Digital mental health interventions are a breakthrough in psychological care. Programs delivered through smartphones and computers can now teach cognitive skills, reduce symptoms or anxiety, and reach people who might never seek traditional treatment. Research shows that many of these digital interventions do work, but they don’t work equally well for everyone.
Rather than simply asking whether digital interventions are effective, researchers are increasingly trying to understand a more informative question: Why do some people benefit more than others?
Our recent study examined a digital intervention designed to promote more positive interpretations of ambiguous social situations. The findings suggest that two factors, namely how flexible people are in updating their interpretations and how well they tolerate uncertainty, may help explain differences in who benefits most from these tools.
Anxiety often lives in the meaning we assign to events
Many of the situations that trigger anxiety are not clearly negative. They are ambiguous. Think about everyday situations, such as when someone does not immediately respond to a message, when a person’s facial expression is difficult to read, or when a colleague walks past without saying hello. These situations require interpretation. People’s minds must answer a simple question: What does this mean?
For people vulnerable to anxiety, the answer often tilts toward the negative: “I must have done something wrong,” “They’re upset with me,” or “I didn’t perform well.”
This tendency to assume threatening meanings when situations are ambiguous is referred to as negative interpretation bias. This bias can maintain anxiety by making everyday situations feel more threatening than they actually are.
Can interpretation habits be retrained?
One promising approach involves digital cognitive training that specifically targets interpretation patterns. These interventions typically present users with everyday ambiguous scenarios and guide them toward more balanced or benign interpretations. The goal is not to promote unrealistic positivity, but to help people consider multiple possible meanings rather than automatically defaulting to the worst one. This reflects an important psychological principle. Our thinking patterns are habits. Like other habits, they can change with practice.
At the same time, changing thinking patterns is not easy. Success may depend on the cognitive tendencies people bring into the intervention.
Not everyone starts from the same cognitive starting point
Our study examined two characteristics that may help explain why people respond differently to digital interpretation training. The first was interpretation inflexibility, or how difficult it is for someone to revise an initial negative impression.
People naturally form quick interpretations. But while some people readily update these interpretations when new information becomes available, others find that their first impression tends to “stick.” For example, if you initially think someone reacted negatively to something you said, do you easily adjust that assumption if they later behave warmly? Or does your first impression remain dominant?
Our findings showed that individuals who had more difficulty revising negative interpretations tended to show smaller benefits from the digital intervention. This does not mean the intervention cannot help these individuals. Rather, it suggests they may need more practice, different pacing, or additional strategies aimed at strengthening cognitive flexibility.
The challenge of living with uncertainty
The second factor was intolerance of uncertainty. This is how stressful someone finds not knowing what will happen. Some people can accept uncertainty as part of life. Others experience it as uncomfortable and feel a strong need to resolve ambiguity quickly.
Yet, digital interpretation training asks people to remain open to alternative possibilities. For someone who finds uncertainty especially distressing, this openness may feel difficult at first. In our study, individuals with higher intolerance of uncertainty tended to benefit somewhat less from the training. This suggests that helping people gradually build tolerance for uncertainty may sometimes be an important step in maximizing the benefits of cognitive interventions.
Toward more personalized digital mental health interventions
These findings reflect a broader shift happening across mental health research. The field is moving beyond asking whether interventions work on average and toward understanding individual differences in response. The future of digital mental health may not just be about building effective tools, but about building adaptive tools and identifying what works best for whom.
We may eventually see digital interventions that adapt to a user’s level of cognitive flexibility, include modules that specifically target uncertainty tolerance, adjust pacing based on individual progress, and identify when additional support may be helpful.
These findings highlight the importance of moving beyond average treatment effects toward understanding predictors of treatment response. Identifying factors such as interpretation inflexibility and intolerance of uncertainty may help explain variability in outcomes and inform the development of more targeted digital interventions.