DIFD Mach-Gaensslen Chair in Suicide Prevention Research Update

Dr. Kaminsky’s 2019 Update

Clinical studies

In January 2019, we began a 6-month contract with the Department of National Defence. Our goal was to find a biological identifier (biosignature), of PTSD and more specifically, the dissociative subtype of PTSD.

After recruiting a group of 32 former Canadian Military personnel, we generated an algorithm (or a classifier model) that can predict the status of PTSD as well as the dissociative subtype. Here are the accuracy results:

  • Algorithm predicting PTSD: 96 per cent accurate
  • Algorithm predicting the dissociative subtype of PTSD: 91 per cent accurate

During the summer, we successfully competed for a $1,000,000 contract, which now allows us to continue our work this year. By pairing our algorithm (classifier model), with other various longitudinal metrics, we hope to generate a model that predicts long-term outcomes. 

Epigenomic analyses

Made up of chemical compounds and proteins, the epigenome attaches to DNA. It may manipulate functions and direct actions like turning genes on or off, or controlling the production of proteins in particular cells.

We have performed the largest epigenome-wide association study on alcohol use disorder to date, examining over 600 individuals and pairing our findings with brain imaging data and HPA axis metrics. The HPA axis is a term used to represent the interaction between the hypothalamus, pituitary gland, and adrenal glands; it plays an important role in the stress response.

As a result, we identified an association in a gene, which has implications for HPA axis function. Our article is under revision at Molecular Psychiatry and we are hopeful we will be successful. We also hope to publish our suicide-associated biomarker found in this group in the coming year.

Paper publication activities

In addition to the above article, we published a work replicating our postpartum depression biomarkers in an additional two cohorts. The article is in press at psychiatry research.

Social media artificial intelligence development 

Using our social media artificial intelligence (AI) tool, we have developed an algorithm that not only assesses an individual’s risk of suicidal ideation, but also when they will be at risk, based on their Twitter posts. By assessing the results of an individual’s score on the AI report, we can determine the likelihood of suicidal thought within the next 5 days.

We also developed a method to adapt our AI scoring to the regional assessment of risk. This method only takes about 2 weeks’ worth of data to generate these metrics, as compared to the year of indicator currently used as the state of the art. As such, we envision a method by which to assess regional risk in Ottawa and to direct suicide prevention resources to particular areas.

Critically, this tool will also enable us to evaluate the effect of these efforts in almost real time, allowing us to adapt and optimize our methods.

Click here to listen to Dr. Kaminsky's interview with CTV, during Bell Let's Talk!

Suicide Prevention Ottawa (SPO)

I have been a member of the Suicide Prevention Ottawa (SPO) steering committee for about a year and a half and this year I took the role of inaugural chair of the Research and Knowledge Exchange committee of SPO. There, I have been working with my committee to create a resource for local suicide prevention researchers to gain SPO’s endorsements in exchange for evaluation and incorporation of SPO’s mission, values, and agenda into their research.

I am currently also spearheading the creation of a resource initially dubbed ‘People With Experience Lived and Living’ (People WELL), which will be a working group of individuals with lived and living experience that local researchers can interact and work with. The idea is to bring this voice to the table for the inception and co-development of research projects, making them more client centered.