Investigating intention to use as a mediator of associations between behavioural determinants and caregiver engagement with a parenting chatbot in South Africa

Imagine the world of public health as a giant, bustling library. For years, librarians have tried to get people to read important books about staying healthy, raising happy kids, and avoiding danger. They've built beautiful reading rooms, but many people can't get there because of long bus rides, busy jobs, or just not having the time. So, scientists started building a magical "digital library" that lives in your pocket—a phone app or a chatbot that sends you stories and advice right when you need them. This is the exciting corner of science called Digital Health Interventions.

But here's the tricky part: just because you have the library in your pocket doesn't mean you'll actually go inside and read. This is the puzzle of Engagement. Scientists have long believed that if you can make people want to use the app (their Intention) and if you can make the app feel useful and easy to use (their Behavioral Determinants), they will definitely log in and stay. It's like thinking, "If I tell someone a movie looks amazing and they have a ticket, they will definitely watch it." But does that always happen? Sometimes, life gets in the way, or the movie isn't as good as the trailer promised. This study dives into that gap between "I plan to do it" and "I actually did it," specifically looking at how parents in South Africa interact with a parenting chatbot.

---

The Story of the Chatbot and the Busy Parents

In the rural and semi-urban towns of South Africa, a team of researchers decided to test a digital parenting tool called ParentText. Think of this chatbot as a friendly, invisible mentor that lives on WhatsApp. It sends daily messages to parents, teaching them how to handle tricky situations with their teenagers, like how to talk about stress, how to spend quality time together, or how to keep their kids safe. The goal was to see if this digital mentor could help parents stay engaged, even when they were busy or didn't have perfect internet access.

The researchers gathered a huge group of 1,034 caregivers (mostly moms, about 94% of them) who had teenage girls between the ages of 10 and 17. Before the chatbot even started sending messages, the team asked the parents a bunch of questions. They wanted to know:

1. What makes you tick? (Did you think the app would be useful? Did you think it was easy to use? Did your friends think it was cool?) These are the Behavioral Determinants.
2. What do you plan to do? (Did you intend to use the app? Did you intend to spend your own mobile data money to read the messages?)

Then, they watched what actually happened. They tracked every single time a parent opened the app, finished a lesson, or did a "homework" activity with their teen.

The Big Surprise: Intentions Aren't the Magic Key

The team expected a specific story to play out. They thought the "good feelings" about the app (the determinants) would make parents plan to use it (intention), and that this plan would be the bridge that led them to actually using it (engagement). It was like expecting a map to lead a traveler to a treasure chest.

But the map didn't work the way they thought.

Here is what they found:

• The Good News: Parents who had positive feelings about the app (thought it was useful, easy, and supported by friends) did sign up and did use it more. If you liked the idea of the app, you were more likely to join the party.
• The Twist: The researchers thought that the plan to use the app was the secret sauce that made people actually do it. They thought, "If you intend to use it, you will use it." But the data said no. The "intention" didn't act as the bridge for the long haul. It didn't explain why some people kept using the app while others stopped.

In fact, the study found that intention did not mediate the relationship for sustained engagement. In plain English: Just because a parent said, "I plan to use this chatbot," didn't guarantee they would actually stick with it over time. The "plan" and the "action" were disconnected for the ongoing use of the app.

Why Did Some Parents Stay and Others Leave?

The study looked closely at two different types of "plans":

1. The Plan to Use: "I intend to open the chatbot."
2. The Plan to Spend Data: "I intend to use my mobile data to read the messages."

They found that having a plan to use the chatbot did help parents get through the front door (enrollment). If you said you wanted to use it, you were more likely to sign up.
However, staying in the room (sustained engagement) was a different story. The only thing that seemed to predict if a parent would keep reading modules and doing goals was their intention to spend mobile data. Since data costs money in South Africa, parents who were willing to spend their own cash on data were the ones who kept coming back.

But here is the biggest mystery the paper highlights: Even with all these plans and positive feelings, the researchers couldn't fully explain why people engaged.
The math showed that the "behavioral determinants" (like thinking the app is useful) and the "intentions" only explained a very small slice of the puzzle—R-squared values ranged from 0.6% to 2% of the reasons why people stayed or left. It's like having a weather forecast that predicts rain, but it only explains a tiny fraction of why the grass is wet. The rest of the story is missing. The paper explicitly concludes that the factors driving engagement are unclear and that the variation in engagement is likely influenced by unmeasured factors rather than the plans or attitudes the researchers measured.

The Takeaway: It's Not Just About "Wanting" to Do It

So, what does this mean for the future of digital health?

The paper suggests that we can't just rely on asking people, "Do you plan to use this?" and assuming they will. The gap between wanting to do something and actually doing it is real and stubborn. In this study, the "intention" didn't do the heavy lifting for long-term engagement.

The researchers realized that for parents in low-income areas, the biggest barrier might not be their attitude or their plans, but the reality of their daily lives. Maybe they ran out of data. Maybe they were too tired after work. Maybe the internet was too slow. The study hints that the "unmeasured factors"—the things the survey didn't ask about—are actually the ones driving the show.

The Bottom Line:
This study is a wake-up call for anyone building apps for parents. You can't just build a great app and hope people's good intentions will carry them through. You have to build apps that work even when the internet is spotty, that don't cost a fortune in data, and that fit into the messy, real lives of parents. The "intention" is just the starting line; the race is won by removing the obstacles that stop people from crossing the finish line.

The authors conclude that to really understand why people engage (or don't), we need to stop just looking at surveys and start listening to the stories of the parents themselves. We need to ask, "What actually stopped you?" rather than just "Did you plan to do it?" Because in the real world, plans are easy, but keeping them is hard, and the reasons why people stay or leave are often more complex than a simple survey can capture.