The Potential of Small Effects at the Right Time, on a Large Scale: Commentary on Linardon et al. (2025)

ABSTRACT

The meta‐analysis of self‐help intervention for eating disorders (ED) by Linardon and colleagues showed significant, albeit small, effects favoring self‐help over the control condition on depression, anxiety, distress, and self‐esteem. Despite modest effect sizes, pure self‐help offers the potential for high accessibility at low cost, which may lead to a meaningful impact on public health in terms of mental health symptoms that are often co‐occurring with ED. There are opportunities to present and package pure self‐help in more creative ways than what is currently available (e.g., by integrating brief instructional and experiential videos, infographics, storytelling, and compelling patient narratives). To fully harness the potential of pure self‐help, disruptive innovations are necessary in both the packaging and delivery methods. These innovations can help to accommodate various needs, learning styles, and preferred delivery formats. A consortium dedicated to pure self‐help for symptoms of ED and its prevention can play a vital role in testing, delivering, collecting big data, understanding moderators of outcomes, and facilitating adaptation and further development, thereby improving access to these interventions and leading to better mental health.

The meta‐analysis conducted by Linardon et al. (2025) revealed that self‐help interventions for individuals at risk of eating disorders (ED) or exhibiting symptoms of ED have favorable, albeit small, effects on secondary mental health and well‐being outcomes, including depression, anxiety, distress, and self‐esteem. Although the significant findings for clinically diagnosed samples regarding distress, quality of life, and psychosocial impairment were less robust, the study's results are important and illuminating in terms of informing future research and implementation. Despite several limitations noted by the authors, the findings warrant attention. In this commentary, we aim to highlight the following key points:

1. Potential for Substantial Public Health Impact

Linardon et al. (2025) suggest that small effects may have a substantial public health impact, given the ease with which pure self‐help interventions can be disseminated. This notion cannot be overstated. Although guided self‐help generally produces better outcomes than pure self‐help for ED symptoms, the effect sizes are relatively similar, and the ease of implementing pure self‐help makes it a more viable candidate for large‐scale implementation of early treatment of uncomplicated cases with ED symptoms. As a research community, we must ask ourselves why pure self‐help has not been implemented on a large scale yet. One possible explanation lies in the recommendation of several national guidelines for ED treatment, which suggest guided self‐help as the first step. This recommendation, although completely adequate and rooted in empirical findings, may have diverted attention and investment away from pure self‐help. Guided self‐help is available within the national healthcare systems in some countries, but not freely accessible within a non‐disease or health‐promotive framework. Such availability and framing can be more easily achieved for pure self‐help, and thus increasing access to intervention. Given that pure self‐help shows modest but still reasonable effects not only on ED symptoms (Perkins et al. 2006) but also on secondary mental health outcomes, as demonstrated by Linardon et al. (2025), it is time to reevaluate its potential to make a significant impact on public health at a relatively low cost at an early stage. In assessing efficacy, cost‐effectiveness, and the potential for broad implementation of pure self‐help and increasing access, several factors should be considered, including how pure self‐help is presented, the channels through which it is delivered, adaptation to different needs of target groups, moderators of outcome, and the future use of artificial intelligence (AI).

2. Packaging of Pure Self‐Help

Initial studies on pure self‐help for ED typically provided patients with a book or equivalent written material in other formats, expecting them to read the material independently and follow the instructions. The advent of internet‐delivered self‐help largely followed the same format, with written material provided weekly as PDF documents, accompanied by automatic reminders and occasional requests for patient responses without feedback. The potential of using pedagogical packaging of pure self‐help, incorporating brief instructional and experiential videos, infographics, storytelling, influential patient experiences, or other innovative strategies, remains to be evaluated. Although reading text in various formats may be suitable for a limited percentage of individuals in need of ED treatment, increased flexibility in packaging the intervention may make it more appealing to a larger group, leading to higher completion rates, reduced dropout, and better outcomes.

3. Modes of Delivery

Self‐help interventions for ED are often delivered as books, internet platforms providing similar text, and occasionally combining text with pictures and videos, or as mobile apps. To further elucidate the efficacy of pure self‐help (i.e., reduced incidence of ED, and long‐term effects on correlated symptoms) and to broadly implement self‐help when reasonable efficacy and cost‐effectiveness are established, disruptive innovations are necessary. Many principles and processes of change involved in these programs can be delivered through alternative channels, adapted for specific target groups. Social media platforms constitute a potentially efficient way of reaching the target groups to provide self‐help for both prevention and early treatment of ED symptoms that we now know have an impact on other mental health symptoms that are often co‐occurring with ED.

As the mean survival time of mental health apps is short, and most users stop using them after the fifth interaction, future self‐help app design for ED should focus on limited interactions. Research on continued app use and its determinants is helpful in this context, but the general trend is a preference for brief, highly instructional, or entertaining apps. Methods used in single‐session interventions for ED, which have shown efficacy, might be the strongest candidates for designing future self‐help apps for ED. Short movies, gamification, and enhancing ED literacy and advocacy, delivered as part of apps primarily intended for entertainment, are other examples of disruptive innovations for delivering pure self‐help for ED, or to use its potential for early prevention. A key question in this context is “Who should organize and build the foundation for implementation of such efforts, after adequate research on their efficacy?” This raises the idea of a consortium that will be discussed later.

4. Adaptation

ED can emerge at any age and in any socioeconomic group. Neither prevention nor treatment is typically a one‐time, limited intervention to stop the emergence of ED or achieve full recovery; rather, it is a process. While some risk factors for ED may exert a linear effect across time, others may be more critical at specific points than others. Individuals with ED symptoms may also be more or less open to considering their symptoms as problematic and seeking self‐help at different times. Consequently, an array of reasonably efficacious self‐help programs should be available to meet the diverse needs and susceptibility of individuals at various levels of risk, ages, and contexts. The interventions should be tailored based on the potentially specific needs of groups or individuals. Some adaptations may increase the likelihood of individuals entering and engaging in such programs. For example, as homosexual and bisexual men have a higher risk of ED compared with heterosexual men, self‐help programs should consider factors that make these programs more appealing to this group and also address specific risk factors, such as minority stress. Similarly, given the marked co‐occurrence of autistic traits in ED, preventive self‐help programs for this group should consider their unique needs and modes of delivery with specific attention to how the content is communicated to avoid misunderstanding. Efficient and clear‐cut rules, straightforward storytelling, and thoughtful use of psychoeducation with unambiguous illustrations might be more helpful than demanding frames such as cognitive dissonance, Socratic style, or complex metaphors. In the process of further adapting self‐help programs, it will be crucial to engage the end users and collaboratively work on refining the programs. This will provide an excellent opportunity to consider diversity and sensitively incorporate elements that the end users perceive as important while also considering the efficacy and overall coherence of these elements through actual iterations and data.

It is essential to note that producing hundreds of self‐help interventions may not be the most effective approach. It will soon become impossible for the individual in need of self‐help to find the right one. A first step should involve mechanistic research into the maintaining variables in ED and processes of change. Next, a mechanistically developed program can be adapted based on the specific needs of subgroups of individuals with specific characteristics (not more than a few versions, all available at the same well‐known resource center/consortium).

Self‐help does have a role to play in clinical settings as well. The accumulated knowledge on the implementation of self‐help and automated tools in addressing mental health (e.g., Chan et al. 2019) combined with disruptive innovations may provide valuable insights into the adaptation and evaluation of pure self‐help for clinical populations. This may be successfully achieved within a consortium focused on self‐help for ED.

5. Moderators of Outcome in Self‐Help Treatment

Although considering moderators of outcome in ED treatment can be seen as part of the adaptation process in developing future self‐help interventions, it deserves separate discussion, as a new infrastructure may be needed to identify these moderators on a large scale before they can be used to adapt interventions. Our knowledge of moderators of outcome is limited because of the small samples in our trials, which are often not representative of the target population at large, low power, and limited variance of the potential moderators. The outcome variables are also often defined in different ways, as does the operationalization of moderators in different trials. Perhaps, it is time to establish a consortium for large‐scale investigation and implementation of pure self‐help. If the most efficacious self‐help prevention and intervention programs are culturally adapted and made available in different languages, pending anonymous sharing of data, then cumulative big data can help us understand moderators of outcome more robustly, while also making the currently best interventions available on a large scale. The AI tools are increasingly simplifying the process of translation and adaptation of both interventions and assessment tools. Accumulated data can then be used, partly by AI models, to further enhance and adapt such programs.

In addition to understanding the moderators of outcome, awareness of potential iatrogenic effects and collection of such data are also important to ensure that these effects can be detected and addressed. The notion of iatrogenic effects raises a series of other ethical issues that need consideration and are beyond the scope of a short commentary. Ethical guidelines and value‐based decision‐making, with the ultimate goal of improving the health of people and safeguarding such an initiative from financial interests and profit, should be at the core of building a consortium. Finally, the issues surrounding who controls the big data and how we ensure that those included in the consortium represent diverse consumers as well as diverse clinical and research representatives are examples of other key questions in need of another commentary paper.

6. Artificial Intelligence

The rapid development of AI presents opportunities to accelerate the development of informed self‐help programs. Although most AI models are currently not sophisticated enough to provide brief, sensitive, context‐specific, and adequate feedback and guidance on time, the current and future path of development in AI is promising. Perhaps single‐session interventions for ED (Schleider et al. 2023), such as those focusing on a critical view of the thinness or beauty ideals of today's society, or focusing on regular and healthy eating, accompanied by AI‐based feedback, could be a suitable starting point for incorporating AI into future prevention and intervention self‐help programs, without expecting AI to act at the same level as a trained therapist. AI‐informed or assisted self‐help can still be framed as part of the pure self‐help, as there is no therapist or facilitator involved, but the individual is using available resources to help herself/himself. AI can be used to help build communities and online support groups and will be a reliable source for curating and delivering educational content. Within a consortium with a focus on pure self‐help for ED, AI can also help to analyze large data sets and identify trends.

7. Concluding Remarks

Pure self‐help brings the promise of high accessibility at low cost, and thus the potential of a significant impact despite small effect sizes, if its efficacy and cost‐effectiveness are further established. The meta‐analysis by Linardon et al. (2025) further shows its value given the significant effect of self‐help on mental health symptoms that are often comorbid with ED. Pure self‐help for ED can be presented and packed in more innovative ways than currently by incorporating, for example, brief instructional and experiential videos, infographics, storytelling, and influential patient experiences in its presentation. Disruptive innovations are needed to fully utilize its potential in both packaging and modes of delivery. Self‐help can be adapted for people with different needs, learning styles, and preferred modes of delivery, as well as pedagogical presentations (level of instructions, use of audiovisual material, level of details and reference to facts, etc.). A consortium focusing on pure self‐help can contribute to testing, delivery, big data collection, understanding the moderators of outcome, and thus adaptation and further development, as well as increased access to such interventions. Finally, AI may play a role in several steps of the process of further evaluating and increasing access to evidence‐based self‐help for ED symptoms and for prevention of ED without reaching the level of acting almost as a human therapist.

Author Contributions

Ata Ghaderi: conceptualization, writing – review and editing, writing – original draft, methodology.

Conflicts of Interest

The author declares no conflicts of interest.

Ghaderi, A. 2025. “The Potential of Small Effects at the Right Time, on a Large Scale: Commentary on Linardon et al. (2025).” International Journal of Eating Disorders 58, no. 9: 1686–1689. 10.1002/eat.24472.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analysed during the current study.