Talk:Marijuana

Part of m conversation with Gemini AI, in case I lose the chat:

Let's go with: 'High' Risk: Weed

And let's sketch out the choices for the main page. I'm thinking: a dropdown with choices for the common ways of consuming: Smoking/Vaping/Edibles. A dropdown for the frequency of consumption (what are the common ways of measuring that in the literature?). And maybe a dropdown for dosage, which would only appear if edibles. Am I missing any inputs that would matter? I think instead of asking "how long" we should just have a long-term risk model that tells the user "if you keep smoking every day for 20 years research says you have a 1 in XXX chance of developing lung cancer, a 1 in YYY chance of ... etc". What do you think?

Your Proposed Inputs (and my thoughts) Method of Consumption (Dropdown): Smoking | Vaping | Edibles. This is perfect. These are the three main methods, and their risk profiles differ significantly. Smoking carries respiratory risks, while edibles carry a higher risk of acute overdose/psychosis due to dosage difficulty. This is a foundational choice.

Frequency of Consumption (Dropdown): You asked what the literature uses. Research often groups frequency into categories that correlate with different levels of risk, especially for Cannabis Use Disorder (CUD). For a user-friendly dropdown, I'd suggest:

Rarely (A few times per year)

Occasionally (1-3 times per month)

Regularly (1-3 times per week)

Daily or Near-Daily

Dosage (Dropdown, for Edibles): Your instinct here is spot on. Dosage is the single most important factor for the acute risks of edibles. A conditional dropdown is the right way to do it. The standard unit is milligrams (mg) of THC. I'd suggest these choices:

Low Dose (2.5mg - 5mg THC)

Standard Dose (10mg THC)

High Dose (25mg+ THC)

Are You Missing Any Inputs That Would Matter? Yes, based on the data, there are two other crucial factors that dramatically influence risk, especially long-term risk:

Age (Especially Age of First Use): This is arguably the single most important factor for long-term cognitive and mental health risks. The adolescent brain is uniquely vulnerable to the effects of THC. Heavy use before age 18 is linked to much higher risks of addiction and potential long-term cognitive changes. We should probably have a dropdown for the user's current age group (e.g., <18, 18-25, 26+).

Potency (THC Concentration): The difference between smoking 1980s-era cannabis (~4% THC) and using modern concentrates (~85% THC) is immense. This is the equivalent of comparing a light beer to a shot of Everclear. This is a harder for a user to know, but we could offer a simplified choice as a proxy:

Standard Potency (Traditional flower)

High Potency (Concentrates, "dabs," most vape oils)

Your Long-Term Risk Model Idea Your proposal to frame the output as "if you keep smoking every day for 20 years... you have a 1 in XXX chance..." is an excellent way to communicate cumulative risk.

It's far more intuitive and impactful than asking for a user's history. It turns the page from a simple calculator into an educational tool about future consequences.

For Addiction (CUD): This works perfectly. The data is strong that daily users have a 30-50% chance of developing a substance use disorder. We can present that as a "1 in 2 to 1 in 3 chance."

For Other Risks (e.g., Lung Cancer): This is where we'd have to be careful. The link between smoking cannabis and lung cancer is much weaker and more controversial than for tobacco. We would need to reflect that uncertainty, perhaps by showing the range of estimates as you suggested.

In summary, a great sketch. I think the ideal set of inputs for the main page would be:

Method: Smoking / Vaping / Edibles

Frequency: Rarely / Occasionally / Regularly / Daily

Your Age: <18 / 18-25 / 26+

Typical Potency: Standard (flower) / High (concentrates)

Edible Dose (if applicable): Low / Standard / High Gavinandresen (talk) 17:09, 10 October 2025 (UTC)====