PMA - Pratiyush Karki

.pdf
School
University of California, Davis **We aren't endorsed by this school
Course
ECI 114
Subject
Statistics
Date
Aug 9, 2023
Pages
4
Uploaded by teiser957 on coursehero.com
ECIV 114 Pratiyush Karki PMA: How accurate is baby's due date? Intro A baby's due date is an occasion that is heavily anticipated by parents. This podcast discusses how accurate the predicted due date actually is. Methods of Calculating Due Date The commonly used method for the calculation of a baby's due date is called the Naegele's Rule. This rule incorporates a formula which involves subtracting 3 months from the last period and then adding 7 days (plus 1 year) to estimate a baby's due date. For example, let's say a woman had her last period on June 21, 2023. The estimated due date would be March 28, 2024. This method has reliably been used by medical professionals all over the world. In conjunction, ultrasound is often used early in pregnancy to calculate the due date. This method is used during the first trimester in which doctors measure the baby from the head to the bottom. This is the
most accurate method to calculate the due date because most babies are the same size during this time of pregnancy. Analysis of Methods However, even utilizing these methods in conjunction still will not guarantee an accurate due date. This is because there are an abundance of factors that affect the due date. Consequently, it's called an "estimated due date" because it's not meant to be precise. Major factors that make a due date inaccurate is if a woman has an irregular period, doesn't remember their period date or has a medical condition that mitigates the due date. This is proven since women who've gone through IVP (In vitro fertilization) generally have more accurate due dates because of the transparency of information. Therefore, the estimated due date is usually not very accurate. The error is estimated to be within 5 days early or late from the estimated due date (04:30). So why do we even estimate a due date if it is going to be significantly inaccurate? Because the birth of a baby isn't the only important event during a pregnancy. Other important events such as prenatal care, interventions and tests can be done using the due date estimation. In the U.S, over 3.5 million babies were born during 2021 in which 10.5% arrived before week 37, around 29% arrived around weeks 37-38, 56% arrived around weeks 39-40, and 4.6% arrived weeks 41 or more (05:30). Interestingly, it is discussed at (07:40) that black infants are twice as likely to be born early than white infants. Consequently, we can infer that race and environment play a factor that impacts the estimated due date. Also we can infer that the fetus, especially the size of the baby has a major input on the due date. Another spontaneous statistic discussed was that mother's under 5 '3 or older than 35 years were more likely to have a baby early. Possible Solutions to Limitations In order to mitigate the limitations of predicting a baby's due date using the stated methods, a multitude of various methods should be used instead of just one. Furthermore, analysis of the mother should be done before in order to factor in elements such as medical conditions, irregular periods or genetics as discussed in the article by Pristyncare. However, a precise prediction is basically impossible; therefore, a close interval range is the goal. Another method that hasn't bloomed yet is machine learning. In the research article it's stated that: "...application of machine 2
learning methods has shown promising results in efficient due date prediction based on ultrasound data, and artificial neural networks have demonstrated high accuracy in predicting due dates." So the combination of computerized models that utilize data of a population and ultrasound (average length of baby in 1st trimester) will be the most accurate in the future. To conclude, we need all the information to accurately predict a due date. Application of Methods to Other Scenarios How can we apply these methods to other scenarios in the real world? Using machine learning as an example, we can create a model that uses a sample of average length of a baby during the first trimester of a pregnancy along with ultrasound to most accurately predict a due date. This is discussed around (08:20). We can transfer this on to another real life scenario such as using machine learning models to predict for example traffic forecasts. When you use Google Maps or Apple Maps it gives you an estimated time of arrival (eta). This estimation comes from historical traffic data (regression models) extracted from satellites as discussed by Javatpoint. Furthermore, the model then uses data from your chosen trip and levels of traffic to predict the best route according to the model. This is beneficial to society because it can optimize transportation systems, improve road safety and reduce the congestion of traffic. 3
Page1of 4
Uploaded by teiser957 on coursehero.com