Boosting adherence in AHA
The digitalization of healthcare interventions has the potential to alleviate the burden of health systems caused by the increasing ageing population and chronic diseases. However, these types of technological solutions require high rates of adherence to be effective and as well as in any other application, users frequently abandon the solution due to various reasons.
In a context like this, the early identification of users with risk of lower adherence rates and usage patterns that indicate risk of dropout is an invaluable opportunity to apply tailored intervention strategies aimed at recovering from disengagement. Traditionally, acceptance was assessed using static methodologies such as SUS or UTAUT, therefore, medium and long-term acceptance and engagement have not been systematically analysed. The objective is to predict early dropouts.
In this challenge, participants are given access to a dataset of more than 300 users that have tested the impact of a digital AHA app to improve their quality of life for at least 6 months in the MAHA network (Moving Active & Healthy Ageing in Madrid).
The dataset represents the baseline (before to start to use the app) and final evaluation (after 6 months of usage) questionnaires (quality of life and acceptance of the solution) from the participants, internal measurements of the solutions, resulted from the usage of each of the functionalies and interventions in the app. these results evaluate the success and the adherence of the intervention; and the log phenotype, that is the elements with the participant interacts when using the application (buttons, menus, etc). During this six-month period, the participants were using the application addressing specific problems of their ageing profile (cognitive, physical, socialization or depressive). They were asked to use the solution frequently (at least twice a week) according to their needs.
The goal is to create the best possible models to identify early dropouts One possible application of such a model could be minimizing adherence problems that can compromise the impact of these types of healthcare solutions, observing the application’s measurements, usage patterns and evaluation results.
How to participate
- Register below, you will receive a confirmation from the Science Challenge Committee.
- Access here to the dataset.
- Download the sample file (Phyton .py or Matlab .m) and develop your solution.
- Submit here your modified compressed entry file to firstname.lastname@example.org. You will have 15 attempts (5 in phase I, 10 in phase II). You will receive (limited) feedback to help you improve your models with each attempt.
- Submit here an overview paper with the best solutions to IUPESM WC 2022 (please select “scientific challenge” after login).
- Attend IUPESM WC 2022 (June 12-17) and present your work there (at least one team member must be registered and attend).
The challenge comprises two phases with the following deadlines in GMT (UTC):
Phase I (max. 5 attempts)
From 1 October 2021 at 10:00, to 17 December at 23:59
Phase II (máx. 10 attempts)
From 3 January 2022 at 10:00, to 10 March at 23:59
In the interest of fairness to all participants, late entries will not be accepted. Submissions not used in Phase I cannot be transferred to Phase II. Submissions that cannot be scored due to missing components or inadequate formatting, will not be counted towards the limit of attempts.
To be eligible, you must follow these rules
- For registration, fill in team name, members’ names and emails.
- You must submit at least one attempt that can be scored in each phase and within the marked deadlines.
- The submission can be in Matlab or Python (name your main file challenge.m or challenge.py respectively). This submission must be self-contained: if some special open-source, non-standard toolbox or library is needed, they must be provided (All commercially available toolboxes from Mathworks for MATLAB R2020a are included). Please follow the detailed information on data, input and output formats provided in the example files.
- If the entry includes all the necessary files and they are executed as expected, it will be evaluated. The evaluation will take place on a weekly basis. When the test is complete, you will automatically receive an email with your provisional score or an error message if it did not work. Please read it carefully and correct any problems before resubmitting it.
- After Phase I, you will receive additional Phase II datasets. You can use these to improve your models for Phase II, if you wish.
- In Phase II (before 10 March 2022), you must submit a full paper describing your work for the Science Challenge following the IUPESM rules, including the best score you have achieved. You will receive a confirmation email if your paper has been accepted from IUPESM WC 2022 by 15 March 2022.
- You must submit the final paper presentation and copyright transfer forms by 20 March 2022.
- Please do not submit an analysis of this year’s Scientific Challenge data to other conferences or journals until after IUPESM WC2022 has taken place, so the competitors are able to discuss the results in a single forum. We expect the submission of an overview paper with the best solutions to a prestigious Scientific Journal, co-authored by the corresponding participants of the challenge and taking into account the publications and discussions at IUPESM WC2022.
Registration is still open! (until 16 December)
Please, only one member of the team needs to register.
The dataset provided for this challenge comes from the ACTIVAGE pilot at the Madrid Deployment Site. The data was collected through a set of mobile applications that make up the Madrid Active and Healthy Ageing (MAHA) app. The information consists of various assessments on the quality of life, acceptance and daily use of various applications and technologies from real users over the course of the project, so that the evolution can be measured.
The MAHA database is separated into 9 DATASETS, all interconnected through the ID field which represents the unique identifier of each user.
You can access an example of this database here:
Patient sociodemographic information, includes also the type of device they use to access the applications and solution in the MAHA as well as the information about their engagement with the solution (dropout or not).
QoL data that represent the self-perceived quality of life and well-being status of the user at baseline of the study and at the end (six months later).
Data of usability and acceptability assessment of the solutions. Two separate data set are included here:
UTAUT: that follows the UTAUT model approach results to assess the acceptability of the solution at the end of the study.
SPQ: that includes fields that represent the self-perceived influence of the tested solutions on the users’ quality of life.
Includes the logs of usage of the different applications.
Represent internal measurements of the solutions depending on their final purpose (i.e. duration of the session, difficulty of the session, final score, session finalized, etc.).
Moving Active & Healthy Ageing
MAHA is located in the fruitful and internationally renowned ecosystem: Madrid Reference Site, which improves the autonomy of the elderly and their quality of life by establishing a strong network within the neighbourhoods, contributing to the sustainability of the healthcare system.
MAHA’s main objective is to prolong and support the independent living of older adults through technology in their own environments. Through ACTIVEAGE.org, MAHA continues the legacy of the large European ACTIVAGE Pilot in Madrid, consolidating a network of stakeholders and institutions where healthy ageing and the development of innovative and personalised solutions intersect to promote the quality of life of people within their homes and across the city.