All About Due Dates

What About Ultrasounds?

An ultrasound can be used to determine your due date if your is last menstrual period is unreliable.  Ultrasound due dates are found by obtaining ultrasound images and taking measurements of the baby.  The underlying assumption of an ultrasound due date is that there is a statistical correlation between certain ultrasound measurements of the baby and it's gestation.  The term "gestation" means how far along is the pregnancy.  If we obtain ultrasound measurements in thousands of pregnancies where we know the exact gestation, we can find the average measurement for any given gestation.  For example, it is known that for a measurement known as the crown-rump length (CRL), the average is 31 millimeters at 70 days (10 weeks) gestation. 

An ultrasound due date reverses this observation to conclude that if a baby's CRL is 31 millimeters, then the pregnancy must be 70 days gestation.  If you know the current gestation, you can calculate the due date by subtracting the current gestation from the average length of a pregnancy and adding the result to the date of the ultrasound.  For example, if you did an ultrasound on a pregnant woman on June 23, 2002 and found a CRL of 31 millimeters, you would conclude that the pregnancy was 70 days gestation.  If you subtract 70 days from 280 days, you get 210 days.  If you add 210 days to June 23, 2002 (the date of the ultrasound) you get a due date of January 19, 2003.

Unfortunately , people tend to give a lot more credibility to ultrasound due dates than is warranted.  That is, many people assume that if an ultrasound due date is different from their due date based on a last menstrual period, that the ultrasound date is more accurate.  However, this isn't necessarily true.  Remember that the basic assumption behind an ultrasound due date is that a given measurement corresponds to a given gestation.  To put this into the proper perspective, imagine that you're told to figure out when a young child will turn 10 years old and are given the following tools:  a tape measure and a pediatrician's growth chart for children.  You measure the subject child's height, and consult the growth chart.  From the growth chart, you find that children are, on average, the same height as the subject child when they are 5 years and 4 months of age.  From this evidence, you conclude that the subject child is currently 5 years and 4 months old. Using simple math, you then calculate that the subject child will turn 10 years old in 4 years and 8 months. 

Only the most gullible individuals would be willing to believe that they can actually predict when a young child will turn age 10 with any degree of accuracy using a tape measure and a growth chart.  The reason for this is fairly obvious.  If you lined a dozen five-year-olds against a wall, they would not all be the same height, even if they all had the same date of birth.  Five-year-olds come in many different heights because children grow at different rates.  This is called normal biological variation.  This normal biological variation introduces a potential for error to the tape measure method for calculating when a young child will turn age 10 because a particular child might be taller or shorter than average.  If a child is taller than average, the tape measure method would over-estimate the child's current age.  If the current age is wrong, then the estimate of when the child will turn 10 years old will also be wrong.

Like older children, babies in the uterus also grow at different rates.  Since ultrasound due dates are based on averages, all ultrasound due dates have a built-in potential for error because a particular baby might be larger or smaller than average.  The amount of possible error depends on the amount of biological variation.  The greater the biological variation, the greater the potential for error.  It turns out that babies in the uterus have the least amount of biological variation in early pregnancy and the greatest amount of biological variation in late pregnancy.  Therefore, the amount of potential error in an ultrasound due date is smallest in early pregnancy and greatest in late pregnancy.  In fact, an ultrasound due date based on a crown-rump length obtained at less than 10 weeks gestation has a built-in error of ± 5 days.  On the other end of the scale, the range of error for an ultrasound due date based on a combination of 4 different measurements obtained at 36 weeks gestation is ± 21 days!!  The statistical term for the built-in error of an ultrasound due date is it's "Standard Error of Mean," or "SEM."

If you have an ultrasound and there is only a 3 day difference between the ultrasound due date and a due date based on a last menstrual period, there is no medical justification for changing your due date to the ultrasound date if the built-in error for the ultrasound date is only ± 5 days.  Statistically, these dates are identical since the 3 day difference is less than the 5 day Standard Error of Mean.  It only makes sense to change a due date to an ultrasound due date if the difference between the due date based on a last menstrual period and the ultrasound due date exceeds the built-in error for the ultrasound due date. 

Many care providers use the "7-14-21" rule for changing a due date based on a last menstrual period to an ultrasound due date. The 7-14-21 rule means that you change the due date to the ultrasound date only if the difference between the ultrasound date and due date based on a last menstrual period is more than 7 days for first trimester ultrasounds, 14 days for second trimester ultrasounds and 21 days for third trimester ultrasounds.  A trimester is about one-third of a pregnancy, or about 13 weeks.  In close cases, care providers are more inclined to accept the ultrasound date if the last menstrual period is unreliable.  As a general rule, if there are two ultrasounds with two different ultrasound dates (a very common occurrence), the earliest ultrasound date is considered more reliable.  If you are uncertain about what date is being used for your due date, discuss it with your care provider.  Your obstetrical care provider should be willing to explain what date is being used for your due date and why that date is being used.

The real "take home" message is DO NOT ASSUME THAT YOUR DUE DATE HAS BEEN CHANGED JUST BECAUSE AN ULTRASOUND COMES UP WITH A DIFFERENT DUE DATE!  Most care providers will only change a due date to an ultrasound due date if there is a "significant" difference between a due date based on a last menstrual period and the ultrasound-based due date, or if the last menstrual period is less reliable than the ultrasound date.  A "significant" difference between a due date based on a last menstrual period and an ultrasound-based due date is a difference that exceeds the "Standard Error of Mean," or built-in error for the ultrasound date.