.png)
Believe it or not, engagement is the missing link in digital health. Most member-facing health products collect data through apps, wearables, connected devices, or health assessments, but struggle to act on it in ways that feel relevant to the individual. The result is a familiar pattern: strong onboarding numbers, a spike in early activity, and a long tail of members who stopped engaging weeks after they signed up.
The problem is not that members do not care about their health. It is that most engagement strategies are not actually personalized. They are segmented, at best. Wearable data, combined with AI, changes what personalization at scale can look like, not by making content marginally more targeted, but by grounding every interaction in what is genuinely happening with that specific member's health right now. This is exactly what we are doing at Thryve with our Member Connection Engine!
The standard toolkit for member health engagement consists of push notifications, wellness challenges, content libraries, and points systems. But it was designed for populations, not for individuals. A campaign that performs well at the cohort level still misses the majority of the people in it, because the intervention is timed and framed for an average that no individual actually represents.
The structural problem is that most engagement logic is built on static inputs: demographic segment, diagnosis code, risk tier from last year's claims. These inputs describe who a member was at some prior point. They say nothing about what is happening with that member today.
This matters for a simple reason. A push notification reminding a member to exercise sent during a week when their sleep has been severely disrupted, and their HRV has dropped, is not a motivating nudge. It is noise. A risk flag raised for a member whose physiological signals have actually improved over the past month undermines trust. When the content does not match the context, members stop paying attention, and that disengagement is hard to reverse.
Wearable devices generate continuous physiological and behavioral signal between the touchpoints that traditional engagement systems can see. That continuity is what makes genuinely personalized engagement possible.
Several specific signals are most relevant at the engagement layer:
Continuous wearable data at member population scale generates a volume of signal that cannot be acted on through manual review or rule-based logic. AI is what converts raw signal into actionable engagement triggers.
Anomaly detection on individual baselines identifies members whose physiological signals are deviating from their own established pattern, rather than crossing a fixed population threshold. This is the technical foundation for personalized outreach that is actually relevant to the individual receiving it.
Engagement timing optimization uses member-level behavioral data to determine when an individual is most likely to respond to an outreach, which content type is most likely to land, and which channel they are most active on. The same intervention delivered at the right moment to the right member performs significantly better than the same intervention delivered to a segment on a fixed schedule.
Risk trajectory segmentation moves beyond static risk tiers to identify members whose wearable signals are trending toward a higher-risk state, before that transition has been captured by claims or clinical data. These members represent an intervention window: the opportunity to change a trajectory rather than respond to an outcome.
Content and recommendation matching uses current physiological context to surface content, challenges, or clinical recommendations that are relevant to what a member is experiencing right now. A sleep-focused intervention for a member in the middle of a sleep deterioration trend is more likely to engage than the same content delivered off-cycle.
Personalization at the engagement layer works differently depending on where a member sits in their health lifecycle. A single engagement model applied uniformly across a member population misses the distinct needs of each group.
This is the segment where prevention investment has the highest long-term return and the lowest urgency. Engagement here is about sustaining healthy behaviors, building trust in the platform, and detecting early signals before they become clinical risk. Wearable-based engagement for well members focuses on positive reinforcement, baseline tracking, and surfacing insights that help members understand their own data.
These are the members whose wearable signals or behavioral patterns indicate movement toward a higher-risk state, without a clinical event or diagnosis yet attached. This is the segment where timely, relevant outreach can change a trajectory. Engagement here needs to be specific enough to feel relevant, a generic wellness tip does not land for a member who is showing consistent sleep disruption and declining activity, and connected to a clear next step, whether that is a clinical touchpoint, a behavioral nudge, or an educational resource.
They require engagement that is coordinated with their care pathway, not layered on top of it independently. Wearable signals for this group are most valuable when they can be connected to medication adherence, care gap detection, and clinical outreach workflows, surfacing members whose physiological data suggests their condition is not well-managed even when their claims record shows they are receiving care.
Personalized engagement fails in predictable ways, and understanding those failure modes is as important as getting the technical architecture right.
Timing is everything. An alert that fires four hours after the window when it was actionable is not a late notification. It is a trust problem. If a member receives an engagement message that feels disconnected from where they are right now, the implicit message is that the platform does not actually know them. Engagement logic needs validity windows: if the triggering condition has passed or changed by the time the message sends, the message should not send.
Relevance requires context, not just data. Having access to a member's wearable data does not automatically make engagement relevant. The signal has to be interpreted correctly, connected to the right intervention, and communicated in language that makes the connection clear to the member. "Your recovery score dropped" is data. "Your sleep over the past five days suggests your body needs recovery time before your next high-intensity session" is context.
Trust is built through accuracy and restraint. Over-notification is as damaging as irrelevance. Members who receive too many alerts, or alerts that do not match their experience, stop engaging with all of them. The goal is not to maximize outreach frequency. It is to maximize the proportion of outreach that feels timely, accurate, and useful to the member receiving it.
Thryve's Member Connection Engine is a purpose-built solution for insurers and digital health providers who want to turn wearable data into sustained member engagement. It is the infrastructure layer that connects continuous health signals to the personalized journeys, milestones, and rewards that keep members active in a program year-round.
The engine works across four connected layers:
The results partners have seen with the platform reflect the underlying logic. Disengagement, not dissatisfaction, drives nearly half of insurance churn. By creating a continuous, personalized connection between a member's daily health behavior and the program they are enrolled in, Thryve reduces disengagement-driven churn by up to 40% and has helped partners like Techniker Krankenkasse and AOK grow their engaged member base tenfold.
Personalized health engagement is not a content problem or a channel problem. It is a data infrastructure problem. The platforms that get engagement right are the ones that have built the continuous, individual-level signal layer underneath it.
If you are building a member engagement product on wearable data and want to see how the Member Connection Engine works in practice, book a demo with Thryve.
Paul Burggraf, co-founder and Chief Science Officer at Thryve, is the brain behind all health analytics at Thryve and drives our research partnerships with the German government and leading healthcare institutions. As an economical engineer turned strategy consultant, prior to Thryve, he built the foundational forecasting models for multi-billion investments of big utilities using complex system dynamics. Besides applying model analytics and analytical research to health sensors, he’s a guest lecturer at the Zurich University of Applied Sciences in the Life Science Master „Modelling of Complex Systems“
.png)
.png)