A V E G E N

Loading

Types of data

There are broadly three types of data that Avegen leverages for better health outcomes of patients enrolled in the various care programs:

  • A web app or a mobile app usage data of a patient enrolled for the care program. Needless to say, the data aggregated and masked from any sort of personally identifiable information, used for an analytical view that would allow care program stakeholders to develop or evolve a care program that would have better outcomes for the enrolled users.
  • Tracking data & responses provided by users enrolled in the care programs as well as from HCPs involved.
  • An external public or a research data repository.

Care-program’s analytics

A detailed analytics dashboard illustrates user behaviour analysis, user acquisition and user retention sliceable by user segments, where users are patients enrolled in digital care programs. The analytics provides a direct view and measure of the success of the care program, allowing the care program managers to act and decide for app development or program enhancements for better outcomes. 

Broadly a digital care program requires a direct view of the following:

  1. User Acquisition
  2. User Retention 
  3. User Behaviour and User Engagement
  4. Outcome Analysis

User retention and engagement remains a challenge for most digital care programs. Many references point to the study that higher engagement is the likelihood of better outcomes for patients completing the care program. A detailed analytics view results in a well-thought product and care program development that enhances user retention and user engagement. 

Avegen also uses machine learning models to identify users at the brink of uninstalling the app, identifies the prominent underlying factors for such users, for actionable insight, such as an AI-powered push notification

Content Recommendation System (CRS) & Chatbot

The diversity of demands of content across user phenotypes calls for customization in a digital care program. Manually creating subtle variations of a generic digital care program for each user is not a viable option. 

Furthermore, the likelihood of a user liking a piece of content is temporal in nature. In simple terms, as a patient progresses through the disease for which s/he is being treated, the questions or information that s/he may seek would change depending upon the instantaneous prominent symptoms. Data science and artificial intelligence-based CRS provides the much needed solution. 

Additionally, CRS can be integrated with a chatbot. A chatbot is a user-friendly way to allow the users to input their requirements as if they are just chatting with a human. Content knowledge and FAQs can be stacked at the backend, and the output processing assisted by CRS can point either to the most relevant video or an FAQ. 

A digital care program with an effective CRS integrated chatbot, not only has a direct impact on user acquisition, retention and engagement, but also opens a door for a real-time data collection of user needs. 

Impact analysis

This is subtly different from the outcome analysis in the sense to measure the impact of the digital care programs on the results which were not the primary or a direct objective. For example, a digital maternal health care program might be designed to improve the awareness in pregnant women, but might also end up improving their mental health. Some more examples of impact analysis would:

  • Assessing if there is a sudden spike for keyword search that can be majorly attributed to a care program’s content or a topic that has been recently digitally marketed in a targeted geographical area
  • Assessing if patients exhibit more adherence to the recommended frequency of counsellor-patient visits  (behaviour impact of patients)
  • Assessing if people are more proactively getting diagnosed for the targeted disease, that benefits cure rate for patients due to in-time or early stage diagnosis

Tracking disease progression & risk predictions

Advanced data science and machine learning use disease progression and adverse condition risk prediction from the tracked symptoms to alert the HCPs using the HealthMachineTM about these patients in real-time. Additionally, the identified risks and patient disease progression stage prediction can be used as an input to the CRS. 

Authored by – Rajanikant Ghate

References
  1. https://www.jmir.org/2021/6/e27218/
  2. https://journals.sagepub.com/doi/full/10.1177/2055207618824727
  3. https://www.jmir.org/2019/9/e14567/ 
  4. https://www.prnewswire.com/il/news-releases/new-dariohealth-study-shows-increased-digital-therapeutics-engagement-leads-to-marked-reduction-in-blood-glucose-lower-high-readings-among-type-2-diabetic-seniors-301176772.html 
  5. https://www.ahajournals.org/doi/abs/10.1161/CIRCRESAHA.121.318224

Leave a Comment