Rahul Gaitonde

(Part 2 – What happens where you run a mental health service as a Valley-type startup)

Several years ago, I had had first-hand experience with poor mental health and sought help via in-person therapy and medication. During that time, I realized the following: 

• mental health help is binary: either there’s nothing or there are sessions with a mental health professional
• the financial cost puts it out of the reach of most people
• the time investment makes it difficult to balance work, home and this

At the same time, I had observed how my health had deteriorated over time, beginning with mild depressive symptoms with a decrease in drive and discipline. Today, I know that in others depression first manifested as increased anxiety. At that point, you know something’s not right, but doesn’t seem anywhere close to needing to see a psychotherapist. So you plod along until things begin to slide faster. By the time you seek help or someone does on your behalf, you’ve suffered quite a bit. 

Therefore, I understood that

• people needed – still need – something handy that didn’t warrant full-fledged medical care but provided some minimal level of support for those with mild symptoms. There are now several studies showing that a plurality of the population, especially urban, have poor mental health
• because people would likely end up being somewhat dependent on it, it needed to be low-overhead enough to be offered free. 
• and counterintuitively, it needed to be low-tech enough that its limits would be clear right away so that it would not disappoint later.

(Part 4 – My 2017 manifesto for such a service)

(Part 1 – the NYT investigates a mental health tech service for practices that aren’t in patients’ best interests)

Continuing our conversation about the NYT’s investigation into the practices at Talkspace, the remote mental health therapy app.

The problem stems from running a healthcare service like this as a business. Worse, a startup. Having spent the vast majority of my professional life at startups, I am all too familiar with the incentives to scale, to perform according to metrics, to employ ‘growth hacks’. When a healthcare startup employs these, the effects are much worse than a game or a social media app.

[the company] has questionable marketing practices and regards treatment transcripts as another data resource to be mined. Their accounts suggest that the needs of a venture capital-backed start-up to grow quickly can sometimes be in conflict with the core values of professional therapy, including strict confidentiality and patient welfare.

In 2015 and 2016, according to four former employees, the company sought to improve its ratings: It asked workers to write positive reviews. One employee said that Talkspace’s head of marketing at the time asked him to compile 100 fake reviews in a Google spreadsheet, so that employees could submit them to app stores.

When convenient, the company spins itself as a healthcare provider: “users can’t delete their transcripts, for example, because they are considered medical records.”, but those transcripts are used by the company for customer engagement and retention:

[A therapist on the platform] said that after she provided a client with links to therapy resources outside of Talkspace, a company representative contacted her, saying she should seek to keep her clients inside the app… “I was like, ‘How do you know I did that?’” Ms. Brennan said. “They said it was private, but it wasn’t.”

Finally, the disconnect between offering a private healthcare service and the exigencies of operating a startup is clear in the company’s public statements:

On Nov. 9, 2016, the morning after the election of Donald Trump, Mr. Frank wrote on Twitter: “Long night in NYC. Woke up this morning to record sales.” The Trump election tweets are examples of the sometimes unfiltered social media presence of Mr. Frank and Talkspace — an irreverence familiar among start-ups but unusual among organizations devoted to mental health care.

In 2016, a man named Ross complained on Twitter that the company’s subway ads “were designed to trigger you into needing their services.” Talkspace’s official Twitter account responded, “Ads for food make people hungry, right?” and added, “I get what you’re saying, Ross, but medical professionals need people to buy things.”

The problems of misincentives of private healthcare are well known and well debated, though nowhere close to being resolved. And startups get a lot more attention than a more traditional healthcare provider would. And Talkspace may have begun with noble ambitions. Regardless, what it has built isn’t anywhere what I think mental health care tech should look like.

How it approaches conflict between what is good for itself versus what is good for its customers – patients – diminishes trust. When it come to data privacy, to marketing, to its incentives for therapists, its responses indicate that it will choose its financial health over people’s mental health. Unfortunately, it’s likely similarly funded venture-backed tech companies in healthcare have similar conflicts and pressure.

(Part 3 – How I’d approached a simple mental-health chat service)

Earlier in August, the NYT published a longform investigation into the data and marketing practices of the mental health therapy provider Talkspace. It hit home for me.

So many of the NYT’s discoveries dismayed, even horrified me. I have sought mental health therapy myself several years ago, and about three years ago contemplated working on a different approach for a text-based service for people with anxiety and mild depression. We will talk about the specifics some time later. However, I thought in detail about, and spoke to mental health professionals about the technology, the ethics, the right audience, among other such issues. The way I thought about all of these was so very different from Talkspace.

The first is the way the service is marketed to potential customers and to therapists:

Talkspace is advertised to users as unlimited, “24/7” messaging therapy. “Your therapist will see your messages and respond to you throughout the day,” the company says. Therapists get a different pitch: “Set your business hours, and check in on your clients daily, five days per week.”’

This immediately reminded me of the worst of Groupon, which pitched to users that they’d never have to pay full price for anything again, while pitching to businesses that a single one-time mass discount would convert to loyal users willing to pay, well, full-price.

But Talkspace went beyond, adding this:

Talkspace introduced a new feature: a button that users could press after sending a message that required the therapist to respond within a certain time frame. If the therapists don’t respond in time, their pay can be docked.

Some therapists on the platform were alarmed, in part because the function required them to work on demand, rather than on their own schedule. More significantly, they asked: Is it harmful to give clients with anxiety and boundary issues a button to press for immediate gratification?

First, this reminded me of the worst of Uber. I have been told my more than one driver that they could only turn down a limited number of rider assignments a day, and had to be online at least a certain number of hours a day. After the rollout of the feature where drivers are assigned their next ride towards the end of the ongoing one, those drivers said they couldn’t even take time out to visit the restroom, leave alone lunch and snack breaks, because they also had to reach the next pickup in the time that the Uber app had estimated for the waiting rider.

Troubling as this is, these constraints are worse in the context of mental health because they introduce added anxiety for both the therapist and the patient, the former needing to be equanimous and the latter already suffering from some anxiety/depression.

(Part 2 – more about conflicts)