The ad-tech industry has made significant progress in creating tools and methods to use data to optimise ad placements and provide the appropriate information to the appropriate audience at the appropriate time. These lessons can all be useful in the field of digital healthcare. Digital healthcare is a quickly expanding, popular industry with the goal of meeting the daily requirements of people everywhere, enhancing the health and wellness of millions of people, and making healthcare more advanced and accessible than ever.
Using deidentified data whenever possible, digital transformation in healthcare entails working with data at scale, sometimes to provide real-time insights on several clinical issues for many populations. Ad-tech, finance, gaming, and other digital start-ups have gained greatly from data, analytics, and AI in the online tech sector, whereas conventional healthcare organisations have been on the other side of the street. There is much to be learned from their best practises and technologies, despite the fact that they are not directly involved in the business of saving lives and enhancing people’s health.
Because the proper ad must be served to the right individual within milliseconds of that person’s entry into a particular web page or mobile application, the ad-tech sector requires that data and analytical tools be maintained up to date. Additionally, it must be based on that person’s past experiences and preferences. Because so many of these advertisements are sent to their intended audience every day in the billions, even the smallest improvement—or error—can have a significant financial impact.
The right data must travel quickly across the network in order for it to be prepared, cleaned, processed, utilised to train AI models, produce predictions and insights, and deliver adverts for the entire operation to run smoothly. In order to guarantee effective and responsible use, avoid mistakes, and minimise potential losses, all of this needs to be carefully vetted and monitored.
I entered the healthcare industry after a protracted career in the advertising technology field. Finally, a crucial insight emerged: why not attempt to implement some of these extremely effective techniques into my present digital healthcare company? Such a multidisciplinary approach might produce some extremely intriguing and out-of-the-box outcomes.
Utilising ad tech to solve a problem in digital healthcare
Ad-tech professionals want to be able to find viewers more precisely and economically than ever before. The capacity to create larger audiences out of smaller consumer segments is utilised and unlocked by lookalike models, which has already been profitable for advertisers. Extensive AI models are internally trained to comprehend the traits of that subset of successful clients using pertinent statistics.
Advertisers can reach audiences they might not have otherwise thought to target thanks to lookalike modelling. It has successfully established itself as a crucial AI-based method that has a lot to offer a complete sector.
Returning to digital healthcare now. The right consumer demographic must be reached as the first stage in the adoption of a digital healthcare service. The majority of this group will be able to gain from our solution. Assume we have developed an application that helps individuals with Crohn’s disease better manage their chronic condition. We can then get in touch with these people to convince them to download and use our software.
Our app might be quite helpful for some people. They will take pleasure in interacting with the range of features, be able to appropriately control their condition, and prevent deterioration. Some users can cease using the programme soon after it is installed. It might be due to the content we offer, a UX issue, or simply because they don’t have the time to invest. Can we narrow our focus to the patients who will use and profit from our app the most?
We can develop several criteria (or traits) that will characterise this segment of users who might profit from the app with the use of lookalike modelling. Using deidentified data where necessary for the protection of health plan members, we may responsibly analyse their medical history (prior diagnoses, procedures, prescription drugs, test results, and more), app engagement metrics, demographics, and behavioural profiles. Then, we turn each of these traits into a dense vector for each individual using a variety of Deep Learning approaches based on natural language processing (NLP). This vector is an individual’s digital fingerprint. Using this information, we can search for individuals whose fingerprints like those of our “successful” Crohn’s patient the most.
An example of an embedding cloud for Crohn’s patients. Each patient is represented by a dot. Clinical histories that are similar bring people together. All people with serious comorbidities are highlighted in red.
Once our strategy is established, we are able to do more than just reach out to new audiences. We may be able to access a completely new level of analytical power now that we have our consumer “map” in place.
We can divide Crohn’s disease sufferers into subgroups based on characteristics like demographics and severity using clustering techniques. Then, we may modify our software to fulfil the requirements of each segment, increasing the likelihood that we’ll benefit our consumers.
What does the future hold?
Just the tip of the iceberg exists here. Healthcare faces numerous problems that have yet to be resolved, but leveraging the perspective of another industry may help find solutions. To assist healthcare in making this shift, there will also be a need for creative thinkers and leaders from numerous industries.
The ability to transition from one-off, ad-hoc analysis performed and presented in spreadsheet format to developing automated Extract/Transform/Load (ETL) processes, creating cloud-based feature stores, and data warehouses, which could then serve as data sources for models and analytics presented via interactive visualisation tools, is likely to define the future of digital healthcare.
For programme management or intervention reasons, these may eventually be converted into insights and actions that could be automatically sent to the appropriate stakeholder or individual. The future is bright, but such projects are already taking shape in the healthcare sector. It may be good to seek inspiration from related IT sectors to narrow gaps and get there more quickly.
