From the Editor
More and more organizations use AI; today, a chatbot might assist you in ordering a pizza or tracking a package. But could a chatbot help our patients find the mental health care that they need? Could it help self-identified members of ethnic groups – who historically do less well in getting services – with access?
Johanna Habicht (of Limbic) and her co-authors try to answer these questions in a new study for Nature Medicine. They looked at the use of a chatbot for self-referral against the standard option in the UK’s NHS when patients seek psychological care. The resulting multisite observational study involved almost 130 000 people. They found that AI increased referrals (especially, in terms of diversity). “Here we demonstrate that digital tools can reduce the accessibility gap by addressing several key barriers.” We look at the study and mull its implications.
As we move past the pandemic, we ask: is virtual care routinely offered for mental disorders? In the second selection from JAMA Health Forum, Jonathan Cantor (of the RAND Corporation) and his colleagues consider mental telehealth – or telepsychiatry, to use the older term – in the United States. With a secret shopper approach, trained callers phoned more than 1 400 US clinics, posing as potential clients with mental health problems. They found most offered virtual care. Further: “There were no differences in the availability of mental telehealth services based on the prospective patient’s clinical condition, perceived race or ethnicity, or sex.”
And, finally, we explore the latest in the news with recent articles from The Guardian and The New York Times. Among the topics: the mental health struggles of rising political star Lina Hidalgo, privacy and mental health apps, and help for those with schizophrenia and homelessness in Cameroon.
DG
Selection 1: “Closing the accessibility gap to mental health treatment with a personalized self-referral chatbot?”
Johanna Habicht, Sruthi Viswanathan, Ben Carrington, et al.
Nature Medicine, 5 February 2024
Mental health is a global health priority as recognized by the World Health Organization with disorders such as anxiety and depression affecting 29% of the global population in their lifetime… The burden can be alleviated with adequate support, yet access remains limited due to structural issues such as underfunding and understaffing. Moreover, not everyone experiencing mental health problems seeks help, or they delay seeking it for years, resulting in unmet needs and inadequate support at the right time.
The first step of many mental healthcare pathways is for individuals to seek help and be referred to the appropriate healthcare service. Ensuring easy and inclusive access at this initial step regardless of an individual’s socioeconomic status, ethnicity, gender or other such factors underpins a fair and equitable healthcare system. And evidence suggests that individuals from minority groups often face higher barriers and stigma in accessing care…
Digital technologies and artificial intelligence (AI) have been proposed as potential remedies to these challenges. While we and others have found evidence that digital technologies can reduce the workload of mental healthcare staff and make services more efficient, less is known about the marginal impact of digital technologies in supporting individuals of differing demographics seeking help…
So begins a paper by Habicht et al.
Here’s what they did:
- They conducted a multisite observational study.
- They used real-world data from the UK – “a suitable environment for testing the self-referral chatbot due to its standardized referral process and a high percentage of self-referrals to NHS Talking Therapies.”
- Participants were given the usual option or Limbic Access, an AI program “which can optimize the standard referral process by autonomously gathering patient information to inform suitability for the service and initial presenting problem…”
- Outcomes: patient referral volume and diversity in ethnicity, gender, and sexual orientation.
Here’s what they found:
- 129 400 people were followed.
- Number of referrals. “We found that services that used this digital solution identified substantially increased referrals (15% increase versus 6% increase in control services).”
- Gender and referrals. “Referrals increased for females and males by 18% and 16%, respectively, which was higher than 6% and 5% for matched services (females: OR = 1.10…; males: OR = 1.12…).” Also, “referrals from individuals who identified as nonbinary increased by 179% in services that used the personalized self-referral chatbot, compared to a 5% decrease in matched control services…”
- Ethnic minorities and referrals. “We identified a 29% increase in referrals for ethnic minorities in the services that used the personalized self-referral chatbot, which was significantly higher than the 10% increase in the matched control services (OR = 1.18…).” See figure below.
- Qualitative feedback. Drawing on 42 332 individuals: “89% of the free-text feedback was classified as positive, 7% neutral and 4% negative.” The human-free nature of the chatbot was mentioned.
A few thoughts:
1. This is an impressive study, drawing on a real-world application with a robust dataset, and published in a major journal.
2. The findings in a sentence: the chatbot led to increased referrals, particularly among those who identify as belonging to ethnic minority groups (up 29%). The authors quantify the difference: “∼33,000 additional patients have found their way into treatment owing to this technical solution.” (!!)
3. Why the difference? “Individuals can refer to the service outside of the usual hours and take the necessary time to complete the referral process without feeling rushed while at the same time being encouraged through the empathetic chatbot responses. Another advantage is that individuals can more comfortably discuss their mental health issues without facing potential communication difficulties with a human and the feelings that they might be judged while still providing empathetic responses.”
4. What was involved in the AI experience? That’s difficult to understand without trying the chatbot (which I haven’t), but the authors describe a tailored experience based on symptoms and the use of scales.
5. This is a British study involving a British mental health service. Still, it’s possible to see the use in other systems and suggests that AI may have a significant future in clinical service delivery.
The full Nature Med paper can be found here:
https://www.nature.com/articles/s41591-023-02766-x
Selection 2: “Availability of Mental Telehealth Services in the US”
Jonathan Cantor, Megan S. Schuler, Samantha Matthews, et al.
JAMA Health Forum, 2 February 2024
Telehealth utilization in the US expanded considerably during the COVID-19 pandemic, enabled by changes in federal and state policies pertaining to financial reimbursement of these services. This shift was particularly pronounced for mental health services: the number of outpatient treatment facilities providing telehealth and the proportion of visits that were conducted using telehealth rose substantially…
During the pandemic, telehealth utilization rose and then returned to nearly prepandemic levels in most fields of medicine; however, it remained much higher than prepandemic levels in mental health care. Although studies have evaluated utilization of telehealth throughout the pandemic, availability and composition of telehealth services remain largely undocumented including ease of access to appointments, mental health conditions treated, types of telehealth services offered, and accepted types of insurance coverage…
So begins a paper by Cantor et al.
Here’s what they did:
“Cross-sectional analysis of a secret shopper survey of mental health treatment facilities (MHTFs) throughout all US states except Hawaii from December 2022 and March 2023. A nationally representative sample of 1938 facilities were contacted; 1404 (72%) responded and were included. Data analysis was performed from March to July 2023.”
Here’s what they found:
- 1 404 facilities responded, of which 1 221 facilities (87%) were accepting new patients.
- Clinic description. Among those accepting new patients, most offered only outpatient services (80.3%), were private, not-for-profit (66.9%), and were in metropolitan areas (81.7%).
- Telehealth. 80% reported offering telehealth. Of these, 97% reported availability of counseling services; 77%, medication management; 69%, diagnostic services.
- Clinical condition. “Telehealth availability did not differ by clinical condition.” See figure below.
- Private vs. public. Private for-profit and private not-for-profit were more likely to offer telemental health (adjusted odds ratios of 1.75 and 2.20, respectively).
- Patient demographics. “No differences were observed in availability of an appointment based on the perceived race, ethnicity, or sex of the prospective patient.” (!)
A few thoughts:
1. This paper offers good data on mental telehealth in the United States.
2. No real surprise here: the pandemic fundamentally changed mental health services; the vast majority of US clinics offer some sort of virtual care.
3. Interestingly, diagnosis didn’t matter nor did ethnicity. Past work suggests that both may make a difference – but that wasn’t found in this study.
4. The dataset is good but not amazing. One in five clinics didn’t even respond to multiple attempts at contacting them. As well, there are limitations to the “secret shopper” approach since clinics may say one thing on the phone and handle a referral differently, obviously.
The full JAMA Health Forum paper can be found here:
https://jamanetwork.com/journals/jama-health-forum/fullarticle/2814605
In the News
Part of an occasional series.
“A Rising Democratic Star Shares Her Mental Health Journey”
J. David Goodman
The New York Times, 6 February 2024
“As a rising young Democratic star and the top elected official of Harris County, the most populous in Texas, Lina Hidalgo surprised many people last summer when she announced that she had checked herself in at a residential mental health clinic for serious depression.
The article describes the mental health problems of this young politician who speaks candidly about her suicidal thoughts and her decision to get care. She does also mention her sexual assault.
“Ms. Hidalgo, 32, has added her name to a growing list of politicians – most of them Democrats – who have chosen to be public about their mental health. She benefited from such openness herself, she said, taking inspiration from Senator John Fetterman of Pennsylvania, who announced that he was receiving inpatient treatment for depression several months before Ms. Hidalgo sought that kind of care.”
The article wonders if her candor will prove to be problematic in the future. Why then speak out? Hidalgo explains: “The more we talk about it, the more it’s going to help somebody else.” Her resolve is impressive.
https://www.nytimes.com/2024/02/06/us/lina-hidalgo-harris-county-texas-mental-health-treatment.html
“‘They thought they were doing good but it made people worse’: why mental health apps are under scrutiny”
David Cox
The Guardian, 4 February 2024
“‘What if I told you one of the strongest choices you could make was the choice to ask for help?’ says a young, twentysomething woman in a red sweater, before recommending that viewers seek out counselling. This advert, promoted on Instagram and other social media platforms, is just one of many campaigns created by the California-based company BetterHelp, which offers to connect users with online therapists.”
So begins a long, detailed essay about BetterHelp, which was fined millions of dollars last year by the FTC for selling private data to third parties. Unfortunately, other apps do the same: in a review of 32 leading mental health apps, 19 failed to protect patient security and privacy. The article notes that these apps have become big business, predicted to be worth $17.2 billion (US) by 2030. But are they effective? The article describes a trial involving 19 000 people with some randomized to using a DBT app; these individuals had more self-harming than those who didn’t use the app.
The writing here is excellent and considers options, including better regulation of apps. As more and more of our patients look to apps, this article is a must read.
“‘Love is our first medicine’: treating mental health in Cameroon’s unique refuge”
Josiane Kouagheu
The Guardian, 8 February 2024
“Eloisa Pentecotisa was begging on the streets of Yaoundé, hearing voices and eating out of bins, when a team of health workers came across her and suggested she go with them. The 28-year-old had no idea how long she had been sleeping rough in Cameroon’s capital.
“But with severe mental health problems and no immediate family to take her in, she received threats, abuse and risked contracting diseases such as cholera. Acutely mentally ill people like Pentecotisa are often rejected by their families, their conditions exacerbated by being left to roam the streets alone.”
The article discusses Pentecotisa – who becomes a resident of Le Village de L’amour (the Village of Love), receiving care and medications. There, a team of volunteers including physicians tends to her needs and those of another 100 or so people who are homeless and suffering from mental illness (most have schizophrenia). Pentecotisa has had a difficult life: orphaned and then passed from family member to family member before becoming homeless; on the street, she was often mocked, leaving her wanting to die. It notes that she now has “a lot of dreams.”
The article touches on the incredible problems of mental health services in a country that has just 12 psychiatrists for a population of 28 million. As is so often the case when reading about mental health care in low-income nations, we are left realizing the depth of the challenges but, at the same time, the creativity of those in health care.
Reading of the Week. Every week I pick articles and papers from the world of Psychiatry.
Recent Comments