The Word We Never Defined

We talk about digital wellbeing constantly, and we have almost never agreed on what it is. We measure technostress here, problematic internet use there, sleep loss somewhere else, each in its own study with its own scale. What if the reason the advice never quite lands is that we have been treating a whole as a pile of parts.

A scoping review published in BMC Medical Education in May 2026 makes a quietly radical observation about the entire field: most research into digital wellbeing examines related constructs, such as technostress, problematic internet use, or digital addiction, in isolation, rather than conceptualising digital wellbeing as a multidimensional and integrative construct [1]. That sentence sounds like academic housekeeping. It is closer to an admission that we have been mapping a coastline one rock at a time and calling it a country.

Why does this matter to you, sitting with a phone that has interrupted you three times since you started reading. Because the shape of the question determines the shape of the answer. If your problem is defined as screen time, you download a screen-time tracker. If it is defined as addiction, you look for abstinence. If it is defined as technostress, you look for workplace boundaries. Each frame produces a genuine intervention, and each quietly assumes the others do not exist. The review's argument is that digitalisation has changed how people interact with themselves, each other, and their environment, and that these effects are not separable line items [1]. Wellbeing is the thing that emerges across them.

The cost of the fragmented approach shows up when you try to compare treatments. A separate systematic review and meta-analysis in JMIR Mental Health, also from May 2026, set out to evaluate and compare the effectiveness of therapeutic interventions for problematic use of digital technology across various behavioral domains [2]. The authors state plainly that despite growing research, evidence on treatment effectiveness across digital behaviors remains fragmented [2]. Read those two reviews together and a pattern appears. We cannot yet compare the cures cleanly because we never agreed on the disease. The construct is scattered, so the trials are scattered, so the meta-analysis inherits the scatter.

This is not a call for one grand unified theory before anyone is allowed to help. It is a reason to be skeptical of any single-metric promise. The population in [1] is specific, learners in higher education, and a scoping review maps what exists rather than proving what works, so it cannot tell you which dimension of digital wellbeing matters most for you. What it can do is name the blind spot. When a study reports that heavier engagement links to poorer outcomes, ask which dimension it measured and which it left out.

There is a mechanism worth naming underneath all of this, and the Center for Humane Technology has put it sharply: in AI, what gets measured gets optimized [3]. Their argument is about AI capability metrics, but it generalises to attention itself. We spend our effort measuring how capable and powerful systems are while ignoring downstream consequences [3]. The same asymmetry runs through digital wellbeing research. We have precise instruments for the easy-to-count parts, minutes of use, notification volume, and blunt or absent instruments for the parts that actually constitute a good day: whether you felt present, whether the tool served your intentions or its own. CHT's proposed inversion, to measure what technology does to us rather than only what it can do [3], is the same move the BMC authors are making inside the wellbeing literature. Stop optimising the fragment. Define the whole first.

What is still contested is whether an integrative construct can be measured at all without collapsing back into its parts the moment someone builds a questionnaire. That tension is unresolved, and honest reviews say so. But the reframe changes what you do this week. Instead of asking how many hours you spent, ask a harder and more integrated question: after the screen, did you feel more yourself or less.

Research Radar

1. A scoping review in BMC Medical Education set out to systematically map the functional elements of digital wellbeing among higher education learners, arguing the field has studied related constructs like technostress and digital addiction in isolation rather than as one integrative whole [1]. The value is in naming the fragmentation, not in a single fix.

1. A PRISMA-guided systematic review and meta-analysis in JMIR Mental Health aimed to compare therapeutic interventions for problematic digital technology use across behavioral domains, and concluded that evidence on treatment effectiveness remains fragmented [2]. Comparing cures is hard when the underlying category is not settled.

1. A study of 491 South African university students tested whether smartphone addiction is a pathway linking social media use to psychological distress, using the Social Media Use Integration Scale among other measures [4]. It treats distress as a route, not just a correlation, echoing the case for looking beneath surface metrics.

One Thing To Try

Tonight, replace your usual screen-time check with one sentence in a notebook: how did today's device use leave me feeling. No number, no target. Just the integrated read that the trackers cannot give you.

Worth Your Attention

• The digital wellbeing among learners in higher education: a scoping review (BMC Medical Education) — the clearest recent argument for treating digital wellbeing as one construct, not a pile of separate problems [1].
• Therapeutic Interventions Targeted at Problematic Use of Digital Technology (JMIR Mental Health) — a meta-analysis honest about how fragmented the treatment evidence still is [2].
• We Measure What AI Can Do. We Should Measure What It Does to Us. (Your Undivided Attention) — the case for flipping our metrics from capability to human consequence [3].
• The Negative Mental Health Consequences of Social Media Use in South Africa (Behavioral Sciences) — a look at addiction as the pathway between use and distress [4].

We opened by asking what digital wellbeing actually is. The most useful answer this week is not a definition but a discipline: what gets measured gets optimized [3]. If we only measure the fragments, we will keep optimising fragments and wondering why the whole day still feels frayed. Name the whole first.