Week 2: Big Idea -> Smaller Focus
0: Summary of Last Week
Last week we presented our research on mental health care. Through literature review and some interviews, we identified the following issue and opportunity:
- There is a global shortage of mental health workers. In low-income countries, the ratio of provider-to-patient can go as low as 2:100,000.
- It is possible to decentralize mental health service. Delegating tasks from specialists to non-specialists, through short training, has been proven effective and sustainable.
We also had a rough proposal of how we might facilitate the decentralizing of mental health care. However, there were still a lot of questions we needed to answer:
- Does this service have to be built on blockchain?
- How should we narrow the scope down?
- What medium would be most appropriate?
- How can we validate our idea?
1: More Interviews & Research
To answer these questions, we conducted even more interviews and research.
Interview #1: Joerg Blumtritt, creator of Building Blocks
Building Blocks was one of the applications we studied last week. It is an aid distribution system built on blockchain, and it was implemented in a refugee camp in Jordan. We met with the creator of the system, Joerg Blumtritt, while he was in NYC.
We explained to him our idea and we asked the question: does our service have to be built on blockchain?
His answer: no, it does not.
First, privacy can be afforded by conventional backend system, through homomorphic encryption. This encryption allows a signature (e.g. an iris scan) to be stored and verified when needed, but this encrypted signature cannot be reverse-engineered and used to recreate the original signature. Although in practice it is computationally expensive, it is a viable method.
Second, blockchain helps with peer-to-peer transactions that cannot have a 3rd party overseer, or an organization that substitutes or compensates for trust (e.g. Uber or Airbnb). Because our project will involve mental health organizations, building our solution on blockchain is not necessary.
Interview #2: Jasrin, Mental Health Counselor at NYU
Because there is an immediate and increasing need for crisis helpline, volunteers are sent out without adequate training…Some volunteers lack cultural sensitivity and the vocabulary to communicate empathy.
Interview #3: Rosi, Grad Student at NYU Wagner’s Health Policy & Management Program
Models and training programs are largely legacy programs. They do not translate well into local communities and subcultures…The toolkit given to workers and patients are low tech. There is definitely room for improved efficiency.
Interview #4: Amrin (not real name), Mental Health Service Client
I called the suicide prevention helpline multiple times. In some of these calls, the responder was properly trained and I felt heard. In some other times, I got offended and not heard… I felt worse after…I don’t feel the responder was genuine in their responses.
Interview #5: Elias (not real name), Helpline Volunteer
Sometimes I don’t know what to say, and sometimes I forget the protocol…I also had so many prank calls. It was a waste of my effort.
Research #1: Reaching out to Organizations
We reached out to ThriveNYC, Vibrant, and ICNYU. They have yet to get back to us, so this is still in progress!
Research #2: Helpline Training Manuals
Lastly, we did a quick review of the training manuals for suicide prevention and crisis helpline.
While these manuals are very comprehensive, they usually are really really long. For example, the manual that we found (which is one part out of 6) is 30-pages long and extremely text-heavy.
2: Defining a Smaller & Addressable Problem
From our 2nd round of research and interviews, it is clear that in NYC, there have been efforts to decentralize mental health care through enlisting volunteers for first aid and crisis helplines.
However, we also noted that due to the increasing need for such helplines, the training systems have not been able to keep up. These training processes appear to not have been designed to be user-friendly. As a result, volunteers are being sent out before getting adequate practical experience. This is problematic because when a suicide or crisis helpline does not perform up to standard, there could be fatal consequences.
Decentralizing mental health care as it turns out, even in a resource-rich city like New York, is not easy. There definitely needs to be a way to conduct more efficient training. Also, time & resource efficiency would mean an increased likelihood that such programs can be adopted in areas with fewer resources.
3: Our Proposed Solution (v2)
We are planning to design an AI-powered, voice-chatbot to effectively train helpline volunteers.
This chatbot will improve the training process by:
- Emulating practical experience. The trainee will talk to the bot as if they’re talking to a real caller.
- Improving empathy training. Powered by machine-learning, the bot will analyze the vocal tone of the trainee, as well as identify keywords that are deemed insensitive.
- Tracking volunteer progress. Because the training is digitalized, supervisors would be able to monitor the progress of numerous trainees.
The goals of this bot are to increase the efficiency (time & volume) of the training process and to ultimately train volunteers to deliver better services.
4: Persona & Next Steps…
We created a persona for our user, based on the interviews and insights we have gathered. Meet Angela the volunteer:
And now that we have a persona we will center our design around, these are our next steps in the project:
- Talk to a program coordinator or lead. We need to understand the organization’s perspective on conducting volunteer training. Furthermore, it is important to get buy-in from existing organizations.
- More volunteers. More training manuals. So that we can identify the exact areas we can improve on: keywords to track, empathetic expressions to enforce, etc.
- Survey training bots & voice chatbots. What products and services are out there? What are the best practices? What is or will be possible?
