Support the news
It's unlikely that your therapist will be replaced by a computer program anytime soon.
That's the takeaway of recent study out of Britain looking at the effectiveness of computer-assisted therapy for depression.
The bottom line: The computer programs offered little or no benefit compared to more typical primary care for adults with depression. That's largely because the patients were generally “unwilling to engage" with the programs, and adherence faltered, researchers conclude, adding that the study "highlighted the difficulty in repeatedly logging on to computer systems when [patients] are clinically depressed."
In an accompanying editorial, Christopher Dowrick, a professor of primary care medicine at the University of Liverpool, stated what may seem obvious: Many depressed patients, he wrote, don't want to interact with computers; rather, "they prefer to interact with human beings." He noted that the poor result “suggests that guided self help is not the panacea that busy [primary care doctors] and cost conscious clinical commissioning groups would wish for."
As part of the study, published in the BMJ, 691 patients suffering from depression were randomly assigned to receive the usual primary care, including access to mental health care, or the usual care plus one of two computer-assisted options that offer cognitive behavior therapy (CBT), a form of therapy that encourages patients to reframe negative thoughts. Patients were assessed at four, 12 and 24 months; those using the computer programs (one called "Beating the Blues" and the other "MoodGYM") were also contacted weekly by phone and offered encouragement and technical support.
The context of all this is that demand for mental health services generally exceeds supply around the globe, and health systems are seeking ways to bridge the gap. According to the new paper, demand for cognitive behavioral therapy, for instance, "cannot be met by existing therapist resources." So, the thinking goes, maybe a computer can ease some of the caseload. And in some cases, it works. Indeed, Britain's National Institute of Health and Care Excellence (NICE) guidelines recommend computerized CBT as an "initial lower intensity treatment for depression...." based on studies that showed it can be effective.
However, results of this latest study may nudge clinicians and policymakers to rethink the computer's role in therapy.
Here are the results, summed up in BMJ news release:
Results showed that cCBT offered little or no benefit over usual GP care. By four months, 44% of patients in the usual care group, 50% of patients in the Beating the Blues group, and 49% in the MoodGYM group remained depressed....
...only 18% of patients completed all eight sessions of Beating the Blues, and 16% completed all six sessions of MoodGYM.
Almost a quarter of patients dropped out of the study by four months...
“Participants wanted a greater level of clinical support as an adjunct to therapy, and in absence of this support, they commonly disengaged with the computer programmes,” explain the authors.
In his editorial, Dowrick asks: "How much human guidance is needed? It is perhaps not surprising that a few minutes of telephone contact with a technician, however well trained, did not have appreciable effects on participants’ depressive symptoms....When considering how to help patients presenting with mild to moderate depressive symptoms, GPs have many evidence based interventions to choose from. They should remember that therapeutic benefit is derived from their own contact with their patients, especially if they convey a sense of hope and optimism and establish a positive relationship with the patient."
From Face To Face, To Split Screen
This is not to say there's no place for technology in therapy. Some mental health providers say that a sort of hybrid approach, with a combination of live interaction and an online component, can be extremely effective.
D'Arcy, a 48-year-old mother in Hopkinton, Mass., suffered from high levels of anxiety and stress, aggravated by a variety of health conditions that, in one year alone, required six operations.
Things got so bad, she said, that she developed an intense reaction to ambulance sirens: Every time she heard one, she hyperventilated and her body froze in place with fear.
She didn't seek counseling, though, until the summer of 2013, when a doctor treating her broken wrist commented that she was particularly high strung.
D'Arcy was referred to cognitive behavioral therapist Ana-Maria Vranceanu, an associate professor in psychology at Harvard Medical School and a clinical psychologist at Massachusetts General Hospital. But D'Arcy's drive to Boston's MGH from her home was grueling, requiring her to spend four hours in the car for a 45 minute therapy session. So, after traditional CBT for 12 weeks, she switched to virtual therapy, and now all of her visits are done on her iPad.
"It's a split screen — half the screen is her, half the screen is me, and we talk," said D'Arcy, who asked that her full name not be used. (She also gets homework assignments at the end of each session.) "I thought I'd get distracted by the stuff around me, but it doesn't happen at all. I actually think I'm more focused because we're just looking at each other. And it's less stressful because I don't have to drive."
Now D'Arcy only connects with her therapist every third week. "I really like the tele-therapy thing, but my personality would require creating a trusting relationship with the therapist first. I wouldn't just automatically trust someone that I only met online. My teenagers might be different, though. They're much more willing to use technology as a primary vehicle for doing everything. Face-to-face means less to them right now."
Dr. Vranceanu, D'Arcy's therapist, says she's had "great success" with an Internet-based program called Vydio, used in the MGH Telepsychiatry program, which allows face-to-face sessions as well as the ability to exchange documents. "My patients love it, and adherence to treatment is almost perfect," she wrote me in an email. "Generally there is growing support that therapy over the Internet works just as well as face-to-face therapy, and adherence rates are higher. So we know for sure that being in the same room with a patient is not necessary for adherence and efficacy."
Regarding the British study, Vranceanu writes:
Things are trickier with cCBT platforms. I think cCBT is useful for teaching general CBT skills that are useful in daily life (e.g., understanding that perceptions are skewed, restructuring negative automatic thoughts, learning about cognitive errors, engagement in pleasurable and mastery activities, accessing social support, etc). cCBT can also be very useful for teaching and monitoring adherence to a medical regimen. However, when dealing with mental health issues like clinical depression or anxiety, symptoms such as lack of motivation, low energy can make it really hard for someone to follow through on their own with a protocol.
Further, depression can encompass difficulties with concentration, and so this can make some of the CBT concepts difficult to grasp. There are studies that show that cCBT can help with mild and moderate depression, but some meta analysis and meta regression has suggested that these effect sizes may have been overestimated...
It seems to me that a lot of work needs to be done to understand whether there are some patients for whom cCBT works, based on type of symptoms and personality. The busy highly educated VP with mild depression may do well with cCBT while the moderately depressed socially isolated student may need the therapeutic relationship to instill motivation for learning and adhering to CBT, and might drop out or not adhere to cCBT. In sum, I think the inconsistency in research findings and difficulties with adherence are a clue that we need to look at person specific factors that predict adherence.
I think we can't discount the therapeutic alliance. Many decades ago it was found to be the active ingredient in all therapies. However, CBT types of therapy rely on the [therapeutic alliance] much less than other forms of therapy such as psychodynamic or interpersonal. I often describe CBT as taking a class. Further, going along with my point above, some patients may need the face to face contact more than others. Generally, I think cCBT could be very helpful for prevention of depression in patients with family history and highs stress. For patients in primary care, cCBT supplemented by phone calls or with report back to the PCP (though email, etc) can help a patients (with good relationship with the PCP) develop the motivation necessary to adhere to treatment. A model of combined face to face or Skype like CBT sessions followed by cCBT may also be successful.
Indeed, it may come down to the specifics of the patient, and his or her particular diagnosis and demographic. For instance, studies have found that age can make a difference, with younger people finding technology-based treatment more effective compared to older adults.
Joseph Greer, a clinical psychologist at MGH's Cancer Center, reiterates that due to the very nature of depression, characterized by a lack of energy and motivation, using technology can be a challenge. "So, perhaps this isn't going to be effective for people with extreme depression, but it may be better for mild or moderate depression," he said in an interview.
However, he said, technology could be a great resource for people who would never go see a therapist or access mental health care otherwise. Or it can be used as the first step in a therapeutic program: if the person doesn't respond initially, there might be followup with face-to-face interventions, he said.
Greer is in the process of testing a self-administered mobile app that offers cognitive behavioral therapy to advanced cancer patients suffering from anxiety. "We wanted to use technology as a way to reach more people," Greer said. "But the question still remains, is adherence going to be a problem?"
Because whether it's anxiety, depression or some other disorder, Greer said, treatment tends to be better "when there's a human being involved — the research shows that...A human, in real time, can tailor the intervention to the patient's specific needs — that's much more difficult with technology. You can do your best to make an algorithm with some tailoring in it, but that's very hard to create, and at some point you can only go so far in that tailoring logic. When there's a person involved, that tailoring is an essential part of of the job."
Support the news