Canada’s Mental Health Crisis

Our mental health system faces significant challenges such as a shortage of psychiatrists, long wait times, and stigma. 1 in 5 Canadians experience a mental illness, and it’s the country’s leading cause of disability (Source: CAMH). 

The COVID pandemic compounded these problems with added stress and forced isolation aggravating mental health issues like depression and anxiety and making it more challenging for mental health workers to provide the necessary care. 

One part of addressing these issues is further relying on digital transformation. The pandemic unlocked a new world of online care and showed what technology could achieve. This includes not only remote treatment but also a more dynamic use of the power of networks to support mental health, Artificial Intelligence (AI).

AI is being introduced to the mental health field to deliver better, more personalized care, shorten wait times, remove barriers to accessing treatment, and boost efficiency in medical and therapy settings. So far, the results have been promising. 

Artificial Intelligence (AI) vs. Machine Learning (ML) vs. Deep Learning (DL): What’s the Difference? 

Before diving into this topic, it’s important to note the differences between AI, ML, and DL, which will frequently appear in this article. 

• Artificial Intelligence (AI): developing machines to mimic human behaviour (behaving the way a human would behave). 
• Machine Learning (ML): a subset of AI that enables a machine to learn by looking for patterns within algorithms and trying to make predictions. 
• Deep Learning (DL): a specific type of machine learning that mimics how a human brain works (thinking the way a human would think). 

It’s important to understand these differences because AI is a catch-all term for ML and DL, but there are significant differences between types of AI. 

More information on the difference between AI, ML, and DL.

How Did the Pandemic Impact Mental Health Services? 

The COVID-19 pandemic took us all by surprise, with countrywide lockdowns forcing the population to remain isolated and swap in-person health services for online ones. 

The increase in telehealth during the pandemic was driven primarily by those seeking mental health services (Source: mHealthIntelligence). According to one survey, 54% of Canadians said their mental health had worsened during the past two years (Source: CBC). Young people were significantly impacted, with 48% of young Canadians sharing that isolation and loneliness were primary challenges for them during the pandemic (Source: Unite for Change). 

From the practitioner’s side, counselling services received overwhelming requests due to rising cases of anxiety and depression. One survey revealed that 84% of psychologists who treat anxiety disorder noticed an increase in the need for treatment since the start of the pandemic (Source: American Psychological Association). Similarly, the demand for depression treatment rose from 60% to 72% over one year. Between long waitlists and working beyond their traditional hours, 41% of psychologists reported being unable to meet the demand for treatment, and 46% reported feeling burnt out (Source: American Psychological Association).

Barriers to Accessing Mental Health Services

The pandemic-related increase in mental health issues among Canadians is made worse by an accessibility problem.

Before discussing how Artificial Intelligence can help, let’s identify the significant issues preventing the accessibility of mental health services. 

Shortage of Mental Health Professionals

There’s a pressing need for more mental health services and support in Canada. Researchers found that 32% of Canadians feel they need professional mental health care but can’t access it (Source: Healthing). A significant shortage of psychiatrists, causing long wait times of between 6 months to a year, is mainly responsible for the lack of access (Source: Healthing). Therapists and counsellors face a similar problem, with 1 in 10 patients waiting up to four months to access counselling services (Source: The Star). 

Barriers for Youth

Youth have difficulty accessing mental health services and are at the center of the mental health crisis in Canada. The risk of mental illness and substance use problems is highest among young people aged 15 to 24 (Source: CAMH). Cost is a significant barrier to youth accessing mental health services, with privatized care fees ranging from around $100 to $225 per hour (Source: The Conversation). 

Stigma

Stereotypes surrounding those with mental illnesses prevent patients from seeking the help they need. According to a World Health Organization study, 30 to 80 percent of those with mental health issues don’t seek treatment (Source: High Watch Recovery). It’s common to hear stereotypes about people with mental health issues like they’re dangerous, incompetent, or responsible for their illness (Source: Progress in Mind). 

Artificial Intelligence’s Role in Mental Health Services 

Because of the issues listed above, mental health professionals have started to introduce AI in their practice, and the results are already proving to be very promising.

Here’s how AI is leveraged at each step of mental health services today.

Custom Care

Natural-language processing (NLP) is a method that uses machine learning to process and understand human language. In mental health care, professionals use NLP to analyze the language used by clients during therapy sessions (Source: ieso). This process pinpoints the specific vocabulary between a therapist and client that are most successful in treating a disorder. With NLP, therapists can ensure that therapy is provided to patients at the greatest quality level. 

The NLP process goes something like this: 

• Machine learning recognizes patterns in patients’ language and their statements.
• A transcript with this data is created and analyzed based on past datasets. Patterns are then automatically identified by the machine learning algorithm.
• Based on these patterns, the practitioner can target the patient’s specific needs and tailor their treatment program, leading to better care.

ieso, a UK-based company, created an online platform to provide mental health care using machine learning to identify language patterns. The organization uses what they call a “treat-data-improve cycle.” Through this process, they collect research patterns from thousands of treatment sessions and use this past data to improve treatment for various conditions such as depression and anxiety. 

Their results are significant, with a 1/3 higher recovery rate than face-to-face therapy, showing the promise of AI-based online treatment. Their services also reduce waiting lists by 30%, getting patients into treatment faster (Source: ieso).  

Monitoring Symptoms & Tracking Progress 

Therapists use AI to better understand their treatment’s effectiveness and modify a patient’s treatment plan depending on their symptoms and progress. 

One example of this is being done by monitoring patients’ speech patterns. For instance, a monotone voice could be a sign of depression, and fast speech could be a sign of mania. Sometimes changes in speech and language aren’t so noticeable. AI algorithms can be trained to identify these subtle changes by flagging concerning speech signals or patterns (Source: TIME). 

The statements made by patients in therapy are also a strong indicator of how they’re progressing. Not only is AI a part of improving treatment, but therapists also use ML to automate transcripts in counselling sessions and analyze the patient statements against data sets to substantiate their progress (Source: Taylor and Francis Online). 

One area in which AI is becoming more prevalent is cognitive behavioural therapy (CBT), a talk therapy that helps manage problems and emotions. In CBT, patient language is measured using observational coding (Source: Taylor and Francis Online). Observational coding segments and transcribes audio and/or video recordings of individual counselling sessions, which is traditionally done manually and requires a lot of time and labour. 

Coding patient language is now automated with deep learning, with statements being coded into one of two categories: “change talk,” any language progressing towards changes in emotion and behaviour, and “counter-change talk,” a movement away from resistance or change. 

For instance, a patient making a statement such as, “I’m never going to get better,” would be an example of counter-change talk. In contrast, a statement such as “I am motivated to put in the work to get better” would be an example of change talk. Keeping track of these statements is crucial since therapists can use them to assess whether adjustments to treatment are required. 

Reducing Wait Times and Combatting Misdiagnosis

In physical healthcare settings (i.e., hospitals, doctor’s offices, etc.), AI is already used to help combat overwhelming wait times. Humber River Hospital in Toronto was the first in Canada to employ AI to track and control patient flow, making hospitals more efficient (Source: The Globe and Mail). AI also expedites pathology assessments by analyzing medical images for illnesses like cancer and eye disease, enabling faster and more efficient treatment. 

Similarly, AI is used in the mental health field to provide diagnoses faster and more accurately, boosting efficiency and speed of treatment, which results in reduced wait times. 

Correct identification of mental illness remains challenging, with one study showing that more than a third of patients with a severe psychiatric disorder have been misdiagnosed (Source: Everyday Health). While more research is needed, there’s promise for video analysis to detect patient symptoms. For example, picking up on signs of anxiety like nail-biting, knuckle cracking, and fidgeting. ML may detect behavioural signs of anxiety with over 90% accuracy (Source: Verywell Mind). Spotting a mental illness right at the onset of symptoms could help lead to a quicker and more accurate diagnosis, meaning patients get access to the proper treatment faster. 

Interested in learning more about how technology will transform healthcare in Canada? Check out 6 Ways 5G is Improving Our Healthcare System.

AI’s Role in Treating Different Mental Health Issues

AI is effective in multiple areas of mental health. Let’s look at some examples of how AI is being used to treat various mental health issues.

Depression and Anxiety

AI therapy apps can significantly reduce symptoms of anxiety and depression. One study examining an AI therapy app called Youper found a decrease in anxiety by 24% and depression by 19% within the first two weeks of treatment (Source: HealthITAnalytics). 

The app combines telemedicine and AI capabilities to foster patient engagement and strengthen mental health services. Products like Youper offer patients various remote services, from free AI mental health assessments to therapy exercises for emotional regulation. While supporting patient health, AI Therapy Apps also empower health professionals by automating triage, gathering patient symptoms and histories, highlighting risk factors, and routinely checking in with patients. 

The Connection Between Physical and Mental Health

Often overlooked, mental health is a significant aspect of those going through treatment for a life-threatening condition. This is an issue with cancer treatment patients whose mental health is at extreme risk of severe anxiety and depression, which can significantly harm recovery outcomes. 

An overwhelming percent of cancer patients experience challenges with their mental health, with one survey revealing that 50.7% of patients had symptoms of anxiety during the pandemic, and 46.8% reported having depression (Source: HealthITAnalytics). By using AI solutions, doctors obtain a more comprehensive understanding of a patient’s mental health and can use that knowledge to better fulfill the patient’s needs. AI gathers feedback by following trends and the patients over time. Physicians then access this data, identify patients experiencing depression or anxiety, and implement solutions. 

It’s evident that physical and mental health are interconnected, and optimism is proven to strengthen this connection. According to a meta-analysis of 83 studies, optimism was found to have better health outcomes for cardiovascular disease, cancer, pain, physical symptoms, and mortality (Source: Verywell Mind). Optimism is also essential for our health, as research shows that optimism can positively impact our health. Those with a more positive outlook on life tend to be healthier and live longer (Source: Positive Psychology).

Post-Traumatic Stress Disorder

Post-Traumatic Stress Disorder (PTSD) is a challenging disorder that impacts up to 25% of veterans (Source: NYU Langone Health). Nearly one in ten people in Canada may have PTSD at some time in their lives (Source: The Canadian Encyclopedia). Since PTSD deals with complex trauma, a patient is often unwilling to disclose such sensitive info to a human. According to one study, 28% of patients admitted giving medical professionals false or omitted information about themselves (Source: Futurism). 

AI is currently used to help treat PTSD through virtual therapy. An AI virtual therapist supports the treatment of mental health issues. Specific to PTSD, an AI virtual therapist helps foster a safe space for patients to communicate openly, removing the barrier of feeling judged.  

Take, for example, Ellie, a virtual therapist created by the University of Southern California’s Institute for Create Technologies, who’s designed to detect signs of depression in PTSD by tracking and responding to visual and verbal cues (Source: News.com). Making the therapist virtual helps patients more inclined to open up because they feel less judged about their experiences (Source: News.com). AI works in tandem with a mental health professional as a data-gatherer, learning things about a patient that would otherwise not be disclosed. 

Addiction 

Those with a mental illness are twice as likely to suffer from addiction (Source: CAMH). Relapse and overdose are two of Canada’s most urgent addiction-related issues. The leading addictions in Canada are opioids, alcohol, cannabis, cocaine, gambling, and methamphetamine (Source: Canadian Centre on Substance Use and Addiction). A countrywide opioid overdose problem particularly impacts Canada. In 2021, there were 7560 fatalities from apparent opioid overdose, which equates to around 21 deaths per day (Source: Government of Canada). This crisis is directly tied to the lack of capacity to treat addiction. Medical professionals are increasingly using technology to address this growing issue, ensuring everyone has access to adequate care.

With an app called Sober Grid, AI is used to predict the risk of relapse by analyzing the language used by the app’s user (Source: Richard Van Hooijdonk). The company collected data from more than 120 000 addicts in different stages of recovery to train an algorithm that recognizes and predicts when relapses occur and what kind of communication precedes it. When the app detects a person is likely to relapse, it immediately offers different types of support, such as cognitive behavioural therapy or getting the user in touch with an available trained coach. 

Trycycle is another example of a company that uses data to make predictions about relapse. Their system acts as an early relapse warning system following three steps: 

1. Self-assess: the person in recovery is regularly promoted to submit journal entries through their mobile device.
2. Algorithm: to assess the client’s health status, their inputs are evaluated using proprietary algorithms, and clinicians have direct access to this information. 
3. Clinician Dashboard: the gathered data determines each person’s risk of relapse in real-time. A dashboard that displays the outcomes allows for human-based decision-making and ongoing monitoring.

Considerations For Creating an AI Mental Health Solution

Canadians should ensure that patient safety and privacy are promoted throughout the development of an AI mental health solution. Here are some considerations. 

Cybersecurity of Our Mental Health Data

According to the University of Maryland, a cyberattack occurs every 39 seconds (Source: UMD). Data breaches often expose sensitive information to the public or use it as ransom. It’s crucial to ensure that the platforms that hold our medical data are fully secure from cyber threats.  

Ethical Transparency of Data Usage

17% of all sensitive files are accessible to employees (Source: Varonis). For this reason, a company’s integrity is of utmost importance. Canada needs to ensure that the companies developing and using AI mental health solutions remain transparent with how data will be used and not cross boundaries with patient privacy. 

Algorithms Developed with Mental Health Specialists in Mind

Lacking the right talent and team collaboration often results in failed algorithms. A 2018 report estimated that 85% of AI projects would fail and deliver erroneous outcomes through 2022 (Source: AI Multiple). Both a mental health expert and an AI expert must collaborate to develop these AI and mental health algorithms.  

Accredited Artificial Intelligence

If humans must go through a significant accreditation process to become a psychologist, shouldn’t machines have to do the same? Algorithms and machines should be tested and validated to ensure they’re operating correctly and are suitable to work in various mental health scenarios. 

Will AI Replace Therapists? 

Using AI in therapy aims to improve care and fill gaps, not take away jobs. AI’s goal is to assist therapists, not replace them. AI should help patients monitor their health, provide preliminary assessments and coping mechanisms, and help therapists deliver more successful treatments. 

We need AI in our mental health services. 

In treating mental health patients, we should see AI as a tool to analyze data that mental health professionals use to provide better services. AI complements therapy and helps to transform how we access mental health services. 

Technology has already significantly impacted our way of life, and mental health treatment may soon follow suit. The mental health crisis will only continue to grow, and we need practical solutions now. It’s estimated that 8.9 million people in Canada could be living with a mental illness by 2041, based solely on demographic changes (Source: Mental Health Commission of Canada). 

AI is establishing its position in the future of mental health care. Let’s have a look at one study that’s looking to show the value of using AI in mental health treatment.

Using AI to Expand Access to Mental Health Care

The University of Illinois Chicago is currently conducting a 5-year study where researchers combine AI voice technology with problem-solving therapy (PST). This will be a systematic approach to helping individuals focus on improving their cognitive and behavioural abilities through AI-delivered mental health treatment (Source: University of Illinois Chicago).  

In their study, an AI virtual agent named Lumen, an Alexa-like app that’s voice-activated, is used to deliver problem-solving skills to patients with moderate anxiety and/or depressive symptoms. Individuals complete eight one-on-one counselling sessions to identify a stressor in their life, and the AI counsellor coaches them to define goals and possible solutions. After solutions are compared, an action is established to put the chosen solution into practice. 

Dr. Jun Ma, a professor and key researcher in this study, says that rather than replacing mental health specialists, the motivation of the study is “trying to leverage the latest technology as a vehicle to improve access to proven psychotherapy” (Source: University of Illinois Chicago). In doing so, researchers hope to “improve the reach and impact of psychotherapy” while minimizing barriers to mental health care for people who need it most (Source: University of Illinois Chicago). Researchers also hope this tool will support currently available mental health services, which are severely in short supply.

The study’s results will be published in 2025 and could demonstrate the future of using AI to improve mental health treatment. 

So no, you shouldn’t expect your future therapist to be a robot unless you prefer it that way. What you can expect, however, is advanced technologies helping to make your treatment more effective and accessible. 

Are you interested in learning how AI and other advanced technologies will help drive innovation and improve vertical sectors?

Download our “Next Generation Network Imperative” Whitepaper to find out what the future of next-generation technologies will look like.

The International Data Corporation (IDC), a world-renowned global market intelligence firm, was commissioned to complete this study.

About the Author

• 

Gabby Mazza is a Content Writer and Marketing Student at CENGN (Summer 2022) and a Psychology Co-op student at the University of Guelph. Her passion is collaborating with others to produce engaging and inspiring content on any topic. Gabby enjoys creative writing, painting, and playing the ukulele in her spare time.

More by Gabriella Mazza