Interpret the data, taking into account the possible effects of:
Were the denominators appropriate? What other rates would make sense? See model answer
Are there any other reproductive outcomes or reproductive health issues of interest? See model answer
The sample sizes are modest, too small for accurate risk estimates.
Back to top
Self reported outcomes such as abortion and birth defects may suffer from recall bias and require confirmation from medical charts. The easiest data to validate will be birth weight from birth certificates, hospital records.
Back to top
A surrogate for exposure might include length of residence, distance from the canal or distance from the swale, (a swale is a natural pathway of water flow). Confounders to be considered include SES, occupation, genetics, parity, alcohol, medications, other illnesses such as diabetes etc.
Back to top
The commonly used denominator is the # of live births, one might argue that the number of women is a better denominator.
Back to top
One could do a more in depth study of birth outcomes, including fecundity or fertility. These researchers studied self reported birth defects.
Back to top