More pregnant women are using marijuana now than a decade ago, and the rate is particularly high among younger pregnant women, according to a new analysis of survey data published in the Journal of the American Medical Association.
For the analysis, researchers from Columbia University combed through data from the National Survey on Drug Use and Health, focusing on more than 200,000 women from 18 to 44 years old. Among that group, 5.3 percent were pregnant at the time of the survey.
The rate of pregnant women who had used marijuana in the past month rose to 3.85 percent in 2014 from 2.37 percent in 2002. That’s an increase of 62 percent. Pregnant women between 18 and 25 years old had the highest use rate, with 7.47 percent of them using the drug within the past month.
The researchers acknowledged that the overall prevalence of marijuana use during pregnancy is not very high, but said the increase should be studied further. The jump coincides with more states legalizing medical and recreational marijuana.
The risks of using marijuana while pregnant aren’t well understood. Some studies in people and animals have suggested that marijuana exposure during pregnancy might be associated with low birth weight and cognitive problems, but the evidence isn’t clear.
Still, the American College of Obstetricians and Gynecologists recommends that women who are pregnant or are thinking about becoming pregnant avoid marijuana.
About the Author Reprints
I am looking for contact information for Dr Melanie Dreher for a project a group of us is currently working on in Canada pertaining to the legalization of medical cannabis. Can anyone provide me with any info? I have just learned she is no longer Dean of Nursing at Rush University College of Nursing in Chicago.
Excellent news! The DEA’s own judge stated that cannabis is one of the safest therapeutic substances known to mankind. An excellent choice for completely non-toxic anti-emetic medication.
Here is a recent meta-analysis study that shows no increased risk of problems with neo-natal cannabis use:
https://www.ncbi.nlm.nih.gov/pubmed/27607879?utm_source=ZohoCampaigns&utm_campaign=Project+CBD+Newsfeed+%2810-13-2016%29_2016-10-12&utm_medium=email
Maternal Marijuana Use and Adverse Neonatal Outcomes: A Systematic Review and Meta-analysis.
>>Maternal marijuana use during pregnancy is not an independent risk factor for adverse neonatal outcomes after adjusting for confounding factors.
When use of Cannabis by pregnant women is discussed no one seems to be aware of Dr. Melanie Dreher of UMass’ study: http://www.os-extra.cannabisclinicians.org/wp-content/uploads/2015/02/Dreher-Lucas1.pdf
I have another concern. The authors say they used “log-Poisson” regression. They likely mean a generalized linear model. The Poisson distribution has equal variance and mean. If that’s not true, the point estimates are consistent, but the standard errors (that the authors use to construct CI and p-values) depart from coverage probability and therefore aren’t usable. I wish I could verify given code + data! Sure it’s public, but unless I can stand on the original author’s work, it’s just too cost prohibitive in terms of time.
https://en.wikipedia.org/wiki/Poisson_regression#Overdispersion_and_zero_inflation
Almost certainly the above problem is a death knell for the analysis, but no way to check!
Mr. Joseph, did you check to see if the results were independently verifiable using the statistical code and data? I don’t see that basic reproducibility was checked.
I’ll reply to my own question here: SUDAAN [RTI International], version 11.0.1 is the proprietary software the authors of the study use to calculate their statistics. It should be noted that the “log-Poisson regression” aka probably a Generalized Linear Model is freely available in base R, or Python’s statsmodels. Instead of using freely available software to produce these basic statistical results, the authors use proprietary code that can’t be verified. The source code is not obtainable, so there’s no way to verify if the results are actually true other than completely redoing the analysis in open source software.
Let’s not forget a simple fat finger mistake can totally change results and therefore we ought to be skeptical when code + data isn’t made available.