Examining Psychedelics Use and Mental Health Conditions: Supervised Learning

[lisagirard]

Thinkful Supervised Learning Capstone Project:

Examining Psychedelics Use and Mental Health Conditions

Google Slides presentation: https://bit.ly/LG-PsychedelicMHCs

An examination in publically-available raw data regarding psychedelics use in US citizens, 2020

Objective: predicting psychedelics use based on certain mental health conditions and/or recent prior knowledge of psychedelic therapeutic use

Google Colab Python notebook code

Conclusions:

• Can predict with 85% accuracy that those with certain mental health conditions AND prior knowledge of psychedelics-as-therapeutic-tool are using psychedelics
• Establishes connection between recreational use as form of self-medicated mental health care
• Connection between psychedelics and mental health care can advance science about:
  • how these compounds work in relation to mental health
  • how to use them safely
  • how to regulate and optimize

Data Analysis:

DATA SOURCE: Psychedelic Mushrooms in the USA: Knowledge, Patterns of Use, and Association With Health Outcomes


Downloadable raw data set

My analysis included:

• Comprehensive EDA
• Hypothesis/Objective formation
• Imbalanced data mitigation:
  • random manual sample generation
  • SMOTE synthetic resampling
  • ROS/RUS synthetic resampling
  • Tomeks/NearMiss synthetic resampling
• Feature engineering
• Benchmarking
• General Logistic Regression predictive modelling
• KNN classification
• Random Forest classification
• Penalized SVM/SVC classification
• Final recommendations based on my findings for further data exploration and industry decisions/progression

December 19, 2022

[CheekeyMonkey]

This seems very useful to practitioners, however, do you have any summary take aways for clinicians who are not too familiar with data science?

[lisagirard]

HI - Thanks for taking a look. Did you perhaps mean "clinicians who are not too familiar" ?

[CheekeyMonkey]

Updated thanks!

[lisagirard]

Hi,

To respond to "clinicians who are not familiar with data science"...

In this project, I am using reported data from recreational psychedelics users. These users had indicated several health conditions, some mental health related, some not. Some of the conditions are noted with self-report, and some are reported with actual testing scores using known testing methods that indicated moderate-to-severe depression and anxiety.

My analysis determined that 67% of recreational psychedelics users (from this data set) had self-reported for mental health conditions such as depression, anxiety, insomnia, chronic pain, migraine, and/or high test scores from other diagnostic tools.

When I ran the data again with an added factor of "knowledge of psychedelics as a therapeutic tool" ((read: think about the communities that are experimenting with microdosing psilocybin, or using cannabis to treat anxiety)), the percentage of use increased to 85%.

Here we can see how the data has helped establish a connection between education of therapeutic tool and use to alleviate mental health issues.

From here, we can begin to emphasize continued education and research about use of psychedelics to treat mental health conditions that can and will:

• Reduce stigma
• Allow for DEA rescheduling & decriminalization
• Provide more tools for those suffering
• Establish statutory and regulatory policies and practices
• Shorten mental health treatment periods to days or weeks
• Increase research funding and testing
• Change methods, modalities, and mindsets around mental health and treatments options

Thank you for your question @CheekeyMonkey and I would be glad to discuss more.

[VoytekBereza]

"_67% of recreational psychedelics users had self-reported for mental health conditions". Interesting, makes me think about intention. I wonder whether although they are taking it recreationally whether they have any mental health benefits from use.