Researchers from Nanyang Technological University in Singapore (NTU Singapore) have discovered that mental health chatbots can help cure the symptoms of depression in people with depression by engaging them in sympathetic dialogues.
Computer programmes that mimic human discussions are known as chatbots or conversational agents. They are increasingly utilised in healthcare, for instance, to support overall wellness and to address mental health problems like sadness and anxiety.
According to a survey conducted in 2021 by Woebot Health, one of the top therapeutic chatbot firms in the US, 22% of adults have used a mental health chatbot, and 47% of them said they would be interested in using one if necessary.
This research, conducted by medical professionals from NTU’s Lee Kong Chian School of Medicine (LKCMedicine), is one of the first to analyse user-Chatbot conversations and assess their efficacy.
Nine mental health chatbots from popular app stores—five of which had at least 500,000 downloads—were examined by researchers to determine whether they provided self-help for depression sufferers.
Four of the chatbots—Marvin, Serenity, Woebot, and 7 Cups—are available for free while Happify, InnerHour, Wooper, Wysa, and Tomo need a subscription or one-time payment to use. Nine chatbots for mental health were included in the study.
The NTU research team assessed the chatbots using scripted user personas that were made to represent various cultures, ages, and genders. Additionally, the personalities displayed actions that show varied degrees of depression illness.
The results of this study, which were published in December in the Journal of Affective Disorders, showed that all of the chatbots interacted with users in an understanding and nonjudgmental manner and provided support and direction through exercises used in psychotherapy by psychologists and counsellors.
The researchers discovered that all chatbots maintained the confidentiality of the user’s personal information and did not transfer or store any of it through study of the app interfaces and their privacy policy legal statements. These details may include names, addresses, or chat histories that they may share in chat.
According to the World Health Organization, depression affects 264 million people worldwide and goes undiagnosed and untreated in 50% of cases. The COVID-19 epidemic has increased mental health issues, including depression, in Singapore.
The study’s principal investigator, Professor Josip Car, Director of the Centre for Population Health Sciences at NTU’s LKCMedicine, stated: “The COVID-19 pandemic has significantly increased the number of people affected by mental health issues, and there are still many stigmas associated with mental health disorders. Healthcare systems around the world are finding it difficult to keep up with the rising demand for mental health services. Chatbots and other digital health tools could help in giving prompt care to patients who may be unable or reluctant to see a doctor. We have demonstrated how chatbots are employed in this study and how they converse in a therapeutic manner.”
Chatting with chatbots to evaluate their efficacy
International study has indicated that chatbots may be able to assist people, but prior studies did not assess the interactions between chatbots and users.
The NTU team’s content study examined the level of personalisation, appropriateness in helping self-management in users with depression, and how they communicated empathy to users, as well as the quality and efficacy of the chatbots’ responses.
The study also observed how the chatbots assisted users in engaging in or completing activities that improved mood, how they tracked moods, and how they handled the danger of suicide.
All of the chatbots, according to the researchers, had a “coach-like” attitude that is upbeat, nurturing, and inspiring. Their investigation, however, revealed that while chatbots could have sympathetic dialogues with users, they were unable to offer tailored guidance. Future chatbots could benefit from this in-depth examination of the conversational flow, according to app developers.
According to the first author, LKCMedicine research fellow Dr. Laura Martinengo, “In order to preserve user privacy, chatbots may not be able to offer individualised advise or ask enough personal questions. However, for those in need, particularly those who cannot obtain medical assistance, these chatbots may still be a beneficial substitute. Some people find it simpler to communicate with a machine than a person.”
The researchers stated that further study is required to enhance chatbots for people at risk of suicide and to assess the long-term efficacy of chatbot-led interventions for mental health, even if chatbots may promote the self-management of depression and other mental health conditions.
To expand the scope, quality, and safety of their research into the efficacy of alternative digital approaches for mental wellbeing, the researchers will carry out more studies.
