Evaluation of four maternal smoking questions.

Abstract

The authors evaluated four questions about maternal smoking during pregnancy for use on birth certificates. Question 1 (yes/no format) and Question 2 (trimester-specific design) were tested among 1171 women who delivered at two Kaiser Permanente medical centers in northern California. Responses to Questions 1 and 2 were compared with smoking information provided by participants in telephone interviews conducted during pregnancy. Question 3 (multiple choice format) and Question 4 (month- and grouped month-specific design) were tested among 900 women who enrolled in a statewide prenatal screening program and who delivered in 20 hospitals in four Central Valley counties. Responses to Questions 3 and 4 were compared with mid-pregnancy serum cotinine levels. The authors evaluated the four questions in terms of conciseness, response rate, data accuracy, and type of data requested. Questions 1 and 2 were the most concise. Response rates could not be calculated for Questions 1 and 2. Response rates were 86.0% for Question 3 and 74.2% for Question 4. Sensitivity was 47.3% for Question 1, 62.1% for Question 2, 83.8% for Question 3, and 86.7% for Question 4. The types of data requested by Questions 2 and 4 seem to best satisfy the needs of the broad audience of birth certificate users. No single question was clearly superior. The authors propose a combination of Questions 2 and 4, which asks about average number of cigarettes smoked per day in the three months before pregnancy and in each trimester of pregnancy.