The paper by Gunther Eysenbach, titled "The Law of Attrition," discusses the phenomenon of high dropout rates in eHealth trials and the need for a "science of attrition." Eysenbach argues that high dropout rates are a distinct characteristic of eHealth trials compared to traditional drug trials, where high dropout rates are seen as detrimental to the credibility of the trial. However, in eHealth trials, particularly those conducted online and involving self-help applications, high dropout rates may be a natural and typical feature. The paper highlights the importance of measuring and analyzing usage metrics and determinants of attrition, including the characteristics of the subpopulation that eventually uses the application. Eysenbach proposes that attrition measures should be reported alongside efficacy measures from intention-to-treat (ITT) analyses to provide a more comprehensive understanding of the impact and uptake of eHealth interventions. The paper also introduces the concept of "attrition curves," which can be used to visualize the proportion of users who are lost to follow-up and those who do not drop out but stop using the application over time. Eysenbach suggests that understanding and describing patterns and predictors of attrition can advance the field of eHealth research. He also discusses the theoretical framework of diffusion research, particularly Everett M. Rogers' model of innovation adoption, and how it can be applied to eHealth interventions. Factors influencing attrition, such as the perceived relative advantage, compatibility, complexity, trialability, and observability of the innovation, are explored. Additionally, the paper proposes a "run-in and withdrawal" trial design to address high dropout rates in Internet-based trials. Finally, Eysenbach emphasizes the importance of overcoming pro-innovation bias, where researchers may overlook or underemphasize discontinuance due to their expectations of innovation success. He calls for a greater focus on attrition, uptake, and diffusion measures in eHealth research to better understand the real-life adoption problems of eHealth innovations.The paper by Gunther Eysenbach, titled "The Law of Attrition," discusses the phenomenon of high dropout rates in eHealth trials and the need for a "science of attrition." Eysenbach argues that high dropout rates are a distinct characteristic of eHealth trials compared to traditional drug trials, where high dropout rates are seen as detrimental to the credibility of the trial. However, in eHealth trials, particularly those conducted online and involving self-help applications, high dropout rates may be a natural and typical feature. The paper highlights the importance of measuring and analyzing usage metrics and determinants of attrition, including the characteristics of the subpopulation that eventually uses the application. Eysenbach proposes that attrition measures should be reported alongside efficacy measures from intention-to-treat (ITT) analyses to provide a more comprehensive understanding of the impact and uptake of eHealth interventions. The paper also introduces the concept of "attrition curves," which can be used to visualize the proportion of users who are lost to follow-up and those who do not drop out but stop using the application over time. Eysenbach suggests that understanding and describing patterns and predictors of attrition can advance the field of eHealth research. He also discusses the theoretical framework of diffusion research, particularly Everett M. Rogers' model of innovation adoption, and how it can be applied to eHealth interventions. Factors influencing attrition, such as the perceived relative advantage, compatibility, complexity, trialability, and observability of the innovation, are explored. Additionally, the paper proposes a "run-in and withdrawal" trial design to address high dropout rates in Internet-based trials. Finally, Eysenbach emphasizes the importance of overcoming pro-innovation bias, where researchers may overlook or underemphasize discontinuance due to their expectations of innovation success. He calls for a greater focus on attrition, uptake, and diffusion measures in eHealth research to better understand the real-life adoption problems of eHealth innovations.