HABIT-Works – how to professionally implement a computer-training!

Cognitive bias modification interventions have been around for a while, and many of them show promise as (add-on)- interventions. However, their implementation is a great challenge. Even though they seem to be easily delivered, they are hardly ever used in clinical practice. We are missing good implementation studies to bridge this gap. HABITworks gives an example of what such studies might look like. The paper describes the development of a CBM-I smartphone app, to be used as an augmentation to acute care. Stakeholders were involved in the development, and several versions were repeatedly tested and refined. First preliminary data were obtained regarding feasibility and acceptability, as well as adherence and target engagement.


Beard, C., Ramadurai, R., McHugh, R. K., Pollak, J. P., & Björgvinsson, T. (2021). HabitWorks: Development of a CBM-I smartphone app to augment and extend acute treatment. Behavior therapy, 52(2), 365-378


Commentary by Alexandra Werntz
University of Virginia & University of Massachusetts Boston, USA

Beard and colleagues’ article focuses on the development and initial pilot testing of HabitWorks, an adjunctive, transdiagnostic CBM-I mobile application (app) for individuals who are transitioning from acute psychiatric care to discharge. I was excited to read that the HabitWorks app was developed following a request from participants who completed a lab-based CBM-I program during a partial hospitalization program; they liked the idea of having the opportunity to continue practice flexible thinking after discharge. As a researcher and clinician, it’s nice to hear that there is a pull for CBM-based interventions by real users!

HabitWorks uses the word-sentence association paradigm (WSAP), and the team was thoughtful in how they implemented the training within the app. The app allows users to personalize the training by matching scenarios to their demographics. Although it sounds like it was labor-intensive to develop a set of 800 stimuli that allowed for personalization, that likely is the direction many of us in the CBM field are moving toward. I really appreciate that the team used advisory boards to initiate app development, then continued refining the program during three waves of initial pilot testing. The team listened to key stakeholders and participants on how to provide feedback on training performance, added GIFs to make the training engaging and rewarding, and created a dashboard for users to be able to track their own statistics on both their WSAP performance and symptom scores (the PHQ-9 and GAD-7 were available to complete within the app whenever users wanted to assess their symptoms). I also appreciate how the team thought critically about how to increase difficulty and change the training over time. A major strength of this work is that the team used key stakeholder feedback to develop and improve HabitWorks. As researchers, we make our best guess at what will seem credible and acceptable to users, however we may be wrong in our predictions. Implementing feedback based on users’ experiences only strengthens the intervention.

HabitWorks was developed in response to the challenge of transitioning from acute psychiatric care to discharge. This work is exciting because the research team is thinking about how CBM can be used to augment existing care pathways. In this context, CBM allows users to continue practicing the cognitive therapy skills that they learned during hospitalization. Beard and colleagues were able to identify a gap in treatment – in this case, structured practice of flexible thinking – and create an app to address the gap. This work should encourage us all to think about how CBM trainings can be implemented in other settings, such as outpatient mental health clinics (to individuals on a waitlist or those who have terminated), schools, or primary care settings. I am excited about how we can encourage individuals to practice ways of thinking flexibly on their own schedule. Learning to think in new ways is a challenge and takes practice, and programs like HabitWorks allows individuals struggling with mental illness to have access to on-demand practice and support.


Commentary by Janna Vrijsen
Radboud Medical Centre & ProPersona, The Netherlands

Beard and colleagues (2021) present the development of a smartphone-based CBM-I app called HabitWorks as possible augmentation tool to acute care and after discharge. They obtained input from stakeholders, tested the app in several feedback rounds, and collected preliminary quantitative and qualitative feasibility and acceptability data in an open trial. All and all, the first data shows that a more thorough evaluation of the feasibility, acceptability and also efficacy of HabitWorks is warranted.
On a general note, it is exciting that our experimental CBM innovations are slowly but surely making their way to standard clinical care, resulting in CBM paradigms fit for implementation. Such true translational initiatives are needed and, in my experience, welcomed by the psychiatric field, although we can work on increasing the understanding of CBM and its potential among our clinical colleagues. Furthermore, taking a transdiagnostic approach to CBM – as done in this study – fits with daily clinical practice, and is hence an important issue to consider when thinking about implementation.

The study itself

Beard and colleagues based the development of this smartphone-based CBM-I intervention on suggestions made by the main stakeholders, the patients. This is such an elegant and valuable approach to innovating on CBM, and in that sense this paper can inspire our field. As with many translational steps, the move from lab to clinic requires letting go of some control. The authors balance the scientific rigor with the demands of clinical practice very well. They follow a roadmap for behavioral therapy development setting a priori benchmarks, hence nicely providing a framework and staging towards implementation. The balance between rigor and practical use is also reflected in details of HabitWorks itself e.g. the use of personally relevant stimuli. This is an important feature of the CBM-I intervention as it can be expected to increase user engagement as well as possible transfer effects to ‘daily life’ processing of information. By basing the selection of stimuli on pre-defined worry domains and demographic details, they again balance control over the trained stimuli with room for individual differences.
The authors included many open science practices enabling replication, such as pre-registration. They also explicitly mention details of data security and features to secure patients’ privacy. Increasing inclusiveness of the app by using GIFs representing an array of ages and ethnoracial backgrounds is also an elegant touch.
While reading the paper, I got curious to know how much feedback the authors received during the iterative development process and what the advisory board’s level of user engagement was. Something else I always struggle with when providing feedback during CBM trials, is the use of the term ‘Correct’. Would it create more awareness about the purpose of CBM (and actually be more accurate) if we would switch to ‘Healthy response’ or something along those lines, I wonder?

Innovating psychiatric care

The authors take and exciting and timely step towards clinical applicability of CBM-I. Specifically, the positioning of CBM as an evidence-based mHealth tool and hence the embedding of CBM within the clinical (treatment) flow is impressive. A particularly strong feature of the study design is the combination of daily prompts during admission and the active scheduling of the at-home CBM-I mHealth sessions. Especially because the authors target the high-risk and for patients often insecure and disruptive phase after discharge. It is interesting that patients differed greatly in how they valued (and made use of) having contact with the staff while using HabitWorks during admission. The adherence during the post-discharge phase was relatively low, but conform other app usage as far as I can tell. Both issues point to also embedding features for personalizing the CBM-I protocol itself.
This paper stimulates us as a field to think about the positioning of CBM in relation to the psychological/psychiatric treatment provision. This includes the positioning of CBM during the waitlist period; providing support for vulnerable individuals and possibly increasing receptiveness of treatment. Secondly, the augmenting effect of CBM on psychotherapeutic (cf. work by Nader Amir, Courtney Beard herself, and several others) and also psychophamacological treatment (cf. the work of Catherine Harmer, indicating cognitive bias as possible mechanism of change of antidepressant medication) is worthwhile exploring more. Thirdly and in line with the current paper, CBM during after-care in the transition to home and/or another care provider seems feasible and valued by the patients themselves.

Scroll to Top