I read with great interest the article by Donovan et al about infant death rate in Ohio in the July issue. I believe that there are some errors in the table. Unadjusted infant mortality rate (deaths/births × 1000) is 3.106 at 38 weeks, 2.317 at 39 weeks, 2.121 at 40 weeks, and 2.997 at 41 weeks. The reference adjusted rate [(269 + 185)/(116,102 + 87,214)] should be 2.233. I don’t have the math skills to be able to calculate their last column of logistic regression.
I would like to see the data manipulated differently. What happens to outcomes when the “plan” is to deliver at a specific gestational age? How do we decide what’s too early and too late for an “elective” delivery at a specific gestational age? A planned delivery at 39, 40, or 41 weeks would have the added mortality of the child who died in the 38th week while awaiting our intervention. There are a number of studies suggesting increasing stillbirth rates weekly at term. What would the data look like if the intent was to deliver infants between 32-41 weeks?
An example by extremes would be if 10% of infants die at 38 weeks but none at 39 weeks. The 39-week data may show the lowest weekly mortality rate, but the right intervention would be delivery at 37 weeks, not 39. The data presented are interesting but not presented in a format that allows decisions to be made to guide management.