Fetal health assessment - expert system

About the author. This page's count says this page has been visited  times since 23rd May 1999. All my calculators will work OFF-LINE if you wish. Instructions for use OFF-LINE You are welcome to copy them over. But check for updates from time to time!

LINKS TO

BMJ article on publishing raw data and real time statistical analysis on e-journals

Demonstation of raw data and real time statistical analysis

Northern Region Fetal/Maternal Medicne Special Interest Group Guidelines
Recent BMJ letter regarding DELAYED CORD CLAMPING


OBGYN.NET
Fetal medicine Unit, St Georges, London        Joseph Woo's Ultrasound       The Fetus.net
Correct application of Bayes theorem in estimating sequnetial conditional risk

Go to OBGYN.net paper on proposal for valid customised charts generation ("Back to the future for Hermanni Boerhaave" published by OBGYN.NET)
In the Hermanni Boerhaave story, I am trying to point out that using a scan measurement only provides a measure of the age of the fetus. Another set of data needs then to be applied to provide an estimate of the date of delivery. The true mean (or modal) interval from LMP to delivery may be closer to 282 days rather than the 280 proposed by Boerhaave. By drawing up the charts in the way I have proposed an estimate for the date of delivery is provided directly by the chart, using the same type of data which was used to generate the chart and from no other data. The precise interval from LMP to delivery becomes irrelevant. Not only this, but the data on which the chart is based is relatively free of error. The date of the scan, the scan measurement and the date of delivery are all known. There is another advantage in this method. Because relatively large numbers of secure data can be obtained, we can look for differences that might exist in sub-populations defined by ethnicity, parity, age, fetal sex and maternal size, all of which have been shown to affect the size of the newborn. This is what I showed at the BMFM society meeting in York, Ref1 with a difference of 4.4 days between the two extreme groups. Adding to the database since has made the differences more convincing. 
Although it is possible to draw up a graphic chart using the methodology I have just described, I am not sure that this is a good way to use the data. For dating, there is no need to have the complete range of measurements throughout pregnancy. There are well recognised times when dating by ultrasound is the most accurate. The data from routine ultrasound services is going to be abundant at these times and therefore most confident at the same gestations as ultrasound should be used to provide an accurate EDD. Data for other gestations is unnecessary for dating.
There is an important difference in generating an ultrasound EDD using this methodology from the traditional ultrasound charts. Taking a BPD of say 40 mm, the data shows us that on average the woman will deliver a healthy normal baby after 150 days. Based on a very large population, which is quite feasible, the confidence interval for this could be very narrow. 
With a traditional chart we can see that a baby with a BPD of 40 mm will on average be 130 days gestation. There will always be a range of gestations for each measurement. ( The traditional centile lines on these charts emphasises this point.) This range will be small if the population is large but it is just not feasible to get large numbers of women who are sufficiently secure in their menstrual data. Within this population there will remain a small number of women whose gestation does not relate as expected to their menstrual period. Finally to provide an estimated date of delivery, the interval from LMP still has to be known and this provides room for further error. 
I believe it is this inherent error in traditional ultrasound dating which convinces the majority of obstetricians and midwives that they should favour an integrated use of ultrasound and menstrual dates. CESDI also recommends this approach. Ref2 Widespread adoption of the Boerhaave method for generating the EDD charts could be expected to speed up the general acceptance of pure ultrasound dating. We have recently shown that is really not possible to accurately date or assess growth using a chart and wheel Ref3 and these will always be necessary until we have either universal computerised maternity systems or programmes built into the ultrasound scanners.
References
1. Hutchon DJR. Customised ultrasound dating charts. British Maternal and Fetal Medicine Society. fourth Annual Conference, University of York. Abstartcs - Journal of Obstetrics and Gynaecology 1999 19:suppl 1;s57
2. Hutchon DJR. Routine ultrasound is the method of choice for dating pregnancy. Br J Obstet Gynaecol 1999:106;616
3. Hutchon DJ, et al. Clinical interpretation of ultrasound biometry for dating and for assessment of fetal growth using a wheel and chart: is it sufficiently accurate? Ultrasound Obstet Gynecol. 1999 Feb;13(2):103-6. Copyright 1999, DJR Hutchon. Restricted use is hereby granted (personal OK) so long as this code is not *directly* sold and the copyright notice is buried somewhere deep in your HTML document.

LINKS