DMIND Application for Screening Depression – An AI Innovation from CU Faculties of Medicine and Engineering Researchers

Introducing a new dimension for consultations with those suffering from depression by the Department of Mental Health and the DMIND AI Innovation from Chula’s Faculties of Medicine and Engineering that provides screening for depression through the Mor Prom Application with greater accuracy, accessibility, and convenience thus reducing the burdens on medical practitioners and psychologists in taking care of patients with depression.

Originating the DMIND Application

Statistics from the Department of Mental Health, Ministry of Public Health are a clear indication that Thailand is faced with the problem of depression leading to suicidal tendencies at a worrisome rate.   In 2021, at least 1.5 million Thais were identified as suffering from clinical depression, and out of a hundred patients, only 28 had access to treatment whereas the rate of those attempting suicide is around 6 persons per hour.  Statistics also show that patients suffering from depression are 20 times more successful in their suicide attempts than the average person, while 70 % of these patients meet an untimely death.  It has been predicted that in 18 years, depression will be the world’s highest form of illness demanding medical attention.  

From the statistics and reasons mentioned here, psychiatrists who are now at the forefront in dealing with the situation such as Associate Professor Solaphat Hemrungrojn, M.D. from the Department of Psychiatry, Faculty of Medicine, Chulalongkorn University hope that some form of innovation can be developed to address this problem.

Associate Professor Solaphat, M.D. shared with us that she “…had the opportunity to talk to other psychiatrists, some of whom work for the Department of Mental Health and there seems to be a common problem we all encounter with a significant rise in the number of patients suffering from depression.  Even if we try to increase the number of practitioners in this field it is still insufficient.  We can see from the news where film stars or celebrities are suffering from depression and yet, in reality, there are so many others who aren’t well known but they too are faced with depression and some decide to put an end to their own lives. This is a problem that worries those of us who are in this profession and has led us to try to find a way to deal with it more effectively.” 

Depression is treatable but first and foremost, we need to provide the opportunity for all patients to have access to the treatment they need. 

“From the way things are at present, even if hospitals were to expand their OPDs or sections to cope with the volume of patients that still won’t be possible.  In some provinces hospitals now see a rise to as many as 200 – 300 patients a day.  This is already discounting the fact that there are also those without access or those who dare not go to the hospital for treatment.  We all know that the Department of Mental Health offers a hotline where one can ask for consultation with a psychologist who can provide initial screening and from what I have heard the number of those assigned to this task is certainly inadequate since there are sometimes thousands of callers waiting for their turn to air their problems.  A large number of people cannot avail of this service and this means they can become a suicide risk.  How is one to know which caller on which line needs urgent help?  Is there some form of technology that might solve this problem?”  Dr. Solaphat poses this crucial question.    

Due to her expertise in psychiatry, Associate Professor Solaphat, M.D. recognizes the signs one usually identifies with depression that can be detected by way of a person’s voice, words, and facial expressions.    Some kind of equipment that can help in this analysis would certainly be helpful.  And this is how the DMIND application came about – an innovation that is the result of collaboration between Chula’s Faculties of Medical Science, Engineering, the Department of Mental Health, the Ministry of Public Health with the Thai Health Promotion Foundation, and Agnos Health Co. Ltd.  The D in DMIND being short for Depression. 

What is DMIND?

DMIND is designed to help psychiatrists but not to replace the psychiatrist.  All that it does is a help screen the urgency and sequence of cases deserving urgent attention.

Dr. Solaphat added, “here at Chula we have held talks amongst us on the possibility of developing such equipment and when the Department of Mental Health became involved, we realized it was time we applied our lab experiment to actual recordings we got from the Mental Health Department.  The process proved to be more difficult than we had anticipated.  We realized that the techniques used or published in foreign countries were not applicable in Thai cases.  This meant having to start again from the beginning and analyze tens of thousands of sample recordings until we achieved a Machine Learning AI that could serve as a template in actual use.  This increased our confidence in the accuracy of figures shown in the data that would improve even more with AI technology.” 

In the meantime, the Mor Prom Application also reported higher numbers of depression cases among its users.  Previously Mor Prom relied on the questionnaire format for screening patients.  It was found that the use of AI analysis increased the level of accuracy with its Direct Bio Tracker capacity which can detect patients’ emotions from their facial expressions that could be a result of changes in the chemical levels in the body.  These are actual biological changes and not mere opinions. Many of the patients aren’t even aware that they have shown those signs.  

The AI will assess depression characteristics on a graded spectrum.  On this scale, green shows the person’s condition is at a normal stage, yellow shows a level of depression where a psychologist will make contact within 7 days, while red means severe depression which means a psychologist will make contact within 1-24 hours.  The criteria derive from existing divisions used by the Department of Mental Health.  Dr. Solaphat explained that when working together one should opt for something simple that is already familiar to all parties to ensure the best results.     

“First of all, we need something that the public, whether it may be the elderly or those who are not familiar with the use of applications, can access.  Secondly, users must be made to feel at ease and are free to say what they like.  In our experimental stage, there were times when patients who appeared normal or only slightly depressed, once given the chance to talk to the DMIND application in the form of an avatar (Dr. Pordee), would pour out all their feelings since they felt there were no barriers and allow us to gain in-depth information.  At first, we feared that our patients would be hesitant to open up to technology, but in reality, the results have been great,”  Dr. Solaphat enumerated the qualities of this application. 

She also added that “What makes me very happy today is that personally, all I know is my own area of specialization which is within the confines of psychiatry.  Right now, we can see that our success has come about through collaboration with lecturers from the Faculty of Engineering.  We have been successful in gaining what we had aspired to and what other countries had achieved.  We should realize that there are many highly capable people in Chula if we could get together to talk and work, we might be able to create world-class innovations.” 

The DMIND application and how it works

Associate Professor Dr. Peerapon Vateekul of the Department of Computer Engineering, Faculty of Engineering, Chulalongkorn University who developed the AI used in this application explained that DMIND is an AI that is used directly for assessing depression.   The project began in 2020 when Associate Professor Solaphat, M.D. invited her colleagues from the Faculty of Engineering to develop this application.  Our collaboration was extended last year to include the Department of Mental Health whereas this year our screening has become available on the Mor Prom Application.    

According to Associate Professor Dr. Peerapon, DMIND comprises three parts – the first is the application in Mor Prom or the Automate Avatar which is a mobile application that patients access to undergo screening.   When a conversation takes place via the app the data will be processed in a video file which will then be sent to the AI for analysis in the second part.  The last part is the Web Base that the psychiatrist can check retroactively.  If any patient shows signs of worrying behavior, a doctor or psychologist from the Department of Mental Health hotline will follow up and see that the patient receives the proper care needed. 

The general public can easily access the application through the Mor Prom application on their smartphones or computer tablets and select the “Talk to Mor Prom” (Chatbot)” then select “Mental health check-up” and choose “Mental health check-up with Doctor Pordee” after which they will find a questionnaire which helps with the screening and self-assessment process (in case the user doesn’t feel at ease with voice recording, camera or video recording).  If an in-depth analysis is needed, the user must first permit a camera recording of the conversation with the avatar to take place and the data of images and voice recordings will be kept confidential.  Dr. Peerapon clarified the fact that questionnaires are used in the screening process and that since many depression patients are not yet ready to seek medical help or to reveal their illness the questionnaire is helpful for the initial stage.    

The DMIND’s Efficiency

“One must admit that the Department of Mental Health’s 1323 hotline is inundated by calls beyond the capacity to attend to all those in need.  We all joined forces in developing DMIND since we could perceive the problem and tried to find an effective solution.  In the past, all calls had to be answered regardless of the severity of the caller’s depression and each call had to be assessed individually.  What was needed, therefore, was a system that could give priority to cases that require urgent attention.    Statistics have shown that the introduction of DMIND for initial screening has been very helpful as it has enabled the hotline to identify those severe cases in need of immediate care and attention,”  Dr. Peerapon added.  

In terms of its accuracy in analyzing facial expressions, gestures, voice, and content seen on DMIND, Dr. Peerapon clarified that currently, the implementation of the experiment is around 75% which is remarkably high compared to the number of patients that reaches the tens of thousands.  

“In terms of maintaining the confidentiality of personal data, the university has its committee overseeing ethics in human research.  No data can be traced to the app users, and after some time, all data will be destroyed.”

DMIND in the future

Dr. Peerapon also expressed his desire that in the future this application should be accessible to the general public.  The project is open to agencies, hospitals, and technology firms that might wish to join in developing or extending the AI to suit their needs since depression is a national problem.  Extending the use of this innovation will help increase its impact while reducing the suicide rates in Thailand.  

“Teaching an AI is much like teaching a child or even a person.  Some points can be further developed and we also hope that the number of users will increase so that the data will also increase leading to more accuracy” he concluded. 

The strengths of DMIND

1.  It uses AI technology which can analyze facial expressions, gestures, and emotions at the Direct Bio Tracker accuracy level.  The large volume of data contributes to its exceptional accuracy. 
2.  It is the only Thai application developed specifically for Thai users that relies on data directly compiled and researched from Thai cases. 
3. It is used as a tool to help reduce the burden of psychiatrists and psychologists whether it be in terms of providing care for patients, follow-up treatment, or changes of prescriptions applied alongside psychotherapy.  The data can be used to extend research aimed at preventing and curing depression in the future. 
4. It is easily accessible via the Mor Prom app, is user-friendly, and enables all those who are at risk or suffering from depression to receive timely help. 

How can the general public access the DMIND application?

The public can access the DMIND Application at https://bit.ly/DMIND_3.  Moreover, the DMIND Application is also connected to Mor Prom communicative platforms like LINE Official Account and Facebook which can be accessed according to these steps: 

1. Select the Mor Prom line and add it as your friend by pressing https://bit.ly/2Pl42qo
2. Select the “talk to Mor Prom” (Chatbot)”
3. Select the menu for “Mental health check-up”
4. Start working on the questionnaire.

The launch of the “DMIND” application on Monday, June 20, 2022, at the Theatre on the first floor of the Library Building, Faculty of Medical Sciences, Chulalongkorn University.

• Facebook
• Twitter
• Line