Creating Happy Minds - In Conversation
During this unprecedented time, stress and other kinds of emotional upheaval have been a common and natural response. With each of us left with limited human interaction and access to facilities of even daily requirements, seeking out psychological support has been extremely tough. While initiatives such as Covid19 helpline India are helping those in need, there have also been recent developments in the tech sphere, especially under the umbrella of Digital India. One such endeavor is HappyMinds, a mobile app aimed at increasing access to support through telecounseling. Talking of the idea and technology behind this initiative, Shefali Mehra and Yastika Shetty are in conversation with the developer and co-founder, Shashwat Aggarwal.
What is ‘HappyMinds’ and why has it been created?
Happy Minds is a mobile application that can be used as a CRM by mental health practitioners to manage their existing clients, provide counseling to new clients and get data driven insights. The application addresses specific problems faced by psychologists and users. Our mission is to make mental healthcare more accessible and improve the efficiency of a tele-counselling session using advanced technologies.
What degree of privacy and confidentiality can the users of mental health apps be provided with, with respect to the current laws in place in India?
This depends on the company. Currently, the laws in India for mental health practitioners are ambiguous. Psychologists other than clinical psychologists do not get any registration number from any government authority. Due to lack of such regulation, some tele-health platforms in the area of mental health might record and store the user interactions with a psychologist for supervision purposes in order to make sure that the quality of service is not compromised. However, certain other technology platforms do not record/store the user’s interactions with a mental health practitioner. In both cases, appropriate security measures are taken to ensure data safety.
In your application HappyMinds,
Is the saved data anonymized?
Is it passed on to any other entity?
Is it removed from the database after a period of time?
Privacy of the user is the core value on which we have built our application. For the same, we have followed certain aspects of the National Health Stack. The user and the psychologist are the sole owner of the data. Certain data is not even transferred to our servers and remains on the smartphone itself. Relevant security measures are taken for securing any data that is stored on the server.
What accuracy and f1 should mental healthcare models strive to achieve? ( For readers, F1 score is the weighted average between precision and recall values; best being 1 and worst being 0. It is a better metric in real-life classification than accuracy.)
The accuracy of the algorithms will differ based on the data that the algorithm deals with. For example the accuracy for the insights provided based on the english words would be relatively lower compared to the insights provided for any other language. This would again decrease in case of verbal insights based on voice owing to the broad spectrum of accents that people from different parts of the country have. However, there should not be any benchmark on the accuracy and we should continue to strive to maximize the accuracy. It is important to incorporate a feedback mechanism to increase accuracy over time. If interested, readers can read about the initiative taken by Sonde Health to improve the accuracy of voice diagnosis in the Indian setting.
How are mental health professionals regulated on the platform of mental health apps in order to verify their professional background and ensure that the ethical standards are followed while dealing with clients?
The qualifications, years of experience and place of work are some of the parameters on the basis of which an eligibility criteria can be set for any mental health practitioner that wishes to join a tele-health platform. Apart from this, some companies might have their own short training program during as well as after onboarding mental health practitioners. However, this is still a challenge for mental health apps and appropriate laws need to be set in place. This is a problem that can be solved by technology. Technology can intervene in a secure manner to provide a “quality” of any tele-health session.
Mental health, like physical health, encompasses a varying range of severity. Today with algorithms for diagnosing the onset of psychosis being developed, what severity of mental illness, within its vast range, should mental health apps take upon themselves to cater to?
Technological strides have been made in diagnosing severe mental illnesses in a remote setting. The words used by the user can help us get an idea of the thought process of the user and the same can be used for diagnosing the mental illness of the user. However, still mental healthcare apps in many countries are not capable of predicting mental illness irrespective of its severity. The reason being, most of the dataset available in this domain is restricted to specific languages and accents. Using the existing datasets for the Indian setting can be difficult because of the diversity in languages in our country and the accuracy with which the words said by a user are identified by the algorithm. Moreover, such technologies, if adapted in the future need to be used as an assistive technology to help psychologists rather than a replacement to psychologists.
These apps use algorithmic techniques such as natural language processing to classify health concerns, and identify key terms to give particular diagnosis. Can this place a stigma on the individual, as opposed to real-life counselling which might avoid labelling but focus on well-being?
Yes and no. One of the biggest pain points in mental illness is that it is intangible. This is one of the biggest problems that technology can solve. For any physical illness, checkups and tests help the user to identify that they have physical illness. Using NLP, platforms can actually help a user identify a sort of their “mental health score”. If this “report” is shown directly to the user then it can cause more harm than any good. However, it can be used to understand and track the user’s condition relatively more accurately within a period of time.
This in turn can be used to suggest a relatively more accurate solution to the user, thereby improving the efficiency of measures that are suggested to the user to improve her/his well-being.
Can misdiagnosis be a bigger problem on such apps, as opposed to physical settings (for instance, because of language barrier)?
Owing to the diversity and language and accent in our country, misdiagnosis is definitely an important issue. Such algorithms should therefore be implemented very carefully and the existing datasets need to be augmented to suit the local setting.
However, remote diagnosis, if implemented efficiently, can be a boon for people who do not want to visit a therapist for diagnosis as well as for mental healthcare practitioners. Currently, there is an acute shortage of mental health practitioners. Due to this, it is difficult for a mental health practitioner to identify all the cues of the user in a remote setting. In fact, a lot of effort is put in by the mental health practitioners during each session and this affects their mental health as well. We are working on providing an assistive technology to the mental health practitioner so as to help them in a tele-counselling session by providing them more insights about the user. This is done only after seeking explicit content from the user. However, care should be taken for not using these technologies as a replacement to the mental health practitioners where the user is empowered for self diagnosis.
In the beginning, HappyMinds application would be available in English alone. However, the psychologists registered on the application would be fluent in multiple languages and the users could filter the psychologists based on the language which they are comfortable speaking in.
What initiatives do you think are required for mental healthcare in the online sphere in the future?
One of the major challenges in leveraging these technologies to revolutionize mental healthcare space is the lack of open data in the Indian setting. I believe that companies in the mental healthcare space need to collaborate with each other to launch a national level initiative aimed at curating anonymised open voice data. The concept of open data needs to be adopted to empower any innovator in the country to build their technologies on top of it and transform mental healthcare space. Apart from mental health, the same data can be used for multiple other use cases.
This is a great step towards helping the community. Tell us a little about your hopes and aims for the society with the launch of your application.
Our mobile application, HappyMinds, will be released in August across Android and iOS platforms. Our mission is to empower mental health practitioners with all the technological capabilities so that the efficiency of tele counselling sessions can be increased. The app would help users to connect with psychologists via different modes of consultation and also track their mental health progress. We aim to equip the mental health practitioner with all the technological capabilities at a nominal monthly fee. The core value on which we have built our application is the privacy of the users interaction with the mental health practitioner and appropriate measures have been taken for the same.