to 12 for the scale. However, the label "NA," which appears in 3 of 18 cells and is defined as "data not available," presumably means "not applicable." Therefore, 2 of the 6 items ("Demographic" and "Socioeconomic") would score a maximum of 1, and "Study design" would never require a score of 0. The relative weightings for the items are not explained or justified. It was also unclear how half-points were scored, but presumably such scoring reflects the averaging of 2 scores. When we applied this scoring system to one of our own studies2 we calculated a score of 9 of 10, instead of the 7 of 10 assigned by Bhutta et al.
Doyle LW, Saigal S, Streiner DL. Prematurity and Later Cognitive Outcomes. JAMA. 2002;288(20):2542. doi:10.1001/jama.288.20.2542-JLT1127-4-1