In 2024, MQ Mental Health Research created an opportunity, developed and funded by Wellcome, to support researchers outside psychiatry, psychology, and neuroscience to apply bold and novel ideas and methods from their discipline to mental health science. In this Transdisciplinary Research Grant series, we will go through each study, and explain why the results are vital to the future of mental health interventions.
Perinatal depression and anxiety (PDA) are mental health conditions that affect mothers during pregnancy and after childbirth. For women living in low- and middle-income countries (LMICs), these problems are often ignored because there aren’t enough resources, people don’t talk about mental health due to stigma, and a lack of awareness. Over the years, there has been growing interest in how artificial intelligence (AI) could identify, diagnose, and even treat these conditions. However, research on using AI to deal with PDA in lower-income countries is just starting, so we still don’t know.
What Did the Research Aim to Do?
A recent study by Dr Uchechi Shirley Anaduaka, Dr Ayomide Oluwaseyi Oladosu and Samantha Katsande analyzed over 2,200 articles published between 2010 and 2024, narrowing down the research to 19 key studies that focus on AI and PDA. They focused on 19 important studies about AI and PDA. Most of these studies happened in developing countries, especially in South and South-East Asia. China and Nepal had the most studies, with four each, and Kenya was the only Sub-Saharan African country included. Interestingly, most of the studies were published in the last five years, which shows this is still a new and growing field.
The studies examined women at different points during pregnancy and after giving birth. Some of the studies focused on pregnant women, while others followed mothers up to three years after giving birth. Most volunteers were recruited from hospitals or maternity clinics; their ages ranged from 15 to 54 years. Many of the studies assessed for depression or anxiety, sometimes more than once, such as during pregnancy and a few months after giving birth.
One of the most promising aspects of AI in this field is its ability to help identify women at risk of depression and anxiety. Using supervised machine learning, a form of artificial intelligence that lets computers learn and get better at tasks automatically, researchers have trained AI systems to recognize patterns, such as certain socioeconomic factors that might predict mental health outcomes. While these systems show potential, they have primarily focused on postpartum depression, leaving gaps in other areas. Additionally, tools like AI chatbots, designed to provide support or treatment, remain largely unexplored in LMICs. One exception is the “Zuri chatbot,” which offers Kenyan mothers access to mental health coaching and advice through Facebook Messenger and text messaging. However, studies examining such tools remain rare.
Where To Go From Here?
Although AI has a lot of potential, there are rising concerns about privacy and ethics. Anaduaka et al found that many studies didn’t clarify whether participants knew how their data would be used or if they gave proper permission. This raises the question of whether they would have agreed to share their information if they fully understood how it would be used.
The researchers also discovered that several studies using chatbots didn’t explain who could see the participants’ data or how it would be protected. Without clear guidelines, this can lead to issues like data leaks or misuse, which could harm the people these tools are supposed to support.
Where To Go From Here?
AI could improve mental health care for mothers in LMICs, but more research is needed. Anaduaka et al point out that researchers should look into more advanced tools like chatbots while making sure ethical standards, like consent and privacy, are followed. As research progresses, it’s important to include the experiences of women and communities affected by postpartum depression to create helpful and respectful solutions.
