Clinical Conversations

Podcast: Play in new window | Download

Subscribe: RSS

This time Dr. Ali Raja and Joe Elia talk with two authors of a study that found disparate effects on traffic deaths from the legalization of recreational cannabis. The two states under study, Colorado and Washington, were compared, not with each other, but with a composite of states that most closely resembled what Colorado and Washington would be if they hadn’t passed legalization. The states were thus compared against their “doppelgangers.”

Colorado showed an increase of roughly 75 additional traffic deaths per year, while Washington didn’t show any substantial effect. How can this be so, and what are the implications for states with legalization already in place or contemplating it?

Links:

JAMA Internal Medicine study

JAMA Internal Medicine editorial

Running time: 20 minutes

 Transcript:

Joe Elia: You’re listening to Clinical Conversations. I’m your host, Joe Elia.

Earlier this week, JAMA Internal Medicine published two reports and an editorial about the association of legalized recreational marijuana and its possible effect on traffic fatalities.

One report found an increased traffic fatality rate in four states that had legalized recreational use of the drug. The increase was measured relative to a control group of states that hadn’t legalized.

The other study examined the effect in two states, Colorado and Washington, and it found disparate results using a so-called synthetic control comparison.

Two authors of that paper have kindly agreed to discuss their results with me and Dr. Ali Raja, Editor-in-Chief of NEJM Journal Watch Emergency Medicine.

Our guests, Dr. Julian Santaella-Tenorio and Dr. Magdalena Cerda, are both with the Center for Opioid Epidemiology and Policy at New York University School of Medicine. Dr. SantaellaTenorio is also with Universidad del Valle in Cali Colombia. I’d like to welcome you both to Clinical Conversations.

Dr. Ali Raja: We both really enjoyed reading your study. For our listeners, can you give us a thumbnail sketch of your conclusions and findings?

Dr. Julian Santaella-Tenorio: We found that after controlling for all of these factors in generating a synthetic control for each one of the exposed states, which are Colorado and Washington state, we found that traffic fatalities were increasing in the post-law period — that’s from 2014 to 2017 — in Colorado but not in the synthetic control, but we didn’t see that in Washington state.

We just saw that in the post-law period the traffic fatality trajectories were the same for Washington state and for the synthetic control for that state.

Dr. Ali Raja: Were you surprised at the disparity between Washington state and Colorado?

Dr. Julian Santaella-Tenorio: It was surprising to see this kind of different effect. We were thinking that we’re going to see the same effect, right?, like increases in fatalities in both states that were enacting these laws. However, we thought that, after looking at the details, that the industry in Colorado is much bigger and it has many more dispensaries compared to Washington state. If you do that per capita, you see that it is much bigger in Colorado than in Washington state, and the other thing that we were trying to plot is for why this is the case in Colorado but not Washington is that we see that many of the states that are around Colorado don’t have these kind of laws and there could be a lot of kind of this tourism going into Colorado, but that’s not likely the case in Washington state.

Dr. Magdalena Cerda: And you also see, in Colorado, you see an increase in use of marijuana after the opening of legal retail sales and you don’t see the same in Washington. So, that also supports, I think, the focused increase in traffic fatalities in Colorado.

Joe Elia: Dr. Cerda, why did you choose synthetic controls which compared each state individually with a kind of, as I was thinking about it, a doppelganger state? You know, an identical state, but that state hadn’t implemented recreational cannabis laws. Why did you take that approach?

Dr. Magdalena Cerda: Yeah. That’s a great question, and so the reason is that we want to make sure –right?, because of the high policy relevance of this question — we want to make sure we get the answer right and we want to make sure that any kind of increase that we see in traffic fatalities after legalization is really due to the law and not to other things that might be different between Colorado, say, and other states, right?

Because we do know that states don’t randomly choose to legalize or not. In fact, there are many reasons why they do, and those reasons are different between states that choose to legalize or not, and it makes it difficult to figure out if it’s really the effect of the law.

So, the nice thing about synthetic control group approach is that we can use a data-driven procedure to figure out which states basically match what would have happened in Colorado, say, or what would have happened in Washington had the law not been passed, and that’s what we want to know, right? We want to figure out okay, well, what would have happened in Colorado had the law not been passed and what actually happened?

So, is what happened greater? Is their increase greater than what would have happened without the law? Yes or no, and that gives us a much better answer about whether it’s really the law and not something else. So, that’s why we chose to use that approach.

Joe Elia: I noticed that you have a table of the states that you used to synthesize the controls from, and Colorado’s state, I believe, the heaviest burden fell on New Jersey, I think, for Colorado, and for Washington, California was the closest comparison. Can you just give me a sense of why, for instance, California related more to Washington than say New Jersey did?

Dr. Magdalena Cerda: We used a range of different characteristics of these states and we combined them to figure out okay, given the combination of multiple characteristics that we think might be related to why certain states passed this law and why other states didn’t, which states were most similar to, say, Washington when you consider the combination of all of these laws together, and so it ended up that California was the best match for Washington and New Jersey was for Colorado, and so it’s not any one…the really nice thing actually about this approach is that it allows you to figure out, you know, not just the single isolated contribution of any one feature but account for the complexity of features, right? The complexity of factors that lead states to choose to enact this law or not, and it’s really a combination of factors that gives more weight to some states than others when you do the comparison.

Dr. Julian Santaella-Tenorio: And the good thing about this is that you’re not picking states to be the controls for your exposed state, for example, Colorado. The algorithm does it for you and it’s trying to find the best combinations of different states that will give the best match for that exposed state, for Colorado, for example. So, it’s an iteration process that will try to find the best combined group of states that will actually give you the synthetic Colorado that you want to use.

Dr. Ali Raja: Probably, hopefully many of our listeners will go back to JAMA Internal Medicine and read your article, but just because some of them may not have yet, let’s talk a little bit about the extent of the effect in Colorado, which we’ve mentioned a few times. You saw an uptick in motor vehicle deaths. What was your estimate of that uptick?

Dr. Julian Santaella-Tenorio: We saw that there was an average for the four years in the post-law period of 75 excess deaths, and if you look at each specific year you see that for 2014 there’s 37 excess deaths, and if you look into 2015, it’s 63 deaths. In 2016, it’s 78, and finally in 2017, it’s 123 excess deaths of the 648 that were reported by the CDC and by the four years’ data that we used. So, it’s an important proportion of the total deaths that were observed for Colorado in those years.

Joe Elia: Can we talk a little bit about, relative to death, are there offsetting benefits to be found in those 75 deaths? For instance, is there less crime surrounding illegal use of cannabis, or is there less driving under the influence of alcohol if you’ve legalized cannabis? Have these tradeoffs been studied?

Dr. Julian Santaella-Tenorio: So, we did a study back in 2016 trying to identify the effects of medical marijuana laws and traffic fatalities, and we did see that there was an overall reduction in traffic fatalities across these states that were enacting these laws.

The hypothesis that we had, and it was based also in another study by Anderson and collaborators showing the same effect, that perhaps marijuana was reducing alcohol use and was reducing the risk of people driving on roads being impaired by alcohol. However, that might be the case for medical marijuana laws, but might not be the case for recreational laws because it’s a totally different scenario.

You have dispensaries that are selling to anyone over the age limit. You have availability just skyrocket, and that could play different when you talk about like driving under the effects of marijuana and also the risk to driving. So, the thing is that marijuana competing with alcohol with medical marijuana laws could be one part of the study, but once you have such great availability of marijuana due to these new recreational laws, then the substitution effect might not be occurring, and you can have like a lot of people driving under the effects of both marijuana and alcohol, and we have different studies, like randomized controlled studies, showing that when you use both of these substances at the same time the risk for having a severe traffic event increases by a lot.

Dr. Magdalena Cerda: I would like to say though that at the same time, it is true that we just looked at one outcome, which is traffic fatalities, and so this doesn’t in any way mean that there aren’t very, you know, clear benefits of legalization in other areas. In no way does it mean that we are, you know, we’re trying to make a case against legalization. It’s just that we were focusing on this one particular outcome.

Joe Elia: You can be forgiven for defending it. However, I’m a bicyclist and I bike in Boston a lot, and I’ve noticed that I’m smelling a lot more marijuana coming out of car windows than I’m comfortable with smelling. So…

Dr. Ali Raja: As a bicyclist, I’m sure that’s of utmost concern. You’re right. It definitely has picked up.

Joe Elia: Dr. Cerda, what do you think are the next steps in your research?

Dr. Magdalena Cerda: So, I think, you know, one of the really interesting pieces of this paper is the fact that we found different effects in different states, right? Because that means that it’s not just legalization yes or no, it’s really about how legalization happens that probably makes a difference, and so I think one of the key questions that we need to figure out is what types of legalization approaches lead to increases in traffic fatalities and which type of legalization approaches don’t, right?

So, we need to figure out things like, you know, in terms of density of stores, taxation and the impact on pricing, advertising. So, really trying to identify the different aspects of legalization policy and figure out which ones are the ones that are driving or what combination of policies are really driving an increase in fatalities so that we can figure out potentially what are the types of legalization models that can avoid such unintended consequences.

Dr. Ali Raja: You know, on a little bit of a lighter note. All of us in academics went into our specific topic area for different reasons, and so if you imagine yourself at a family gathering for a holiday, and your friends and family ask you well, what do you do and why did you go into that, is there a story, or why do you tell them that you started researching this particular topic?

Dr. Julian Santaella-Tenorio: When I was doing my DrPH studies at Colombia University, I had this methods exam at the qualifying exams, right, and this was one of the questions that we were thrown at this exam, and I remember being really interested in answering this question, and then decided to start putting together databases and methods to be able to answer this question. So, to that I would say, like the first term paper that I had put together for passing this exam, and then continue and developed this nice paper that we published back in 2016, and I wanted to see the other effects of recreational marijuana laws, and I got Magdalena to be convinced that this was a good idea and to join in this effort.

Joe Elia: Dr. Cerda, what do you tell families at gatherings about why you’re doing what you do?

Dr. Magdalena Cerda: Well, you know, as a public health researcher I find this fascinating, and I feel like it’s my responsibility to figure out what the public health implications of these laws are. Particularly, I think what really fascinates me is not so much about what happens with legalization, but rather trying to figure out what types of models of legalization have been adopted across the United States and in different countries in the world to figure out: Are there ways that we can legalize safely? That is, are there ways that we can legalize and minimize public health consequences, and so to figure out and to use our capacity as researchers to figure out, you know, what are the unintended consequences and what do different experiences in different states and in different countries tell us about the best way to do it to protect public health?

Joe Elia: So, you’re both assuming, I think, that people are going to use cannabis, and your interest is in how do you make that possible and safe, as well as enjoyable or…

Your department encompasses both epidemiology and public policy, and so Dr. Cerda, what would you like to see change regarding that?

Dr. Magdalena Cerda: In terms of policy for marijuana, you mean?

Joe Elia: Yes.

Dr. Magdalena Cerda: I think it’s a good question. I think as…well, several things.

One is, as states legalize, I think there needs to be tight regulation around zoning of retail outlet stores, tight regulation and constraints on advertising of the product, investment in a lot of the taxes earned from sales in prevention, particularly in schools and targeted towards adolescents so that we can prevent unintended increases in early initiation of marijuana use.

Also, I think there needs to be use of those dollars for greater investment in treatment for people who are experiencing problems with cannabis use. So, I think it’s, you know, tighter regulation of access to marijuana and advertising, as well as investment in prevention and treatment.

Joe Elia: Okay, and just to be clear about the roles on the paper, Dr. Santaella-Tenorio, you’re the first author?

Dr. Julian Santaella-Tenorio: Yes.

Joe Elia: And Dr. Cerda was the senior author?

Dr. Julian Santaella-Tenorio: Yes.

Dr. Magdalena Cerda: Yes.

Joe Elia: Correct, and Dr. Cerda got the money together?

Dr. Magdalena Cerda: Yeah. Yeah. So, yes.

Joe Elia: Okay.

Dr. Magdalena Cerda: That’s part of what I did.

Joe Elia: All right. I just wanted to make the roles clear.

Dr. Julian Santaella-Tenorio: So, there’s something that I wanted to say, and is that so when we published our study on medical marijuana laws back in 2016, a lot of people complained about these results and it was kind of like we are saying that it’s safe to drive under the influence of marijuana, right? Because we saw reduction in fatalities, and we got a lot of emails and complaints like that, and I just want to point this out and is that it’s not that these findings at the population level when we were looking at states or say anything about the individual risk of marijuana use, right, but people sometimes get confused with that.

So, what we’re showing is not that if you smoke marijuana you have an increased risk of getting into an accident that will kill you. We’re saying is that, on average, states that passed these laws, such as Colorado, we see that at the population level there’s an increase in fatalities, which is totally different from the individual risk.

Joe Elia: Okay, and Dr. Raja, you’re an emergency department physician. Have you seen any effect? Do you see more people coming in with effects of having used recreational cannabis?

Dr. Ali Raja: It’s a good question, and, Dr. Santaella-Tenorio, you’re right. This is a very polarizing issue and people bring their own biases to it, and they take your conclusions and based on what they inherently believe, they’ll see them as either positive or negative, whereas you’re presenting data at a population level.

I’ll say individually, my patients…Massachusetts has always, at least that in the 12 years that I’ve been here, there’s been a lot of cannabis use in Massachusetts. That’s gone up over the past few years because of legalization, but there’s still only a handful, less than 10, maybe even less than five, I haven’t kept up, dispensaries in the state. So, it’s much more similar to Washington than it is to Colorado. When you get off the plane at the airport in the Colorado, there’s basically a dispensary at every corner, whereas in Massachusetts, although it is legal, it is still not widely available, and so I haven’t seen that much of a change in the patients that are actually presenting to my emergency department. Good question though, Joe.

Joe Elia: Well, I want to thank you very much, Dr. Santaella-Tenorio and Dr. Cerda, for sharing your insights with us today.

Dr. Magdalena Cerda: Thank you. Thank you so much for this…

Dr. Julian Santaella-Tenorio: Thank you. Yeah, thank you for inviting us.

Joe Elia: Okay. Thank you.

That was our 268th episode. They are all available free at Podcasts.Jwatch.org. We come to you from the NEJM group, and our executive producer is Kristin Kelley. I’m Joe Elia.

Dr. Ali Raja: And I’m Ali Raja.

Joe Elia: Thanks for listening. Dr. Ali Raja: Thanks for listening.